e146e2a86be6078015fce4e06a1e1bf280039c4b
[gcc.git] / gcc / ChangeLog
1 2013-09-03 Meador Inge <meadori@codesourcery.com>
2
3 Revert:
4
5 2013-08-30 Meador Inge <meadori@codesourcery.com>
6
7 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8
9 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10
11 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
12 function descriptor.
13
14 2013-09-03 Richard Biener <rguenther@suse.de>
15
16 * tree-affine.c (add_elt_to_tree): Fix association issue,
17 avoid useless converts and make sure to always return a
18 properly typed result.
19
20 2013-09-03 Richard Biener <rguenther@suse.de>
21
22 PR middle-end/57656
23 * fold-const.c (negate_expr_p): Fix division case.
24 (negate_expr): Likewise.
25
26 2013-09-03 Richard Biener <rguenther@suse.de>
27
28 PR lto/58285
29 * tree-streamer-out.c: Include tm.h.
30 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
31
32 2013-09-03 Jan Hubicka <jh@suse.cz>
33
34 * tree-profile.c (tree_profiling): Cleanup CFG when done.
35
36 2013-09-03 Alan Modra <amodra@gmail.com>
37
38 * config.gcc (powerpc*-*-linux*): Add support for little-endian
39 multilibs to big-endian target and vice versa.
40 * config/rs6000/t-linux64: Use := assignment on all vars.
41 (MULTILIB_EXTRA_OPTS): Remove fPIC.
42 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
43 * config/rs6000/t-linux64le: New file.
44 * config/rs6000/t-linux64bele: New file.
45 * config/rs6000/t-linux64lebe: New file.
46
47 2013-09-02 Jan Hubicka <jh@suse.cz>
48
49 * ipa-inline-transform.c (inline_transform): Do not
50 optimize_inline_calls when not optimizing.
51
52 2013-09-02 Jan Hubicka <jh@suse.cz>
53
54 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
55 duplicated nodes for assembler names.
56 * symtab.c (symtab_unregister_node): Do not attempt to unlink
57 hard registers from assembler name hash.
58
59 2013-09-02 Jan Hubicka <jh@suse.cz>
60
61 * ipa-split.c (execute_split_functions): Split externally visible
62 functions called once.
63
64 2013-09-02 Martin Jambor <mjambor@suse.cz>
65
66 PR ipa/58106
67 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
68 linked list. When finding the correct duplicate, also consider also
69 the caller in additon to its inlined_to node.
70
71 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
72
73 * config/aarch64/aarch64-simd-builtins.def
74 (dup_lane_scalar): Remove.
75 * config/aarch64/aarch64-simd.md
76 (aarch64_simd_dup): Add 'w->w' alternative.
77 (aarch64_dup_lane<mode>): Allow for VALL.
78 (aarch64_dup_lane_scalar<mode>): Remove.
79 (aarch64_dup_lane_<vswap_width_name><mode>): New.
80 (aarch64_get_lane_signed<mode>): Add w->w altenative.
81 (aarch64_get_lane_unsigned<mode>): Likewise.
82 (aarch64_get_lane<mode>): Likewise.
83 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
84 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
85 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
86 (VCON): Change container of V2SF.
87 (vswap_width_name): Likewise.
88 * config/aarch64/arm_neon.h
89 (__aarch64_vdup_lane_any): New.
90 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
91 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
92 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
93
94 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
95
96 PR middle-end/56382
97 * expr.c (emit_move_complex): Do not move complex FP values as parts if
98 the source or the destination is a single hard register.
99
100 2013-09-02 Richard Biener <rguenther@suse.de>
101
102 PR middle-end/57511
103 * tree-scalar-evolution.c (instantiate_scev_name): Allow
104 non-linear SCEVs.
105
106 2013-09-02 Richard Biener <rguenther@suse.de>
107
108 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
109 arithmetic to sizetype.
110
111 2013-09-02 Bin Cheng <bin.cheng@arm.com>
112
113 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
114 Find auto-increment use both before and after candidate.
115
116 2013-09-02 Marek Polacek <polacek@redhat.com>
117
118 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
119
120 2013-09-01 Jan Hubicka <jh@suse.cz>
121
122 * Makefile.in: Add ipa-profile.o
123 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
124 * cgraph.c (struct cgraph_propagate_frequency_data,
125 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
126 ipa-profile.c; replace cgraph_ by ipa_ prefix.
127 * cgraph.h (cgraph_propagate_frequency): Remove.
128 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
129 (inline_update_callee_summaries): Update.
130 * ipa-profile.c: New file.
131 * ipa-utils.h (ipa_propagate_frequency): Declare.
132 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
133 data-streamer.h, value-prof.h
134 (symtab_remove_unreachable_nodes): Update profile.
135 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
136 account_time_size, cmp_counts, dump_histogram,
137 ipa_profile_generate_summary, ipa_profile_write_summary,
138 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
139 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
140 Move to ipa-profile.c
141
142 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
143
144 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
145
146 2013-09-01 Jan Hubicka <jh@suse.cz>
147
148 * common.opt (fdevirtualize-speculatively): New function.
149 * invoke.texi (fdevirtualize-speculatively): Document.
150 * ipa-devirt.c: Include ipa-inline.h
151 (likely_target_p): New function.
152 (ipa_devirt): New function.
153 (gate_ipa_devirt): New function.
154 (pass_data_ipa_devirt): New static var.
155 (pass_ipa_devirt): Likewise.
156 (make_pass_ipa_devirt): New function.
157 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
158 (common_handle_option): Disable devirtualization when
159 value range profiling is available.
160 * passes.def (pass_ipa_devirt): Add.
161 * timever.def (TV_IPA_DEVIRT): New timevar.
162 * tree-pass.h (make_pass_ipa_devirt):
163
164 2013-09-01 Iain Sandoe <iain@codesourcery.com>
165
166 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
167 include sanitize(undefined).
168
169 2013-08-31 Diego Novillo <dnovillo@google.com>
170
171 * Makefile.in (TREE_CORE_H): Define.
172 (TREE_H): Use.
173 (GTFILES): Add tree-core.h.
174 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
175 size the array.
176 * tree-core.h: New file.
177 Move all data structures, enum, typedefs, global
178 declarations and constants from ...
179 * tree.h: ... here.
180
181 2013-08-31 Jan Hubicka <jh@suse.cz>
182
183 * bulitins.c (expand_builtin): Do not early exit for gcov
184 instrumented functions.
185
186 2013-08-31 Marek Polacek <polacek@redhat.com>
187
188 * ubsan.c: Include tm_p.h.
189
190 2013-08-31 Jan Hubicka <jh@suse.cz>
191
192 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
193 warning.
194
195 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
196 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
197 * tree-cfg.c (verify_gimple_label): ... here.
198 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
199 (ipa_merge_profiles): New function.
200 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
201 (lto_input_function_body): Likewise.
202 * ipa-utils.h (ipa_merge_profiles): Declare.
203 * lto-streamer.h (lto_input_function_body): Update prototype.
204 (emit_label_in_global_context_p): Remove.
205 * lto-symtab.c: Include ipa-utils.h
206 (lto_cgraph_replace_node): Use ipa_merge_profiles.
207
208 2013-08-31 Jan Hubicka <jh@suse.cz>
209
210 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
211
212 2013-08-31 Jan Hubicka <jh@suse.cz>
213
214 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
215
216 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
217
218 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
219 "cmp" RTX before signed_comparison_operator check to account
220 for "code" changes.
221
222 2013-08-30 Jan Hubicka <jh@suse.cz>
223
224 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
225 (detect_type_change_1): Rename to ...
226 (detect_type_change): ... this one; early return on non-polymorphic
227 types.
228 (detect_type_change_ssa): Add comp_type parameter; update
229 use of detect_type_change.
230 (compute_complex_assign_jump_func): Add param_type parameter;
231 update use of detect_type_change_ssa.
232 (compute_complex_ancestor_jump_func): Likewise.
233 (ipa_get_callee_param_type): New function.
234 (ipa_compute_jump_functions_for_edge): Compute parameter type;
235 update calls to the jump function computation functions.
236
237 2013-08-30 Teresa Johnson <tejohnson@google.com>
238 Steven Bosscher <steven@gcc.gnu.org>
239
240 * cfgrtl.c (fixup_new_cold_bb): New routine.
241 (commit_edge_insertions): Invoke fixup_partitions.
242 (find_partition_fixes): New routine.
243 (fixup_partitions): Ditto.
244 (verify_hot_cold_block_grouping): Update comments.
245 (rtl_verify_edges): Invoke find_partition_fixes.
246 (rtl_verify_bb_pointers): Update comments.
247 (rtl_verify_bb_layout): Ditto.
248 * basic-block.h (probably_never_executed_edge_p): Declare.
249 (fixup_partitions): Ditto.
250 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
251 * bb-reorder.c (sanitize_hot_paths): New function.
252 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
253 sanitize_hot_paths.
254 * predict.c (probably_never_executed_edge_p): New routine.
255 * cfg.c (check_bb_profile): Add partition insanity warnings.
256
257 2013-08-30 Meador Inge <meadori@codesourcery.com>
258
259 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
260
261 2013-08-30 Marek Polacek <polacek@redhat.com>
262
263 * Makefile.in (ubsan.o): Add.
264 (c-family/c-ubsan.o): Add.
265 (builtins.o): Add ubsan.h dependency.
266 * ubsan.h: New file.
267 * ubsan.c: New file.
268 * common.opt: Add -fsanitize=undefined option.
269 (flag_sanitize): Add variable.
270 (fsanitize=): Add option. Add Driver.
271 (fsanitize=thread): Remove option.
272 (fsanitize=address): Likewise.
273 (static-libubsan): New option.
274 * doc/invoke.texi: Document the new flag and -static-libubsan.
275 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
276 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
277 * builtin-attrs.def (ATTR_COLD): Define.
278 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
279 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
280 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
281 * flag-types.h (sanitize_code): New enum.
282 * opts.c (common_handle_option): Parse command line arguments
283 of -fsanitize=. Add -fsanitize=unreachable option.
284 * varasm.c (get_variable_section): Adjust.
285 (assemble_noswitch_variable): Likewise.
286 (assemble_variable): Likewise.
287 (output_constant_def_contents): Likewise.
288 (categorize_decl_for_section): Likewise.
289 (place_block_symbol): Likewise.
290 (output_object_block): Likewise.
291 * builtins.def: Likewise.
292 * toplev.c (compile_file): Likewise.
293 (process_options): Likewise.
294 * cppbuiltin.c: Likewise.
295 * tsan.c (tsan_pass): Likewise.
296 (tsan_gate): Likewise.
297 (tsan_gate_O0): Likewise.
298 * cfgexpand.c (partition_stack_vars): Likewise.
299 (expand_stack_vars): Likewise.
300 (defer_stack_allocation): Likewise.
301 (expand_used_vars): Likewise.
302 * cfgcleanup.c (old_insns_match_p): Likewise.
303 * asan.c (asan_finish_file): Likewise.
304 (asan_instrument): Likewise.
305 (gate_asan): Likewise.
306 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
307 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
308 (asan_global_struct): Use pointer_sized_int_node instead
309 calling build_nonstandard_integer_type.
310 (initialize_sanitizer_builtins): Likewise.
311 (asan_finish_file): Likewise.
312 * gcc.c: Document %{%:function(args):X}.
313 (static_spec_functions): Add sanitize.
314 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
315 store funcval != NULL there.
316 (do_spec_1): Adjust handle_spec_function caller.
317 (handle_braces): Allow %:function(args) as condition.
318 (sanitize_spec_function): New function.
319 (ADD_STATIC_LIBUBSAN_LIBS): Define.
320 (LIBUBSAN_SPEC): Likewise.
321 (LIBUBSAN_EARLY_SPEC): Likewise.
322 (SANITIZER_SPEC): Handle libubsan.
323 (SANITIZER_EARLY_SPEC): Likewise.
324 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
325 instead of fsanitize=address.
326 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
327 instead of fsanitize=address*.
328 * builtins.c: Include ubsan.h.
329 (fold_builtin_0): Instrument __builtin_unreachable.
330 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
331 instead of flag_asan.
332 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
333 (pointer_sized_int_node): Define.
334 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
335
336 2013-08-30 Mike Stump <mikestump@comcast.net>
337
338 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
339 with RE patterns.
340
341 2013-08-29 Jan Hubicka <jh@suse.cz>
342
343 * cgraph.c (cgraph_function_body_availability): Handle weakref
344 correctly.
345 * passes.def: Remove pass_fixup_cfg.
346 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
347 track when we need to remove functions.
348 (gate_ipa_inline): Execute inlining always; add comment why.
349 (pass_data_ipa_inline): Remove TODO_remove_functions.
350 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
351 do not produce summaries.
352 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
353 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
354 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
355 constant pool and vtable.
356
357 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
358
359 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
360 New arm_neon.h's internal macros to specify 64-bit constants.
361 Avoid using stdint.h's macros.
362
363 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
364
365 * recog.c (verify_changes): Verify that changes[i].old is non-zero
366 before applying REG_P.
367
368 2013-08-30 Jakub Jelinek <jakub@redhat.com>
369
370 PR tree-optimization/58277
371 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
372 after seeing too many stmts with vdef in between dombb and current
373 bb, invalidate everything.
374
375 2013-08-30 Richard Biener <rguenther@suse.de>
376
377 * fold-const.c (fold_single_bit_test): Fix overflow test.
378
379 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
380
381 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
382 and which can live in a register, always retrieve the value on entry.
383 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
384 passed by invisible reference specially.
385 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
386 (vt_add_function_parameter): Correctly deal with a parameter passed by
387 invisible reference.
388
389 2013-08-30 Jan Hubicka <jh@suse.cz>
390
391 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
392
393 2013-08-30 Richard Biener <rguenther@suse.de>
394
395 PR tree-optimization/58228
396 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
397 allow invariant loads in nested loop vectorization.
398
399 2013-08-30 Richard Biener <rguenther@suse.de>
400
401 PR tree-optimization/58223
402 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
403 (has_anti_or_output_dependence): ... this and adjust to also
404 look for output dependences.
405 (mark_nodes_having_upstream_mem_writes): Adjust.
406 (rdg_flag_uses): Likewise.
407
408 2013-08-30 Richard Biener <rguenther@suse.de>
409
410 PR tree-optimization/58010
411 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
412 assert that we have a loop-closed PHI.
413
414 2013-08-29 Jan Hubicka <jh@suse.cz>
415
416 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
417 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
418 * lto-section-in.c (lto_free_function_in_decl_state): New function.
419 (lto_free_function_in_decl_state_for_node): New function.
420
421 2013-08-29 Xinliang David Li <davidxl@google.com>
422
423 * loop-unroll.c (report_unroll_peel): Minor message change.
424 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
425 Emit alignment peeling message with default -fopt-info.
426 (vect_loop_versioning): Emit loop version info message.
427 * tree-vectorizer.c (vectorize_loops): Minor message change.
428 (execute_vect_slp): Ditto.
429
430 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
431
432 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
433 of the clone from the DECL_NAME of the original function.
434
435 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
436
437 * passes.c (register_pass): Add overload.
438 * tree-pass.h (register_pass): Forward declare it. Add comment.
439
440 2013-08-29 Jan Hubicka <jh@suse.cz>
441
442 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
443 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
444 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
445 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
446 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
447 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
448 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
449 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
450
451 2013-08-29 Teresa Johnson <tejohnson@google.com>
452
453 * dumpfile.c (dump_loc): Output column number.
454 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
455 * doc/invoke.texi: Document optall -fopt-info flag.
456 * profile.c (read_profile_edge_counts): Use new dump framework.
457 (compute_branch_probabilities): Ditto.
458 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
459 when pass not in any opt group.
460 * pass_manager.h (pass_manager::get_pass_profile): New method.
461 * value-prof.c (check_counter): Use new dump framework.
462 (check_ic_target): Ditto.
463 * coverage.c (get_coverage_counts): Ditto.
464 (coverage_init): Setup new dump framework.
465
466 2013-08-29 Richard Biener <rguenther@suse.de>
467
468 PR tree-optimization/58246
469 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
470 handle the dominance check inside a basic-block.
471
472 2013-08-29 Richard Biener <rguenther@suse.de>
473
474 PR middle-end/57287
475 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
476 of default defs that appear in abnormal PHI nodes.
477
478 2013-08-29 Richard Biener <rguenther@suse.de>
479
480 PR tree-optimization/57685
481 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
482 single-use operands to avoid exponential complexity.
483
484 2013-08-28 Dehao Chen <dehao@google.com>
485
486 * ipa-inline.c (edge_badness): Fix integer underflow.
487
488 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
489
490 * gtm-builtins.def (_ITM_free): Declare leaf.
491
492 2013-08-28 Jakub Jelinek <jakub@redhat.com>
493
494 PR target/58067
495 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
496 (*tls_local_dynamic_base_64_largepic): Likewise.
497 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
498 Remove predicate from call operand.
499 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
500 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
501
502 2013-08-28 Jeff Law <law@redhat.com>
503
504 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
505 checks for the number of predecessors and successors allowed.
506 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
507 which require copying a joiner block if there is a request which
508 is a subpath that requires no joiner block copying.
509
510 2013-08-28 Jan Hubicka <jh@suse.cz>
511
512 * lto-streamer-out.c (DFS_write_tree_body): Drop
513 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
514 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
515 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
516 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
517 Do not read DECL_ERROR_ISSUED.
518 (unpack_ts_decl_with_vis_value_fields): Do not read
519 DECL_DEFER_OUTPUT.
520 (lto_input_ts_binfo_tree_pointers): Do not read
521 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
522 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
523 write DECL_ERROR_ISSUED..
524 (pack_ts_decl_with_vis_value_fields): Do not write
525 DECL_DEFER_OUTPUT.
526 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
527 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
528 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
529 * tree.h (tree_decl_common): Update comment.
530 (DECL_ERROR_ISSUED): Remove.
531
532 2013-08-28 Jakub Jelinek <jakub@redhat.com>
533
534 PR middle-end/58257
535 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
536
537 2013-08-28 Jan Hubicka <jh@suse.cz>
538
539 * builtins.def (free): Declare leaf.
540
541 2013-08-27 David Malcolm <dmalcolm@redhat.com>
542
543 * gdbhooks.py: New.
544 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
545 * configure: Regenerate.
546
547 2013-08-27 Martin Jambor <mjambor@suse.cz>
548
549 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
550 (ipa_ancestor_jf_data): Likewise.
551 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
552 (ipa_get_jf_pass_through_type_preserved): New function.
553 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
554 (ipa_get_jf_ancestor_type_preserved): New function.
555 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
556 (ipa_get_jf_ancestor_result): Likewise.
557 (propagate_vals_accross_pass_through): Use
558 ipa_get_jf_pass_through_result to do all the value mappings.
559 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
560 type_preserved flag.
561 (ipa_set_jf_cst_copy): New function.
562 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
563 (ipa_set_jf_arith_pass_through): Likewise.
564 (ipa_set_ancestor_jf): Likewise.
565 (compute_complex_assign_jump_func): Set type_preserved instead of
566 punting.
567 (ipa_compute_jump_functions_for_edge): Likewise.
568 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
569 (update_jump_functions_after_inlining): Update type_preserved.
570 Explicitely create jump functions when combining one with pass_through.
571 (ipa_write_jump_function): Stream the type_preserved flags.
572 (ipa_read_jump_function): Likewise.
573
574 2013-08-27 Jakub Jelinek <jakub@redhat.com>
575 Aldy Hernandez <aldyh@redhat.com>
576
577 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
578 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
579 * function.h (struct function): Add has_force_vect_loops and
580 has_simduid_loops.
581 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
582 * gimple.c (gimple_build_omp_critical): Add KIND argument and
583 handle it.
584 * gimple.def: Update CLAUSES comments.
585 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
586 (gimple_build_omp_for): Add argument to prototype.
587 (gimple_omp_for_kind): New.
588 (gimple_omp_for_set_kind): New.
589 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
590 GOVD_DATA_SHARE_CLASS.
591 (enum omp_region_type): Add ORT_SIMD.
592 (gimple_add_tmp_var): Handle ORT_SIMD.
593 (gimplify_var_or_parm_decl): Same.
594 (is_gimple_stmt): Same.
595 (omp_firstprivatize_variable): Same.
596 (omp_add_variable): Only use splay_tree_insert if lookup failed.
597 (omp_notice_variable): Handle ORT_SIMD.
598 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
599 (omp_check_private): Handle ORT_SIMD.
600 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
601 OMP_CLAUSE_SAFELEN.
602 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
603 Handle OMP_CLAUSE_LASTPRIVATE.
604 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
605 OMP_CLAUSE_SAFELEN.
606 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
607 (gimplify_expr): Handle OMP_SIMD.
608 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
609 (expand_GOMP_SIMD_VF): New.
610 (expand_GOMP_SIMD_LAST_LANE): New.
611 * internal-fn.def (GOMP_SIMD_LANE): New.
612 (GOMP_SIMD_VF): New.
613 (GOMP_SIMD_LAST_LANE): New.
614 * omp-low.c: Include target.h.
615 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
616 OMP_CLAUSE_SAFELEN.
617 (check_omp_nesting_restrictions): Same.
618 (omp_max_vf): New.
619 (lower_rec_simd_input_clauses): New.
620 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
621 OMP_CLAUSE_LINEAR.
622 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
623 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
624 (expand_omp_build_assign): New.
625 (expand_omp_for_init_counts): New.
626 (expand_omp_for_init_vars): New.
627 (extract_omp_for_update_vars): New.
628 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
629 and rewrite accordingly.
630 (expand_omp_simd): New.
631 (expand_omp_for): Use expand_omp_simd.
632 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
633 (lower_omp_for): Do not lower the body.
634 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
635 in their own loops.
636 * tree-flow.h (find_omp_clause): Remove prototype.
637 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
638 forcing vectorization of the loop, or if flag_tree_vectorize.
639 (gate_tree_if_conversion): Similarly.
640 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
641 gimple_build_omp_for.
642 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
643 * tree-parloops (create_parallel_loop): Pass kind argument to
644 gimple_build_omp_for.
645 * tree-pretty-print.c (dump_omp_clause): Add cases for
646 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
647 OMP_CLAUSE__SIMDUID_.
648 (dump_generic_node): Handle OMP_SIMD.
649 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
650 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
651 unroll OMP_SIMD loops here.
652 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
653 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
654 loop->safelen.
655 (vect_analyze_data_refs): Handle simd loops.
656 * tree-vect-loop.c (vectorizable_live_operation): Handle
657 IFN_GOMP_SIMD*.
658 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
659 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
660 (vectorizable_load): Same.
661 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
662 (struct simduid_to_vf): New.
663 (simduid_to_vf::hash): New.
664 (simduid_to-vf::equal): New.
665 (struct simd_array_to_simduid): New.
666 (simd_array_to_simduid::hash): New.
667 (simd_array_to_simduid::equal): New.
668 (adjust_simduid_builtins): New.
669 (struct note_simd_array_uses_struct): New.
670 (note_simd_array_uses_cb): New.
671 (note_simd_array_uses): New.
672 (vectorize_loops): Handle simd hints and adjust simd builtins
673 accordingly.
674 * tree-vectorizer.h (struct _stmt_vec_info): Add
675 simd_lane_access_p field.
676 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
677 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
678 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
679 (omp_clause_code_name): Same.
680 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
681 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
682 * tree.def (OMP_SIMD): New entry.
683 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
684 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
685 (OMP_CLAUSE_DECL): Adjust range for new clauses.
686 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
687 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
688 (OMP_CLAUSE_LINEAR_STEP): New.
689 (OMP_CLAUSE_SAFELEN_EXPR): New.
690 (OMP_CLAUSE__SIMDUID__DECL): New.
691 (find_omp_clause): New prototype.
692
693 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
694
695 * config/i386/driver-i386.c (host_detect_local_cpu): Update
696 Haswell processor detection.
697
698 2013-08-27 Christian Widmer <shadow@umbrox.de>
699
700 PR target/57927
701 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
702 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
703 AVX2 capable processors.
704
705 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
706
707 * config/aarch64/arm_neon.h: Replace all inline asm implementations
708 of vget_low_* with implementations in terms of other intrinsics.
709
710 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
711
712 PR middle-end/57219
713 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
714 values to -1, 0 and 1.
715
716 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
717
718 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
719 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
720 (<optab><mode>3_insn): Remove.
721 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
722 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
723 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
724 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
725 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
726 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
727 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
728 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
729 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
730 (ror<mode>3_insn): Likewise.
731 * config/aarch64/predicates.md (aarch64_simd_register): New.
732
733 2013-08-27 Richard Biener <rguenther@suse.de>
734
735 PR tree-optimization/57521
736 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
737 one edge is non-critical.
738 (find_phi_replacement_condition): Make sure to use a non-critical
739 edge. Cleanup and remove old bug workarounds.
740 (bb_postdominates_preds): Remove.
741 (if_convertible_loop_p_1): Do not compute post-dominators.
742 (combine_blocks): Do not free post-dominators.
743 (main_tree_if_conversion): Likewise.
744 (pass_data_if_conversion): Add TODO_verify_ssa.
745
746 2013-08-27 DJ Delorie <dj@redhat.com>
747
748 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
749
750 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
751
752 * function.c (assign_parm_find_data_types): Set passed_mode and
753 nominal_mode to the TYPE_MODE of nominal_type for the built
754 pointer type in case of the struct-pass-by-reference.
755
756 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
757
758 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
759 (UINT16_TYPE): Change default to "unsigned int".
760
761 * config/avr/avr.opt (mfract-convert-truncate): New option.
762 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
763 is set, round negative fractional integers according to n1169
764 when converting to integer types.
765
766 2013-08-26 Jan Hubicka <jh@suse.cz>
767
768 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
769 methods can not be called indirectly when their address is not taken.
770
771 2013-08-26 Jan Hubicka <jh@suse.cz>
772
773 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
774 ctor_for_folding.
775
776 2013-08-26 Jan Hubicka <jh@suse.cz>
777
778 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
779 can be unshared.
780
781 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
782
783 * reload.c (find_valid_class): Allow classes that do not include
784 FIRST_PSEUDO_REGISTER - 1.
785
786 2013-08-26 Jan Hubicka <jh@suse.cz>
787
788 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
789 fix edge count/frequency when speculation failed; fix type check
790 for the direct call.
791
792 2013-08-26 Jan Hubicka <jh@suse.cz>
793
794 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
795
796 2013-08-26 Jan Hubicka <jh@suse.cz>
797
798 * ipa-inline-transform.c (inline_transform): Be ready for basic block
799 to be changed by edge redirection.
800
801 2013-08-26 Jan Hubicka <jh@suse.cz>
802
803 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
804 formating; add sanity check.
805 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
806 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
807 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
808 (resolve_noninline_speculation): Update callee keys, too.
809
810 2013-08-26 Jan Hubicka <jh@suse.cz>
811
812 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
813 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
814
815 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
816
817 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
818 into proper place.
819
820 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
821
822 * config/i386/i386.c (ix86_debug_options): Remove prototype.
823 (x86_64_elf_select_section): Ditto.
824 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
825 arguments.
826 (ix86_pass_by_reference): Ditto.
827 (output_set_got): Ditto.
828 (ix86_unary_operator_ok): Ditto.
829 (ix86_expand_builtin): Ditto.
830
831 2013-08-23 Jan Hubicka <jh@suse.cz>
832
833 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
834
835 2013-08-23 Jan Hubicka <jh@suse.cz>
836
837 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
838 (tree_decl_with_vis): Add FINAL field.
839
840 2013-08-23 Jeff Law <law@redhat.com>
841
842 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
843 the debugging dump when the expression is fully redundant.
844
845 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
846
847 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
848 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
849 * pretty-print.c (pp_formatted_text_data): Likewise.
850 (pp_write_text_to_stream): Likewise.
851 (pp_write_text_as_dot_label_to_stream): Likewise.
852 (pp_append_r): Likewise.
853 (pp_format): Likewise.
854 (pp_flush): Likewise.
855 (pp_clear_output_area): Likewise.
856 (pp_append_text): Likewise.
857 (pp_formatted_text): Likewise.
858 (pp_remaining_character_count_for_line): Likewise.
859 (pp_newline): Likewise.
860 (pp_character): Likewise.
861 (output_buffer::~output_buffer): Define.
862 (pretty_printer::~pretty_printer): Destruct output buffer.
863 * pretty-print.h (output_buffer::~output_buffer): Declare.
864 (pretty_printer::~pretty_printer): Declare virtual.
865
866 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
867
868 PR other/57324
869 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
870 HOST_WIDE_INT_M1U): New macros.
871 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
872 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
873 unsigned -1 for lshift.
874 * cse.c (cse_insn): Likewise.
875 * double-int.c (rshift_double, lshift_double): Likewise.
876 * builtins.c (fold_builtin_bitop): Likewise.
877 * combine.c (force_to_mode): Likewise.
878 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
879 * simplify-rtx.c (simplify_const_unary_operation,
880 simplify_const_binary_operation): Likewise.
881 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
882 check_va_list_escapes): Likewise.
883 * rtlanal.c (nonzero_bits1): Likewise.
884 * expmed.c (expand_smod_pow2): Likewise.
885 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
886
887 2013-08-23 Jan Hubicka <jh@suse.cz>
888
889 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
890 as having address taken.
891
892 2013-08-23 Jan Hubicka <jh@suse.cz>
893
894 * ipa-utils.h (method_class_type): Declare.
895 * ipa-devirt.c (method_class_type): Export.
896
897 * cgraphunit.c (analyze_functions): Do basic devirtualization;
898 do not walk base classes of anonymous types.
899
900 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
901
902 PR rtl-optimization/58220
903 PR regression/58221
904 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
905 handle SEQUENCE insns properly.
906
907 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
908
909 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
910 definition.
911 (pp_newline_and_indent): Likewise.
912 (pp_separate_with): Likewise.
913 * pretty-print.c (pp_newline_and_flush): Define.
914 (pp_newline_and_indent): Likewise.
915 (pp_separate_with): Likewise.
916
917 2013-08-23 Jakub Jelinek <jakub@redhat.com>
918
919 PR target/58218
920 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
921 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
922
923 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
924
925 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
926 * gcc/config/i386/i386.md (*movti_internal): Use
927 predicate to determine if EVEX is needed.
928 (*movsi_internal): Ditto.
929 (*movdf_internal): Ditto.
930 (*movsf_internal): Ditto.
931 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
932
933 2013-08-23 Jakub Jelinek <jakub@redhat.com>
934
935 PR tree-optimization/58209
936 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
937 (find_tail_calls): Give up for pointer result types if m is non-NULL.
938 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
939 emit POINTER_PLUS_EXPR.
940 (create_tailcall_accumulator): For pointer result type accumulate in
941 sizetype type.
942
943 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
944
945 * configure.ac: Add backslashes missing from the last change.
946 * configure: Regenerate.
947
948 2013-08-22 Jan Hubicka <jh@susue.cz>
949
950 * ipa.c (function_and_variable_visibility): First remember function
951 was global and then make it local.
952
953 2013-08-22 Julian Brown <julian@codesourcery.com>
954
955 * configure.ac: Add aarch64 to list of arches which use "nop" in
956 debug_line test.
957 * configure: Regenerate.
958
959 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
960
961 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
962 gnu_libc_has_function.
963 * config/s390/tpf.h: Likewise.
964
965 2013-08-22 Jan Hubicka <jh@susue.cz>
966
967 * timevar.c (validate_phases): Add cast.
968
969 2013-08-22 Jan Hubicka <jh@susue.cz>
970
971 * timevar.c (validate_phases): Use size_t for memory.
972 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
973
974 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
975
976 * pretty-print.h (output_buffer::output_buffer): Declare.
977 (pretty_printer::pretty_printer): Likewise.
978 (pp_construct): Remove.
979 * pretty-print.c (output_buffer::output_buffer): Define.
980 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
981 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
982 (print_gimple_expr): Likewise.
983 (print_gimple_seq): Likewise.
984 (gimple_dump_bb): Likewise.
985 * sched-vis.c (dump_value_slim): Likewise.
986 (dump_insn_slim): Likewise.
987 (dump_rtl_slim): Likewise.
988 (str_pattern_slim): Likewise.
989 * tree-mudflap.c (mf_varname_tree): Likewise.
990 * graph.c (print_graph_cfg): Likewise.
991 (start_graph_dump): Likewise.
992 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
993 placement-new.
994 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
995 pretty printer initialization.
996 * coretypes.h (diagnostic_context): Remove superflous type alias
997 declaration.
998 (pretty_printer): Likewise. Declare directly as a class.
999 (pretty_print_info): Remove declaration as class.
1000 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
1001 and pp_clear_output_area.
1002 (asan_add_global): Likewise.
1003
1004 2013-08-22 Jan Hubicka <jh@suse.cz>
1005
1006 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
1007 * ipa-utils.h (update_type_inheritance_graph): Declare.
1008 (possible_polymorphic_call_target_p): Declare.
1009 (possible_polymorphic_call_target_p): New.
1010 * ipa-devirt.c: Update toplevel comments.
1011 (cached_polymorphic_call_targets): Move up.
1012 (odr_type_d): Move ID down.
1013 (polymorphic_type_binfo_p): Update comment.
1014 (odr_hasher::remove): Likewise;
1015 (get_odr_type): Set anonymous_namespace.
1016 (dump_odr_type): Dump it.
1017 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
1018 (maybe_record_node): Record node in cached_polymorphic_call_targets.
1019 (record_binfo): Add comment.
1020 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
1021 built.
1022 (devirt_node_removal_hook): Do not iCE when cache is freed.
1023 (possible_polymorphic_call_target_p): New predicate.
1024 (update_type_inheritance_graph): New function.
1025
1026 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1027 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1028 Sergey Lega <sergey.s.lega@intel.com>
1029 Anna Tikhonova <anna.tikhonova@intel.com>
1030 Ilya Tocar <ilya.tocar@intel.com>
1031 Andrey Turetskiy <andrey.turetskiy@intel.com>
1032 Ilya Verbin <ilya.verbin@intel.com>
1033 Kirill Yukhin <kirill.yukhin@intel.com>
1034 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1035
1036 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1037 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1038 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1039 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1040 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1041 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1042 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1043 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1044 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1045 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1046 OPT_mavx512pf, OPT_mavx512er cases.
1047 * config/i386/constraints.md (v): New constraint.
1048 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1049 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1050 (bit_AVX512CD): New.
1051 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1052 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1053 * config/i386/i386-c.c (ix86_target_macros_internal):
1054 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1055 __AVX512PF__.
1056 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1057 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1058 * config/i386/i386.c (regclass_map, dbx_register_map)
1059 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1060 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1061 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1062 -mavx512pf options.
1063 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1064 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1065 -mavx512pf options. Fix formatting.
1066 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1067 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1068 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1069 -mavx512cd, -mavx512pf options.
1070 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1071 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1072 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1073 ALL_SSE_REGS.
1074 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1075 (ix86_set_reg_reg_cost): Ditto.
1076 (x86_order_regs_for_local_alloc): Ditto.
1077 (MAX_VECT_LEN): Extend to 64-byte.
1078 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1079 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1080 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1081 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1082 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1083 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1084 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1085 (SSE_REG_MODE_P): Support new modes.
1086 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1087 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1088 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1089 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1090 (REG_CLASS_CONTENTS): Add new registers.
1091 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1092 (EXT_REX_SSE_REGNO_P): New.
1093 (HI_REGISTER_NAMES): Add new registers.
1094 * config/i386/i386.md: Define constants for new registers.
1095 (mode): Add new 512-bit modes.
1096 (prefix): Support evex prefix.
1097 (isa): Support avx512f, noavx512f, fma_avx512f.
1098 (ssemodesuffix): Add new 512-bit modes.
1099 (movxi): New.
1100 (*movxi_internal_avx512f): Ditto.
1101 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1102 Support MODE_XI.
1103 (*movsi_internal): Likewise.
1104 (*movdf_internal): Likewise.
1105 (*movsf_internal): Likewise.
1106 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1107 (<code><mode>3): Likewise.
1108 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1109 New.
1110 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1111 with the new constraint "v".
1112 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1113 modes.
1114 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1115 with the new constraint "v".
1116 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1117 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1118 (*<plusminus_insn><mode>3): Likewise.
1119 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1120 (*mul<mode>3): Likewise.
1121 (<sse>_vmmul<mode>3): Likewise.
1122 (<sse>_div<mode>3): Likewise.
1123 (<sse>_vmdiv<mode>3): Likewise.
1124 (<sse>_sqrt<mode>2): Likewise.
1125 (<sse>_vmsqrt<mode>2): Likewise.
1126 (*<code><mode>3_finite): Likewise.
1127 (*<code><mode>3) <smaxmin>: Likewise.
1128 (<sse>_vm<code><mode>3): Likewise.
1129 (*<code><mode>3) <any_logic>: Likewise.
1130 (*fma_fmadd_<mode>): Likewise.
1131 (*fma_fmsub_<mode>): Likewise.
1132 (*fma_fnmadd_<mode>): Likewise.
1133 (*fma_fnmsub_<mode>): Likewise.
1134 (*fma_fmaddsub_<mode>): Likewise.
1135 (*fma_fmsubadd_<mode>): Likewise.
1136 (*fmai_fmadd_<mode>): Likewise.
1137 (*fmai_fmsub_<mode>): Likewise.
1138 (*fmai_fnmadd_<mode>): Likewise.
1139 (*fmai_fnmsub_<mode>): Likewise.
1140 (sse_cvtsi2ss): Likewise.
1141 (sse_cvtsi2ssq): Likewise.
1142 (sse_cvtss2si): Likewise.
1143 (sse_cvtss2si_2): Likewise.
1144 (sse_cvtss2siq): Likewise.
1145 (sse_cvtss2siq_2): Likewise.
1146 (sse_cvttss2si): Likewise.
1147 (sse_cvtss2siq_2): Likewise.
1148 (float<sseintvecmodelower><mode>2): Likewise.
1149 (sse2_cvtsd2si_2): Likewise.
1150 (sse2_cvtsd2siq_2): Likewise.
1151 (*<plusminus_insn><mode>3): Likewise.
1152 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1153 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1154 (ashr<mode>3): Likewise.
1155 (<shift_insn><mode>3): Likewise.
1156 (avx2_<code><mode>3): Likewise.
1157 (*avx2_<code><mode>3): Likewise.
1158 (*andnot<mode>3): Likewise.
1159 (*<code><mode>3) <any_logic>: Likewise.
1160 (abs<mode>2): Likewise.
1161 (avx2_permvar<mode>): Likewise.
1162 (avx2_perm<mode>_1): Likewise.
1163 (*avx_vpermilp<mode>): Likewise.
1164 (avx_vpermilvar<mode>3): Likewise.
1165 (avx2_ashrv<mode>): Likewise.
1166 (avx2_<shift_insn>v<mode>): Likewise.
1167 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1168 -mavx512cd.
1169 * doc/rtl.texi: Document XImode.
1170
1171 2013-08-21 Jeff Law <law@redhat.com>
1172
1173 * tree-flow.h (register_jump_thread): Pass vector of edges
1174 instead of each important edge.
1175 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1176 thread path into a vector and pass that to register_jump_thread.
1177 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1178 passed in edge vector to the current 3-edge form.
1179
1180 Revert:
1181 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1182
1183 * dce.c (fini_dce): Call df_analyze again just in case
1184 delete_unmarked_insns removed anything.
1185
1186 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1187
1188 * reload.h (struct reg_equivs): Rename to ..
1189 (struct reg_equivs_s): .. this.
1190
1191 2013-08-20 Martin Liska <marxin.liska@gmail.com>
1192
1193 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
1194
1195 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1196
1197 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
1198
1199 2013-08-21 Jeff Law <law@redhat.com>
1200
1201 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
1202 simplify assignments too. If the RHS collapses to a singleton
1203 range, then return the value for the range.
1204
1205 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
1206
1207 * config/i386/sse.md (V16): Rename to...
1208 (VMOVE): this.
1209 (mov<mode>): Update iterator name.
1210 (*mov<mode>_internal): Ditto.
1211 (push<mode>1): Ditto.
1212 (movmisalign<mode>): Ditto.
1213
1214 2013-08-20 Jan Hubicka <jh@suse.cz>
1215
1216 PR bootstrap/58186
1217 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
1218 entry for direct edges.
1219 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
1220
1221 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1222
1223 Revert my last two changes, r201865 and r201864:
1224
1225 Revert r201865:
1226 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1227
1228 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1229 instances can own GC refs.
1230
1231 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1232 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1233 (gcc::context::gt_pch_nx): Likewise.
1234 (gcc::context::gt_pch_nx): Likewise.
1235 * ggc.h (gt_ggc_mx <T>): New.
1236 (gt_pch_nx_with_op <T>): New.
1237 (gt_pch_nx <T>): New.
1238 * passes.c (opt_pass::gt_ggc_mx): New.
1239 (opt_pass::gt_pch_nx): New.
1240 (opt_pass::gt_pch_nx_with_op): New.
1241 (pass_manager::gt_ggc_mx): New.
1242 (pass_manager::gt_pch_nx): New.
1243 (pass_manager::gt_pch_nx_with_op): New.
1244 (pass_manager::operator new): Use
1245 ggc_internal_cleared_alloc_stat rather than xcalloc.
1246 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1247 (pass_manager::gt_ggc_mx): New.
1248 (pass_manager::gt_pch_nx): New.
1249 (pass_manager::gt_pch_nx_with_op): New.
1250 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1251 (opt_pass::operator new): New.
1252 (opt_pass::gt_ggc_mx): New.
1253 (opt_pass::gt_pch_nx): New.
1254 (opt_pass::gt_pch_nx_with_op): New.
1255
1256 Revert r201864:
1257 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1258
1259 * Makefile.in (GTFILES): Add context.h.
1260 * context.c (gcc::context::operator new): New.
1261 (gcc::context::gt_ggc_mx): New.
1262 (gcc::context::gt_pch_nx): New.
1263 (gcc::context::gt_pch_nx): New.
1264 * context.h (gcc::context): Add GTY((user)) marking.
1265 (gcc::context::operator new): New.
1266 (gcc::context::gt_ggc_mx): New.
1267 (gcc::context::gt_pch_nx): New.
1268 (gcc::context::gt_pch_nx): New.
1269 (g): Add GTY marking.
1270 (gt_ggc_mx (gcc::context *)): New.
1271 (gt_pch_nx (gcc::context *)): New.
1272 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1273 void *cookie)): New.
1274 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1275
1276 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1277
1278 * dce.c (fini_dce): Call df_analyze again just in case
1279 delete_unmarked_insns removed anything.
1280
1281 2013-08-20 Teresa Johnson <tejohnson@google.com>
1282
1283 PR rtl-optimizations/57451
1284 * final.c (reemit_insn_block_notes): Prevent lexical blocks
1285 from crossing split section boundaries.
1286
1287 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1288
1289 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
1290 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
1291 with MULTLIB_DEFAULTS.
1292
1293 2013-08-20 Nick Clifton <nickc@redhat.com>
1294
1295 * target.def (narrow_volatile_bitfield): Note that the default
1296 value is false, not !TARGET_STRICT_ALIGN.
1297 * doc/tm.texi: Regenerate.
1298
1299 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
1300
1301 Fix LIB_SPEC for systems without libpthread.
1302
1303 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
1304 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
1305 for Android.
1306 * config/i386/linux-common.h: Likewise.
1307 * config/mips/linux-common.h: Likewise.
1308
1309 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1310
1311 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
1312 checks.
1313
1314 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1315
1316 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1317 instances can own GC refs.
1318
1319 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1320 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1321 (gcc::context::gt_pch_nx): Likewise.
1322 (gcc::context::gt_pch_nx): Likewise.
1323 * ggc.h (gt_ggc_mx <T>): New.
1324 (gt_pch_nx_with_op <T>): New.
1325 (gt_pch_nx <T>): New.
1326 * passes.c (opt_pass::gt_ggc_mx): New.
1327 (opt_pass::gt_pch_nx): New.
1328 (opt_pass::gt_pch_nx_with_op): New.
1329 (pass_manager::gt_ggc_mx): New.
1330 (pass_manager::gt_pch_nx): New.
1331 (pass_manager::gt_pch_nx_with_op): New.
1332 (pass_manager::operator new): Use
1333 ggc_internal_cleared_alloc_stat rather than xcalloc.
1334 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1335 (pass_manager::gt_ggc_mx): New.
1336 (pass_manager::gt_pch_nx): New.
1337 (pass_manager::gt_pch_nx_with_op): New.
1338 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1339 (opt_pass::operator new): New.
1340 (opt_pass::gt_ggc_mx): New.
1341 (opt_pass::gt_pch_nx): New.
1342 (opt_pass::gt_pch_nx_with_op): New.
1343
1344 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1345
1346 * Makefile.in (GTFILES): Add context.h.
1347 * context.c (gcc::context::operator new): New.
1348 (gcc::context::gt_ggc_mx): New.
1349 (gcc::context::gt_pch_nx): New.
1350 (gcc::context::gt_pch_nx): New.
1351 * context.h (gcc::context): Add GTY((user)) marking.
1352 (gcc::context::operator new): New.
1353 (gcc::context::gt_ggc_mx): New.
1354 (gcc::context::gt_pch_nx): New.
1355 (gcc::context::gt_pch_nx): New.
1356 (g): Add GTY marking.
1357 (gt_ggc_mx (gcc::context *)): New.
1358 (gt_pch_nx (gcc::context *)): New.
1359 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1360 void *cookie)): New.
1361 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1362
1363 2013-08-20 Alan Modra <amodra@gmail.com>
1364
1365 PR target/57865
1366 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
1367 (rs6000_emit_epilogue): Likewise.
1368
1369 2013-08-19 Dehao Chen <dehao@google.com>
1370
1371 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
1372
1373 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
1374 Jakub Jelinek <jakub@redhat.com>
1375
1376 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
1377 (BUILT_IN_FABSD64): Likewise.
1378 (BUILT_IN_FABSD128): Likewise.
1379 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
1380 (fold_builtin_1): Likewise.
1381 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
1382 destination and source operands.
1383 (*abstd2_fpr): Likewise.
1384 (*nabstd2_fpr): Likewise.
1385
1386 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
1387
1388 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
1389 JUMP_P and INSN_P.
1390
1391 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
1392
1393 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
1394 incomplete.
1395
1396 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
1397
1398 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
1399 * builtins.c (default_libc_has_function): New.
1400 (gnu_libc_has_function): Ditto.
1401 (no_c99_libc_has_function): Ditto.
1402 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
1403 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
1404 (fold_builtin_sincos): Likewise.
1405 (fold_builtin_cexp): Likewise.
1406 * builtins.def (DEF_C94_BUILTIN): Likewise.
1407 (DEF_C99_BUILTIN): Likewise.
1408 (DEF_C99_C90RES_BUILTIN): Likewise.
1409 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
1410 definitions to using this define.
1411 * config/darwin-protos.h (darwin_libc_has_function): New.
1412 * config/darwin.c (darwin_libc_has_function): Ditto.
1413 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
1414 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
1415 * config/darwin.h: Ditto.
1416 * config/elfos.h: Ditto.
1417 * config/freebsd.h: Ditto.
1418 * config/i386/cygming.h: Ditto.
1419 * config/i386/djgpp.h: Ditto.
1420 * config/i386/i386-interix.h: Ditto.
1421 * config/microblaze/microblaze.h: Ditto.
1422 * config/mmix/mmix.h: Ditto.
1423 * config/gnu-user.h: Ditto.
1424 * config/ia64/hpux.h: Ditto.
1425 * config/pa/pa-hpux.h: Ditto.
1426 * config/pdp11/pdp11.h: Ditto.
1427 * config/picochip/picochip.h: Ditto.
1428 * config/linux.h: Ditto.
1429 * config/netbsd.h: Ditto.
1430 * config/openbsd.h: Ditto.
1431 * config/rs6000/aix43.h: Ditto.
1432 * config/rs6000/aix51.h: Ditto.
1433 * config/rs6000/aix52.h: Ditto.
1434 * config/rs6000/aix53.h: Ditto.
1435 * config/rs6000/aix61.h: Ditto.
1436 * config/rs6000/darwin.h: Ditto.
1437 * config/rs6000/linux.h: Ditto.
1438 * config/rs6000/linux64.h: Ditto.
1439 * config/s390/tpf.h: Ditto.
1440 * config/sol2-10.h: Ditto.
1441 * config/sol2.h: Ditto.
1442 * config/vms/vms.h: Ditto.
1443 * config/vxworks.h: Ditto.
1444 * config/linux-android.c (linux_android_libc_has_function):
1445 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
1446 * config/linux-protos.h (linux_android_libc_has_function):
1447 New declaration.
1448 * config/i386/i386.c (ix86_libc_has_function): New.
1449 * config/i386/i386-protos.h
1450 (ix86_libc_has_function): New declaration.
1451 * config/i386/i386.md
1452 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
1453 ("isinf<mode>2): Likewise.
1454 * convert.c (convert_to_integer): Using new target hook
1455 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
1456 TARGET_C99_FUNCTIONS.
1457 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
1458 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
1459 * coretypes.h (function_class): New enum for different
1460 classes of functions.
1461 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
1462 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
1463 (TARGET_HAS_SINCOS): Likewise.
1464 (TARGET_LIBC_HAS_FUNCTION): New.
1465 * doc/tm.texi: Regenerated.
1466 * targhooks.h (default_libc_has_function): New declaration.
1467 (no_c99_libc_has_function): Ditto.
1468 (gnu_libc_has_function): Ditto.
1469 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
1470 and TARGET_HAS_SINCOS.
1471
1472 2013-08-18 Jan Hubicka <jh@suse.cz>
1473
1474 * Makeifle-in (ipa-devirt.o): New.
1475 (GTFILES): Add ipa-utils.h and ipa-devirt.c
1476 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
1477 (analyze_functions): Look into possible targets of polymorphic call.
1478 * dumpfile.c (dump_files): Add type-inheritance dump.
1479 * dumpfile.h (TDI_inheritance): New.
1480 * ipa-devirt.c: New file.
1481 * ipa-utils.h (odr_type_d): Forward declare.
1482 (odr_type): New type.
1483 (build_type_inheritance_graph): Declare.
1484 (possible_polymorphic_call_targets): Declare and introduce inline
1485 variant when only edge is pased.
1486 (dump_possible_polymorphic_call_targets): Likewise.
1487 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
1488 * tree.c (type_in_anonymous_namespace_p): Break out from ...
1489 (types_same_for_odr): ... here.
1490 * tree.h (type_in_anonymous_namespace_p): Declare.
1491
1492 2013-08-18 Jakub Jelinek <jakub@redhat.com>
1493
1494 PR tree-optimization/58006
1495 * tree-parloops.c (take_address_of): Don't ICE if get_name
1496 returns NULL.
1497 (eliminate_local_variables_stmt): Remove clobber stmts.
1498
1499 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
1500
1501 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
1502 error message is issued for an alias to undefined symbol.
1503
1504 2013-08-18 Jan Hubicka <jh@suse.cz>
1505
1506 * cgraph.c (cgraph_create_indirect_edge): Discover
1507 polymorphic calls and record basic info into indirect_info.
1508 * gimple-fold.c (gimple_fold_call): When doing BINFO based
1509 devirtualization, ignore objc function calls.
1510 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
1511 call with no parm index info.
1512 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
1513 * tree.c (virtual_method_call_p): New function.
1514 * tree.h (virtual_method_call_p): Declare.
1515
1516 2013-08-16 Jan Hubicka <jh@suse.cz>
1517
1518 PR middle-end/58179
1519 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
1520
1521 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
1522
1523 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
1524 attribute.
1525
1526 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1527
1528 * gengtype.c (type_for_name): Add special-case support for
1529 locating types within the "gcc::" namespace.
1530 (open_base_files): Emit a "using namespace gcc" directive.
1531
1532 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1533
1534 PR target/58160
1535 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
1536 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
1537
1538 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
1539 array instead of each individual operand as a separate argument.
1540 (emit_fusion_gpr_load): Likewise.
1541 (expand_fusion_gpr_load): Add new function declaration.
1542
1543 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
1544 signature to have the operands passed as an array, instead of as
1545 separate arguments. Allow ZERO_EXTEND to be in the memory
1546 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
1547 depend on the register live/dead flags when peepholes are run.
1548 (expand_fusion_gpr_load): New function to be called from the
1549 peephole2 pass, to change the register that addis sets to be the
1550 target register.
1551 (emit_fusion_gpr_load): Change the calling signature to have the
1552 operands passed as an array, instead of as separate arguments.
1553 Allow ZERO_EXTEND to be in the memory address, and also
1554 SIGN_EXTEND if -mpower8-fusion-sign.
1555
1556 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
1557 unspec enumeration.
1558 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
1559 adjust the register addis loads up in the peephole2 pass. Do not
1560 depend on the register live/dead state when the peephole pass is done.
1561
1562 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1563
1564 * gengtype.c (create_user_defined_type): Ensure that the kind
1565 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
1566 declaration is seen before the GTY((user)) marking.
1567
1568 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
1569
1570 PR target/58105
1571 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
1572
1573 2013-08-16 Jan Hubicka <jh@suse.cz>
1574
1575 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
1576 arugment expected_type.
1577 (gimple_fold_call): Use it.
1578 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
1579 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
1580 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
1581 (try_make_edge_direct_virtual_call): Likewise.
1582 * tree.c (obj_type_ref_class): New.
1583 * tree.h (obj_type_ref_class): Use it.
1584
1585 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
1586
1587 * sched-vis.c (rtl_slim_pp_initialized): Remove.
1588 (rtl_slim_pp): Likewise.
1589 (init_rtl_slim_pretty_print): Likewise.
1590 (dump_value_slim): Don't call it. Use local pretty printer.
1591 (dump_insn_slim): Likewise.
1592 (dump_rtl_slim): Likewise.
1593 (str_pattern_slim): Likewise.
1594 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
1595 Simplify.
1596
1597 2013-08-16 Jakub Jelinek <jakub@redhat.com>
1598
1599 PR tree-optimization/58164
1600 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
1601 walk gimple_goto_dest of GIMPLE_GOTO.
1602
1603 PR tree-optimization/58165
1604 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
1605 bi_call must be the last stmt in a bb, don't split_block, instead
1606 use fallthru edge from it and give up if there is none.
1607 Release conds vector when returning early.
1608
1609 2013-08-14 Xinliang David Li <davidxl@google.com>
1610
1611 * config/i386/i386.c (ix86_option_override_internal):
1612 Remove unused variable and field.
1613
1614 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1615
1616 PR target/57949
1617 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
1618 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
1619 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
1620 and Linux, correct BLKmode alignment when 128-bit alignment is
1621 required and compatibility flag is not set.
1622 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
1623 for zero-size arguments when compatibility flag is not set.
1624
1625 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1626
1627 PR tree-optimization/58145
1628 * tree-sra.c (build_ref_for_offset): If prev_base has
1629 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
1630
1631 2013-08-14 Xinliang David Li <davidxl@google.com>
1632
1633 * config/i386/i386.c (ix86_option_override_internal):
1634 Fix uninitialized variable error.
1635
1636 2013-08-14 Xinliang David Li <davidxl@google.com>
1637
1638 * config/i386/i386.opt: Define two new options.
1639 * config/i386/x86-tune.def: Add arch selector field in macros.
1640 * config/i386/i386.h: Adjust macro definition.
1641 * config/i386/i386.c (ix86_option_override_internal):
1642 Refactor the code.
1643 (parse_mtune_ctrl_str): New function.
1644 (set_ix86_tune_features): New function.
1645 (ix86_function_specific_restore): Call the new helper function.
1646
1647 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
1648
1649 PR rtl-optimization/57662
1650 * sel-sched.c (code_motion_process_successors): When the current insn
1651 is removed after the recursive traversal, break from the loop.
1652 Add comments and debug printouts.
1653
1654 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1655 Alexandre Oliva <aoliva@redhat.com>
1656
1657 PR target/58067
1658 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
1659 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
1660 there also UNSPEC_PLTOFF.
1661
1662 2013-08-14 Marek Polacek <polacek@redhat.com>
1663
1664 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
1665 NUM_CONDITIONS bit positions.
1666
1667 2013-08-13 Cary Coutant <ccoutant@google.com>
1668
1669 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
1670 (attr_checksum): Hash vector contents instead of pointer.
1671 (attr_checksum_ordered): Likewise.
1672
1673 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
1674
1675 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
1676 when Pmode != word_mode. Add length_address attribute.
1677 (sse3_monitor_<mode>): Merge from sse3_monitor and
1678 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
1679 Pmode != word_mode. Update insn length attribute.
1680 * config/i386/i386.c (ix86_option_override_internal): Update
1681 ix86_gen_monitor selection for merged sse3_monitor insn.
1682
1683 2013-08-13 Julian Brown <julian@codesourcery.com>
1684
1685 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
1686 perform invalid legitimization on greater-than-word-size modes for
1687 TARGET_E500_DOUBLE.
1688
1689 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
1690
1691 * ira.c (setup_class_translate_array): Use aclass instead of cl
1692 for classes not fully covered by allocno classes.
1693
1694 2013-08-13 Jakub Jelinek <jakub@redhat.com>
1695
1696 PR tree-optimization/57661
1697 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
1698 * tree-inline.c (tree_function_versioning): Initialize it.
1699 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
1700 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
1701 that is not being copied.
1702
1703 PR sanitizer/56417
1704 * asan.c (instrument_strlen_call): Fix typo in comment.
1705 Use char * type even for the lhs of POINTER_PLUS_EXPR.
1706
1707 2013-08-13 Steve Ellcey <sellcey@mips.com>
1708
1709 * config/mips/mips.md (prefetch): Use lw instead of ld on
1710 loongson in 32bit mode.
1711
1712 2013-08-13 Nick Clifton <nickc@redhat.com>
1713
1714 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
1715
1716 2013-08-13 Jan Hubicka <jh@suse.cz>
1717
1718 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
1719 introduced edge; fix typo in sanity check.
1720 (cgraph_resolve_speculation): Export; improve diagnostic.
1721 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
1722 speculation at type mismatch.
1723 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
1724 (cgraph_resolve_speculation): Declare.
1725 (symtab_can_be_discarded): New function.
1726 * value-prof.c (gimple_ic_transform): Remove actual transform code.
1727 * ipa-inline-transform.c (speculation_removed): New global var.
1728 (clone_inlined_nodes): See if speculation can be removed.
1729 (inline_call): If speculations was removed, we growths may not match.
1730 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
1731 (speculation_useful_p): New function.
1732 (resolve_noninline_speculation): New function.
1733 (inline_small_functions): Resolve useless speculations.
1734 * ipa-inline.h (speculation_useful_p): Declare
1735 * ipa.c (can_replace_by_local_alias): Simplify.
1736 (ipa_profile): Produce speculative calls in non-lto, too;
1737 add simple cost model; produce local aliases.
1738
1739 2013-08-13 David Malcolm <dmalcolm@redhat.com>
1740
1741 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
1742 PASS_MANAGER_H.
1743
1744 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
1745
1746 * config/i386/i386.c (ix86_function_versions): Use error + inform.
1747
1748 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
1749
1750 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
1751 iterator instead of X87MODEF.
1752
1753 2013-08-12 Perez Read <netfirewall@gmail.com>
1754
1755 PR target/58132
1756 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
1757 operand 0 for intel asm alternative.
1758 (*movabs<mode>_2): Ditto for operand 1.
1759
1760 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
1761
1762 * config/aarch64/arm_none.h
1763 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
1764
1765 2013-08-12 Nick Clifton <nickc@redhat.com>
1766
1767 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
1768
1769 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1770
1771 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
1772 expand for QI/HImode operand to produce more effictive code for
1773 unsigned char(short) --> float(double) conversion.
1774
1775 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
1776
1777 * doc/invoke.texi: Mention that -ftls-model does not force the final
1778 model.
1779
1780 2013-08-12 Marek Polacek <polacek@redhat.com>
1781 Marc Glisse <marc.glisse@inria.fr>
1782
1783 PR tree-optimization/57980
1784 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
1785 when creating -1 constant.
1786
1787 2013-08-10 Jan Hubicka <jh@suse.cz>
1788
1789 Workaround binutils PR14342.
1790 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
1791 (gimple_init_edge_profiler): Likewise.
1792 (gimple_gen_ic_func_profiler): Likewise.
1793
1794 2013-08-09 Jan Hubicka <jh@suse.cz>
1795
1796 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
1797
1798 2013-08-09 Xinliang David Li <davidxl@google.com>
1799
1800 * config/i386/stringop.def: New file.
1801 * config/i386/stringop.opt: New file.
1802 * config/i386/i386-opts.h: Include stringopt.def.
1803 * config/i386/i386.opt: Include stringopt.opt.
1804 * config/i386/i386.c (ix86_option_override_internal):
1805 Override default size based stringop inline strategies with options.
1806 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1807 New function.
1808
1809 2013-08-09 Jan Hubicka <jh@suse.cz>
1810
1811 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1812
1813 2013-08-09 Jan Hubicka <jh@suse.cz>
1814
1815 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1816 CGRAPH_FREQ_MAX.
1817 (dump_cgraph_node): Dump profile-id.
1818 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1819 and common_target_probability.
1820 * lto-cgraph.c (lto_output_edge): Stream common targets.
1821 (lto_output_node): Stream profile ids.
1822 (input_node): Stream profile ids.
1823 (input_edge): Stream common targets.
1824 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1825 * ipa.c: Include value-prof.h
1826 (ipa_profile_generate_summary): Turn indirect call statement histograms
1827 into common targets.
1828 (ipa_profile): Turn common targets into speculative edges.
1829
1830 2013-08-09 Jan Hubicka <jh@suse.cz>
1831
1832 * cgraph.h (cgraph_node): Add profile_id.
1833 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1834 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1835 (del_node_map): Update.
1836 (find_func_by_funcdef_no): Replace by ...
1837 (find_func_by_profile_id): ... this one.
1838 (gimple_ic_transform): Do not remove useful histograms when
1839 speculation is not done; dump info when indirect call removal
1840 can happen at LTO.
1841 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1842 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1843 (__gcov_indirect_call_profiler_v2): .. this one.
1844 * profile.h (init_node_map): Update.
1845 * coverage.c (coverage_compute_profile_id): New function.
1846 * coverage.h (coverage_compute_profile_id): Declare.
1847 * tree-profile.c (init_ic_make_global_vars): Make
1848 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1849 (gimple_init_edge_profiler): Update prototype of
1850 __gcov_indirect_call_profiler.
1851 (gimple_gen_ic_func_profiler): Simplify.
1852 (tree_profiling): Use init_node_map
1853
1854 2013-08-09 Jan Hubicka <jh@suse.cz>
1855
1856 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1857 non-speculative refs.
1858 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1859 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1860 (cgraph_set_call_stmt): Likewise.
1861 (cgraph_create_edge_1): Fix release checking compilatoin;
1862 clear lto_stmt_uid.
1863 (cgraph_free_edge): Free indirect info.
1864 (cgraph_turn_edge_to_speculative): New function.
1865 (cgraph_speculative_call_info): New function.
1866 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1867 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1868 (dump_cgraph_node): Dump speculation.
1869 (verify_edge_count_and_frequency): Accept speculative edges.
1870 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1871 (verify_cgraph_node): Handle speculation.
1872 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1873 (cgraph_set_call_stmt): Update prototype.
1874 (cgraph_make_edge_direct): Update prototype.
1875 (cgraph_speculative_call_info): Declare.
1876 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1877 to change; update call of ipa_find_references.
1878 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1879 lto_stmt_uid and speculative flags.
1880 (ipa_dump_references): Dump speculation.
1881 (ipa_clone_references): Clone speculative flag.
1882 (ipa_clone_referring): Likewise.
1883 (ipa_clone_ref): New function.
1884 (ipa_find_reference): Look into lto_stmt_uids
1885 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1886 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1887 (ipa_find_reference): Update declaration.
1888 (ipa_clone_ref): Declare.
1889 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1890 stream speculative flag.
1891 (lto_output_ref): Stream statements uids and speculation.
1892 (input_ref): Likewise.
1893 (input_edge): Stream speuclation.
1894 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1895 (cgraph_set_call_stmt_including_clones): Handle speculation.
1896 * ipa-inline.c (heap_edge_removal_hook): New function.
1897 (inline_small_functions): Register it.
1898 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1899 also initialize refs.
1900 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1901 edge to change.
1902 (try_make_edge_direct_simple_call): Likewise.
1903 (try_make_edge_direct_simple_call): Likewise.
1904 (update_indirect_edges_after_inlining): Likewise.
1905 (remove_described_reference): Look proper lto_stmt_uid.
1906 (propagate_controlled_uses): Likewise.
1907 (propagate_controlled_uses): Liekwise.
1908 * tree-inline.c (copy_bb): Copy speculative edges.
1909 (redirect_all_calls): New function.
1910 (copy_cfg_body): Do redirection after loop info is updated.
1911 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1912
1913 2013-08-09 Jan Hubicka <jh@suse.cz>
1914
1915 * lto-streamer-out.c (output_function): Renumber PHIs.
1916 * lto-streamer-in.c (input_function): Likewise.
1917
1918 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1919
1920 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1921 (get_lane_unsigned): Likewise.
1922 (dup_lane_scalar): Likewise.
1923 (get_lane): enable for VALL.
1924 * config/aarch64/aarch64-simd.md
1925 (aarch64_dup_lane_scalar<mode>): Remove.
1926 (aarch64_get_lane_signed<mode>): Likewise.
1927 (aarch64_get_lane_unsigned<mode>): Likewise.
1928 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1929 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1930 (aarch64_get_lane<mode>): Enable for all vector modes.
1931 (aarch64_get_lanedi): Remove misleading constraints.
1932 * config/aarch64/arm_neon.h
1933 (__aarch64_vget_lane_any): Define.
1934 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1935 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1936 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1937 * config/aarch64/iterators.md (VDQQH): New.
1938 (VDQQHS): Likewise.
1939 (vwcore): Likewise.
1940
1941 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1942
1943 * configure.ac: Add GAS check for LEON instructions on SPARC.
1944 * configure: Regenerate.
1945 * config.in: Likewise.
1946 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1947 sparc*-*-* block.
1948 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1949 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1950 for LEON or LEON3.
1951 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1952 (AS_LEON_FLAG): New macro.
1953 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1954 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1955 Deal with LEON and LEON3 for the memory model.
1956 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1957 (atomic_compare_and_swap<mode>_1): Likewise.
1958 (*atomic_compare_and_swap<mode>_1): Likewise.
1959
1960 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1961
1962 * config/arm/neon.md (vcond): Fix floating-point vector
1963 comparisons against 0.
1964
1965 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1966
1967 * lra-constraints.c (emit_spill_move): Remove assert.
1968 (process_alt_operands): Add more debugging
1969 output. Increase reject for spilling into memory. Decrease
1970 reject for reloading scratch.
1971 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1972
1973 2013-08-08 Steve Ellcey <sellcey@mips.com>
1974
1975 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1976 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1977 micromips incompatible. Add nan2008.
1978 (MULTILIB_DIRNAMES): Add nan2008.
1979 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1980 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1981 and micromips incompatible. Add nan2008.
1982 (MULTILIB_DIRNAMES): Add nan2008.
1983 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1984
1985 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1986
1987 PR rtl-optimization/58079
1988 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1989 simplify_comparison has widened a comparison with an integer.
1990
1991 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1992
1993 * config/arm/neon.md (movmisalign<mode>): Disable when we
1994 don't allow unaligned accesses.
1995 (*movmisalign<mode>_neon_store): Likewise.
1996 (*movmisalign<mode>_neon_load): Likewise.
1997 (*movmisalign<mode>_neon_store): Likewise.
1998 (*movmisalign<mode>_neon_load): Likewise.
1999
2000 2013-08-08 Jan Hubicka <jh@suse.cz>
2001
2002 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
2003 (make_pass_rebuild_cgraph_edges): Also clear references.
2004 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
2005 * ipa-inline-transform.c (inline_transform): Remove all references
2006 after inlining.
2007 * cgraphunit.c (expand_function): Remove all references after
2008 expansion.
2009 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
2010 (ipa_find_reference): Rewrite to iterator.
2011 (remove_stmt_references): Likewise.
2012 (ipa_clear_stmts_in_references): New function.
2013 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
2014 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
2015 clear references.
2016 * ipa-split.c (split_function): Remove references in split function.
2017
2018 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2019
2020 PR target/57431
2021 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
2022 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
2023
2024 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2025
2026 PR target/56979
2027 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2028 suggested mode for the assignment isn't compatible with the
2029 registers required.
2030
2031 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2032
2033 PR target/58065
2034 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2035
2036 2013-08-07 Xinliang David Li <davidxl@google.com>
2037
2038 * config/i386/i386.opt: New option -mtune-ctrl=.
2039 * config/i386/x86-tune.def: New file.
2040 * config/i386/i386.h: include x86-tune.def.
2041 * config/i386/i386.c (ix86_option_override_internal):
2042 Parsing -mtune-ctrl= option and set tune features.
2043
2044 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2045
2046 PR other/12081
2047 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2048 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2049 to gen_2arg_fn_t.
2050
2051 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2052
2053 * rtl.h (update_alignments): Declare.
2054 * final.c (grow_label_align): New function extracted from...
2055 (shorten_branches): ...here. Call it.
2056 (update_alignments): New function.
2057 * reorg.c (sibling_labels): New variable.
2058 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2059 the new label along with it onto the sibling_labels vector.
2060 (fill_simple_delay_slots): Adjust call to get_label_before.
2061 (fill_slots_from_thread): Likewise.
2062 (relax_delay_slots): Likewise.
2063 (make_return_insns): Likewise.
2064 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2065
2066 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2067
2068 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2069 document its semantics.
2070 (diagnostic_report_diagnostic): Adjust accordingly.
2071
2072 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2073
2074 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2075 (sparc_option_override): ...and port to new C++ pass API.
2076 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2077
2078 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2079
2080 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2081
2082 2013-08-06 Caroline Tice <cmtice@google.com>
2083
2084 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2085 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2086 * tree-pass.h: Add pass_vtable_verify.
2087 * varasm.c (assemble_variable): Add code to properly set the comdat
2088 section and name for the .vtable_map_vars section.
2089 (assemble_vtyv_preinit_initializer): New function.
2090 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2091 LINK_ONCE flag.
2092 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2093 * vtable-verify.c: New file.
2094 * vtable-verify.h: New file.
2095 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2096 initialiation levels.
2097 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2098 * passes.def: Insert pass_vtable_verify.
2099 * aclocal.m4: Reorder includes.
2100 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2101 -fvtv-counts options.
2102 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2103 as appropriate, if -fvtable-verify=... is used.
2104 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2105 -fvtable-verify=... is used.
2106 * Makefile.in (OBJS): Add vtable-verify.o to list.
2107 (vtable-verify.o): Add new build rule.
2108 (GTFILES): Add vtable-verify.c to list.
2109 * common.opt (fvtable-verify=): New flag.
2110 (vtv_priority): Values for fvtable-verify= flag.
2111 (fvtv-counts): New flag.
2112 (fvtv-debug): New flag.
2113 * tree.h (save_vtable_map_decl): New extern function decl.
2114
2115 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2116
2117 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2118 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2119 (pass_data_rl78_devirt): ...new pass_data instance and...
2120 (make_pass_rl78_devirt): ...new function.
2121 (rl78_asm_file_start): Port pass registration to new C++ API.
2122
2123 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2124
2125 * coretypes.h (rtl_opt_pass): Add.
2126 (gcc::context): Add.
2127 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2128 (epiphany_init): Port to new C++ pass API.
2129 (epiphany_optimize_mode_switching): Likewise.
2130 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2131 (pass_manager::get_pass_mode_switching): New.
2132 (pass_manager::get_pass_peephole2): New.
2133 * mode-switching.c (pass_mode_switching): Add clone method.
2134 * recog.c (pass_peephole2): Add clone method.
2135 (pass_split_all_insns): Add clone method.
2136
2137 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2138
2139 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2140 (mips_option_override): ...here, porting to new C++ API for passes.
2141
2142 2013-08-06 Jan Hubicka <jh@suse.cz>
2143
2144 * cgraph.c (cgraph_get_body): New function based on lto.c
2145 implementation.
2146 * cgraph.h (cgraph_get_body): Declare.
2147 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2148 LTO paths.
2149 * cgraphunit.c (expand_function): Get body prior expanding.
2150 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2151 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2152 really need.
2153 * passes.c (do_per_function_toporder): Get body.
2154 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2155 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2156
2157 2013-08-06 Martin Jambor <mjambor@suse.cz>
2158
2159 PR fortran/57987
2160 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2161 is not re-finalized. Rename second parameter to no_collect.
2162
2163 2013-08-06 Martin Jambor <mjambor@suse.cz>
2164
2165 PR middle-end/58041
2166 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2167 MEM_REF has proper alignment information.
2168
2169 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2170
2171 PR other/12081
2172 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2173 class insn_gen_fn.
2174 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2175 rtx (*) (rtx, ...) with insn_gen_fn.
2176 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2177 insn_gen_fn::stored_funcptr. Add initializer braces.
2178
2179 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2180
2181 Rewrite how instances of passes are cloned to remove assumptions
2182 about their sizes (thus allowing pass subclasses to have
2183 additional data fields, albeit non-GC-managed ones at this point).
2184
2185 * passes.c (make_pass_instance): Now that passes have clone
2186 methods, rewrite this function to eliminate XNEW and memcpy
2187 calls that used hardcoded sizes. Since this function no longer
2188 creates pass instances, rename it to...
2189 (add_pass_instance): ...this. Document the old way that passes were
2190 numbered and flagged, and rework this function to continue using it.
2191 (next_pass_1): Add an initial_pass argument for use by
2192 add_pass_instance.
2193 (position_pass): When adding multiple instances of a pass, use
2194 the pass's clone method, rather than relying on the XNEW/memcpy
2195 within the former make_pass_instance (now add_pass_instance).
2196 (pass_manager::pass_manager): When invoking next_pass_1, also supply
2197 the initial instance of the current pass within the pass manager.
2198
2199 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2200
2201 This is the automated part of the conversion of passes from C
2202 structs to C++ classes.
2203
2204 Patch autogenerated by refactor_passes.py from
2205 https://github.com/davidmalcolm/gcc-refactoring-scripts
2206 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
2207
2208 * asan.c (pass_asan): Convert from a global struct to a subclass of
2209 gimple_opt_pass along with...
2210 (pass_data_asan): ...new pass_data instance and...
2211 (make_pass_asan): ...new function.
2212 (pass_asan_O0): Convert from a global struct to a subclass of
2213 gimple_opt_pass along with...
2214 (pass_data_asan_O0): ...new pass_data instance and...
2215 (make_pass_asan_O0): ...new function.
2216 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
2217 subclass of rtl_opt_pass along with...
2218 (pass_data_inc_dec): ...new pass_data instance and...
2219 (make_pass_inc_dec): ...new function.
2220 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
2221 a subclass of rtl_opt_pass along with...
2222 (pass_data_reorder_blocks): ...new pass_data instance and...
2223 (make_pass_reorder_blocks): ...new function.
2224 (pass_duplicate_computed_gotos): Convert from a global struct to a
2225 subclass of rtl_opt_pass along with...
2226 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
2227 (make_pass_duplicate_computed_gotos): ...new function.
2228 (pass_partition_blocks): Convert from a global struct to a subclass of
2229 rtl_opt_pass along with...
2230 (pass_data_partition_blocks): ...new pass_data instance and...
2231 (make_pass_partition_blocks): ...new function.
2232 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
2233 struct to a subclass of rtl_opt_pass along with...
2234 (pass_data_branch_target_load_optimize1): ...new pass_data instance
2235 and...
2236 (make_pass_branch_target_load_optimize1): ...new function.
2237 (pass_branch_target_load_optimize2): Convert from a global struct to a
2238 subclass of rtl_opt_pass along with...
2239 (pass_data_branch_target_load_optimize2): ...new pass_data instance
2240 and...
2241 (make_pass_branch_target_load_optimize2): ...new function.
2242 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
2243 of rtl_opt_pass along with...
2244 (pass_data_jump): ...new pass_data instance and...
2245 (make_pass_jump): ...new function.
2246 (pass_jump2): Convert from a global struct to a subclass of
2247 rtl_opt_pass along with...
2248 (pass_data_jump2): ...new pass_data instance and...
2249 (make_pass_jump2): ...new function.
2250 * cfgexpand.c (pass_expand): Convert from a global struct to a
2251 subclass of rtl_opt_pass along with...
2252 (pass_data_expand): ...new pass_data instance and...
2253 (make_pass_expand): ...new function.
2254 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
2255 of rtl_opt_pass along with...
2256 (pass_data_free_cfg): ...new pass_data instance and...
2257 (make_pass_free_cfg): ...new function.
2258 (pass_into_cfg_layout_mode): Convert from a global struct to a
2259 subclass of rtl_opt_pass along with...
2260 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
2261 (make_pass_into_cfg_layout_mode): ...new function.
2262 (pass_outof_cfg_layout_mode): Convert from a global struct to a
2263 subclass of rtl_opt_pass along with...
2264 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
2265 (make_pass_outof_cfg_layout_mode): ...new function.
2266 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
2267 struct to a subclass of gimple_opt_pass along with...
2268 (pass_data_build_cgraph_edges): ...new pass_data instance and...
2269 (make_pass_build_cgraph_edges): ...new function.
2270 (pass_rebuild_cgraph_edges): Convert from a global struct to a
2271 subclass of gimple_opt_pass along with...
2272 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
2273 (make_pass_rebuild_cgraph_edges): ...new function.
2274 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
2275 subclass of gimple_opt_pass along with...
2276 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
2277 and...
2278 (make_pass_remove_cgraph_callee_edges): ...new function.
2279 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
2280 struct to a subclass of rtl_opt_pass along with...
2281 (pass_data_stack_adjustments): ...new pass_data instance and...
2282 (make_pass_stack_adjustments): ...new function.
2283 * combine.c (pass_combine): Convert from a global struct to a subclass
2284 of rtl_opt_pass along with...
2285 (pass_data_combine): ...new pass_data instance and...
2286 (make_pass_combine): ...new function.
2287 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
2288 global struct to a subclass of rtl_opt_pass along with...
2289 (pass_data_compare_elim_after_reload): ...new pass_data instance
2290 and...
2291 (make_pass_compare_elim_after_reload): ...new function.
2292 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
2293 of rtl_opt_pass along with...
2294 (pass_data_rtl_cprop): ...new pass_data instance and...
2295 (make_pass_rtl_cprop): ...new function.
2296 * cse.c (pass_cse): Convert from a global struct to a subclass of
2297 rtl_opt_pass along with...
2298 (pass_data_cse): ...new pass_data instance and...
2299 (make_pass_cse): ...new function.
2300 (pass_cse2): Convert from a global struct to a subclass of
2301 rtl_opt_pass along with...
2302 (pass_data_cse2): ...new pass_data instance and...
2303 (make_pass_cse2): ...new function.
2304 (pass_cse_after_global_opts): Convert from a global struct to a
2305 subclass of rtl_opt_pass along with...
2306 (pass_data_cse_after_global_opts): ...new pass_data instance and...
2307 (make_pass_cse_after_global_opts): ...new function.
2308 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
2309 of rtl_opt_pass along with...
2310 (pass_data_ud_rtl_dce): ...new pass_data instance and...
2311 (make_pass_ud_rtl_dce): ...new function.
2312 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
2313 rtl_opt_pass along with...
2314 (pass_data_fast_rtl_dce): ...new pass_data instance and...
2315 (make_pass_fast_rtl_dce): ...new function.
2316 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
2317 a subclass of rtl_opt_pass along with...
2318 (pass_data_df_initialize_opt): ...new pass_data instance and...
2319 (make_pass_df_initialize_opt): ...new function.
2320 (pass_df_initialize_no_opt): Convert from a global struct to a
2321 subclass of rtl_opt_pass along with...
2322 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
2323 (make_pass_df_initialize_no_opt): ...new function.
2324 (pass_df_finish): Convert from a global struct to a subclass of
2325 rtl_opt_pass along with...
2326 (pass_data_df_finish): ...new pass_data instance and...
2327 (make_pass_df_finish): ...new function.
2328 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
2329 rtl_opt_pass along with...
2330 (pass_data_rtl_dse1): ...new pass_data instance and...
2331 (make_pass_rtl_dse1): ...new function.
2332 (pass_rtl_dse2): Convert from a global struct to a subclass of
2333 rtl_opt_pass along with...
2334 (pass_data_rtl_dse2): ...new pass_data instance and...
2335 (make_pass_rtl_dse2): ...new function.
2336 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
2337 subclass of rtl_opt_pass along with...
2338 (pass_data_dwarf2_frame): ...new pass_data instance and...
2339 (make_pass_dwarf2_frame): ...new function.
2340 * except.c (pass_set_nothrow_function_flags): Convert from a global
2341 struct to a subclass of rtl_opt_pass along with...
2342 (pass_data_set_nothrow_function_flags): ...new pass_data instance
2343 and...
2344 (make_pass_set_nothrow_function_flags): ...new function.
2345 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
2346 subclass of rtl_opt_pass along with...
2347 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
2348 and...
2349 (make_pass_convert_to_eh_region_ranges): ...new function.
2350 * final.c (pass_compute_alignments): Convert from a global struct to a
2351 subclass of rtl_opt_pass along with...
2352 (pass_data_compute_alignments): ...new pass_data instance and...
2353 (make_pass_compute_alignments): ...new function.
2354 (pass_final): Convert from a global struct to a subclass of
2355 rtl_opt_pass along with...
2356 (pass_data_final): ...new pass_data instance and...
2357 (make_pass_final): ...new function.
2358 (pass_shorten_branches): Convert from a global struct to a subclass of
2359 rtl_opt_pass along with...
2360 (pass_data_shorten_branches): ...new pass_data instance and...
2361 (make_pass_shorten_branches): ...new function.
2362 (pass_clean_state): Convert from a global struct to a subclass of
2363 rtl_opt_pass along with...
2364 (pass_data_clean_state): ...new pass_data instance and...
2365 (make_pass_clean_state): ...new function.
2366 * function.c (pass_instantiate_virtual_regs): Convert from a global
2367 struct to a subclass of rtl_opt_pass along with...
2368 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
2369 (make_pass_instantiate_virtual_regs): ...new function.
2370 (pass_leaf_regs): Convert from a global struct to a subclass of
2371 rtl_opt_pass along with...
2372 (pass_data_leaf_regs): ...new pass_data instance and...
2373 (make_pass_leaf_regs): ...new function.
2374 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
2375 subclass of rtl_opt_pass along with...
2376 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
2377 and...
2378 (make_pass_thread_prologue_and_epilogue): ...new function.
2379 (pass_match_asm_constraints): Convert from a global struct to a
2380 subclass of rtl_opt_pass along with...
2381 (pass_data_match_asm_constraints): ...new pass_data instance and...
2382 (make_pass_match_asm_constraints): ...new function.
2383 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
2384 subclass of rtl_opt_pass along with...
2385 (pass_data_rtl_fwprop): ...new pass_data instance and...
2386 (make_pass_rtl_fwprop): ...new function.
2387 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
2388 rtl_opt_pass along with...
2389 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
2390 (make_pass_rtl_fwprop_addr): ...new function.
2391 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
2392 rtl_opt_pass along with...
2393 (pass_data_rtl_pre): ...new pass_data instance and...
2394 (make_pass_rtl_pre): ...new function.
2395 (pass_rtl_hoist): Convert from a global struct to a subclass of
2396 rtl_opt_pass along with...
2397 (pass_data_rtl_hoist): ...new pass_data instance and...
2398 (make_pass_rtl_hoist): ...new function.
2399 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
2400 subclass of gimple_opt_pass along with...
2401 (pass_data_lower_cf): ...new pass_data instance and...
2402 (make_pass_lower_cf): ...new function.
2403 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
2404 from a global struct to a subclass of gimple_opt_pass along with...
2405 (pass_data_strength_reduction): ...new pass_data instance and...
2406 (make_pass_strength_reduction): ...new function.
2407 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
2408 of rtl_opt_pass along with...
2409 (pass_data_rtl_ifcvt): ...new pass_data instance and...
2410 (make_pass_rtl_ifcvt): ...new function.
2411 (pass_if_after_combine): Convert from a global struct to a subclass of
2412 rtl_opt_pass along with...
2413 (pass_data_if_after_combine): ...new pass_data instance and...
2414 (make_pass_if_after_combine): ...new function.
2415 (pass_if_after_reload): Convert from a global struct to a subclass of
2416 rtl_opt_pass along with...
2417 (pass_data_if_after_reload): ...new pass_data instance and...
2418 (make_pass_if_after_reload): ...new function.
2419 * init-regs.c (pass_initialize_regs): Convert from a global struct to
2420 a subclass of rtl_opt_pass along with...
2421 (pass_data_initialize_regs): ...new pass_data instance and...
2422 (make_pass_initialize_regs): ...new function.
2423 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
2424 of ipa_opt_pass_d along with...
2425 (pass_data_ipa_cp): ...new pass_data instance and...
2426 (make_pass_ipa_cp): ...new function.
2427 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
2428 global struct to a subclass of gimple_opt_pass along with...
2429 (pass_data_inline_parameters): ...new pass_data instance and...
2430 (make_pass_inline_parameters): ...new function.
2431 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
2432 subclass of gimple_opt_pass along with...
2433 (pass_data_early_inline): ...new pass_data instance and...
2434 (make_pass_early_inline): ...new function.
2435 (pass_ipa_inline): Convert from a global struct to a subclass of
2436 ipa_opt_pass_d along with...
2437 (pass_data_ipa_inline): ...new pass_data instance and...
2438 (make_pass_ipa_inline): ...new function.
2439 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
2440 struct to a subclass of gimple_opt_pass along with...
2441 (pass_data_local_pure_const): ...new pass_data instance and...
2442 (make_pass_local_pure_const): ...new function.
2443 (pass_ipa_pure_const): Convert from a global struct to a subclass of
2444 ipa_opt_pass_d along with...
2445 (pass_data_ipa_pure_const): ...new pass_data instance and...
2446 (make_pass_ipa_pure_const): ...new function.
2447 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
2448 to a subclass of ipa_opt_pass_d along with...
2449 (pass_data_ipa_reference): ...new pass_data instance and...
2450 (make_pass_ipa_reference): ...new function.
2451 * ipa-split.c (pass_split_functions): Convert from a global struct to
2452 a subclass of gimple_opt_pass along with...
2453 (pass_data_split_functions): ...new pass_data instance and...
2454 (make_pass_split_functions): ...new function.
2455 (pass_feedback_split_functions): Convert from a global struct to a
2456 subclass of gimple_opt_pass along with...
2457 (pass_data_feedback_split_functions): ...new pass_data instance and...
2458 (make_pass_feedback_split_functions): ...new function.
2459 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
2460 global struct to a subclass of simple_ipa_opt_pass along with...
2461 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
2462 instance and...
2463 (make_pass_ipa_function_and_variable_visibility): ...new function.
2464 (pass_ipa_free_inline_summary): Convert from a global struct to a
2465 subclass of simple_ipa_opt_pass along with...
2466 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
2467 (make_pass_ipa_free_inline_summary): ...new function.
2468 (pass_ipa_whole_program_visibility): Convert from a global struct to a
2469 subclass of ipa_opt_pass_d along with...
2470 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
2471 and...
2472 (make_pass_ipa_whole_program_visibility): ...new function.
2473 (pass_ipa_profile): Convert from a global struct to a subclass of
2474 ipa_opt_pass_d along with...
2475 (pass_data_ipa_profile): ...new pass_data instance and...
2476 (make_pass_ipa_profile): ...new function.
2477 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
2478 ipa_opt_pass_d along with...
2479 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
2480 (make_pass_ipa_cdtor_merge): ...new function.
2481 * ira.c (pass_ira): Convert from a global struct to a subclass of
2482 rtl_opt_pass along with...
2483 (pass_data_ira): ...new pass_data instance and...
2484 (make_pass_ira): ...new function.
2485 (pass_reload): Convert from a global struct to a subclass of
2486 rtl_opt_pass along with...
2487 (pass_data_reload): ...new pass_data instance and...
2488 (make_pass_reload): ...new function.
2489 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
2490 subclass of rtl_opt_pass along with...
2491 (pass_data_cleanup_barriers): ...new pass_data instance and...
2492 (make_pass_cleanup_barriers): ...new function.
2493 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
2494 of rtl_opt_pass along with...
2495 (pass_data_loop2): ...new pass_data instance and...
2496 (make_pass_loop2): ...new function.
2497 (pass_rtl_loop_init): Convert from a global struct to a subclass of
2498 rtl_opt_pass along with...
2499 (pass_data_rtl_loop_init): ...new pass_data instance and...
2500 (make_pass_rtl_loop_init): ...new function.
2501 (pass_rtl_loop_done): Convert from a global struct to a subclass of
2502 rtl_opt_pass along with...
2503 (pass_data_rtl_loop_done): ...new pass_data instance and...
2504 (make_pass_rtl_loop_done): ...new function.
2505 (pass_rtl_move_loop_invariants): Convert from a global struct to a
2506 subclass of rtl_opt_pass along with...
2507 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
2508 (make_pass_rtl_move_loop_invariants): ...new function.
2509 (pass_rtl_unswitch): Convert from a global struct to a subclass of
2510 rtl_opt_pass along with...
2511 (pass_data_rtl_unswitch): ...new pass_data instance and...
2512 (make_pass_rtl_unswitch): ...new function.
2513 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
2514 subclass of rtl_opt_pass along with...
2515 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
2516 and...
2517 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2518 (pass_rtl_doloop): Convert from a global struct to a subclass of
2519 rtl_opt_pass along with...
2520 (pass_data_rtl_doloop): ...new pass_data instance and...
2521 (make_pass_rtl_doloop): ...new function.
2522 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
2523 a subclass of rtl_opt_pass along with...
2524 (pass_data_lower_subreg): ...new pass_data instance and...
2525 (make_pass_lower_subreg): ...new function.
2526 (pass_lower_subreg2): Convert from a global struct to a subclass of
2527 rtl_opt_pass along with...
2528 (pass_data_lower_subreg2): ...new pass_data instance and...
2529 (make_pass_lower_subreg2): ...new function.
2530 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
2531 struct to a subclass of ipa_opt_pass_d along with...
2532 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
2533 (make_pass_ipa_lto_gimple_out): ...new function.
2534 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
2535 of ipa_opt_pass_d along with...
2536 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
2537 (make_pass_ipa_lto_finish_out): ...new function.
2538 * mode-switching.c (pass_mode_switching): Convert from a global struct
2539 to a subclass of rtl_opt_pass along with...
2540 (pass_data_mode_switching): ...new pass_data instance and...
2541 (make_pass_mode_switching): ...new function.
2542 * modulo-sched.c (pass_sms): Convert from a global struct to a
2543 subclass of rtl_opt_pass along with...
2544 (pass_data_sms): ...new pass_data instance and...
2545 (make_pass_sms): ...new function.
2546 * omp-low.c (pass_expand_omp): Convert from a global struct to a
2547 subclass of gimple_opt_pass along with...
2548 (pass_data_expand_omp): ...new pass_data instance and...
2549 (make_pass_expand_omp): ...new function.
2550 (pass_lower_omp): Convert from a global struct to a subclass of
2551 gimple_opt_pass along with...
2552 (pass_data_lower_omp): ...new pass_data instance and...
2553 (make_pass_lower_omp): ...new function.
2554 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
2555 of gimple_opt_pass along with...
2556 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
2557 (make_pass_diagnose_omp_blocks): ...new function.
2558 * passes.c (pass_early_local_passes): Convert from a global struct to
2559 a subclass of simple_ipa_opt_pass along with...
2560 (pass_data_early_local_passes): ...new pass_data instance and...
2561 (make_pass_early_local_passes): ...new function.
2562 (pass_all_early_optimizations): Convert from a global struct to a
2563 subclass of gimple_opt_pass along with...
2564 (pass_data_all_early_optimizations): ...new pass_data instance and...
2565 (make_pass_all_early_optimizations): ...new function.
2566 (pass_all_optimizations): Convert from a global struct to a subclass
2567 of gimple_opt_pass along with...
2568 (pass_data_all_optimizations): ...new pass_data instance and...
2569 (make_pass_all_optimizations): ...new function.
2570 (pass_all_optimizations_g): Convert from a global struct to a subclass
2571 of gimple_opt_pass along with...
2572 (pass_data_all_optimizations_g): ...new pass_data instance and...
2573 (make_pass_all_optimizations_g): ...new function.
2574 (pass_rest_of_compilation): Convert from a global struct to a subclass
2575 of rtl_opt_pass along with...
2576 (pass_data_rest_of_compilation): ...new pass_data instance and...
2577 (make_pass_rest_of_compilation): ...new function.
2578 (pass_postreload): Convert from a global struct to a subclass of
2579 rtl_opt_pass along with...
2580 (pass_data_postreload): ...new pass_data instance and...
2581 (make_pass_postreload): ...new function.
2582 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
2583 subclass of rtl_opt_pass along with...
2584 (pass_data_gcse2): ...new pass_data instance and...
2585 (make_pass_gcse2): ...new function.
2586 * postreload.c (pass_postreload_cse): Convert from a global struct to
2587 a subclass of rtl_opt_pass along with...
2588 (pass_data_postreload_cse): ...new pass_data instance and...
2589 (make_pass_postreload_cse): ...new function.
2590 * predict.c (pass_profile): Convert from a global struct to a subclass
2591 of gimple_opt_pass along with...
2592 (pass_data_profile): ...new pass_data instance and...
2593 (make_pass_profile): ...new function.
2594 (pass_strip_predict_hints): Convert from a global struct to a subclass
2595 of gimple_opt_pass along with...
2596 (pass_data_strip_predict_hints): ...new pass_data instance and...
2597 (make_pass_strip_predict_hints): ...new function.
2598 * recog.c (pass_peephole2): Convert from a global struct to a subclass
2599 of rtl_opt_pass along with...
2600 (pass_data_peephole2): ...new pass_data instance and...
2601 (make_pass_peephole2): ...new function.
2602 (pass_split_all_insns): Convert from a global struct to a subclass of
2603 rtl_opt_pass along with...
2604 (pass_data_split_all_insns): ...new pass_data instance and...
2605 (make_pass_split_all_insns): ...new function.
2606 (pass_split_after_reload): Convert from a global struct to a subclass
2607 of rtl_opt_pass along with...
2608 (pass_data_split_after_reload): ...new pass_data instance and...
2609 (make_pass_split_after_reload): ...new function.
2610 (pass_split_before_regstack): Convert from a global struct to a
2611 subclass of rtl_opt_pass along with...
2612 (pass_data_split_before_regstack): ...new pass_data instance and...
2613 (make_pass_split_before_regstack): ...new function.
2614 (pass_split_before_sched2): Convert from a global struct to a subclass
2615 of rtl_opt_pass along with...
2616 (pass_data_split_before_sched2): ...new pass_data instance and...
2617 (make_pass_split_before_sched2): ...new function.
2618 (pass_split_for_shorten_branches): Convert from a global struct to a
2619 subclass of rtl_opt_pass along with...
2620 (pass_data_split_for_shorten_branches): ...new pass_data instance
2621 and...
2622 (make_pass_split_for_shorten_branches): ...new function.
2623 * ree.c (pass_ree): Convert from a global struct to a subclass of
2624 rtl_opt_pass along with...
2625 (pass_data_ree): ...new pass_data instance and...
2626 (make_pass_ree): ...new function.
2627 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
2628 subclass of rtl_opt_pass along with...
2629 (pass_data_stack_regs): ...new pass_data instance and...
2630 (make_pass_stack_regs): ...new function.
2631 (pass_stack_regs_run): Convert from a global struct to a subclass of
2632 rtl_opt_pass along with...
2633 (pass_data_stack_regs_run): ...new pass_data instance and...
2634 (make_pass_stack_regs_run): ...new function.
2635 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
2636 subclass of rtl_opt_pass along with...
2637 (pass_data_cprop_hardreg): ...new pass_data instance and...
2638 (make_pass_cprop_hardreg): ...new function.
2639 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
2640 subclass of rtl_opt_pass along with...
2641 (pass_data_reginfo_init): ...new pass_data instance and...
2642 (make_pass_reginfo_init): ...new function.
2643 * regmove.c (pass_regmove): Convert from a global struct to a subclass
2644 of rtl_opt_pass along with...
2645 (pass_data_regmove): ...new pass_data instance and...
2646 (make_pass_regmove): ...new function.
2647 * regrename.c (pass_regrename): Convert from a global struct to a
2648 subclass of rtl_opt_pass along with...
2649 (pass_data_regrename): ...new pass_data instance and...
2650 (make_pass_regrename): ...new function.
2651 * reorg.c (pass_delay_slots): Convert from a global struct to a
2652 subclass of rtl_opt_pass along with...
2653 (pass_data_delay_slots): ...new pass_data instance and...
2654 (make_pass_delay_slots): ...new function.
2655 (pass_machine_reorg): Convert from a global struct to a subclass of
2656 rtl_opt_pass along with...
2657 (pass_data_machine_reorg): ...new pass_data instance and...
2658 (make_pass_machine_reorg): ...new function.
2659 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
2660 of rtl_opt_pass along with...
2661 (pass_data_sched): ...new pass_data instance and...
2662 (make_pass_sched): ...new function.
2663 (pass_sched2): Convert from a global struct to a subclass of
2664 rtl_opt_pass along with...
2665 (pass_data_sched2): ...new pass_data instance and...
2666 (make_pass_sched2): ...new function.
2667 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
2668 to a subclass of rtl_opt_pass along with...
2669 (pass_data_stack_ptr_mod): ...new pass_data instance and...
2670 (make_pass_stack_ptr_mod): ...new function.
2671 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
2672 to a subclass of rtl_opt_pass along with...
2673 (pass_data_rtl_store_motion): ...new pass_data instance and...
2674 (make_pass_rtl_store_motion): ...new function.
2675 * tracer.c (pass_tracer): Convert from a global struct to a subclass
2676 of gimple_opt_pass along with...
2677 (pass_data_tracer): ...new pass_data instance and...
2678 (make_pass_tracer): ...new function.
2679 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
2680 to a subclass of gimple_opt_pass along with...
2681 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
2682 (make_pass_diagnose_tm_blocks): ...new function.
2683 (pass_lower_tm): Convert from a global struct to a subclass of
2684 gimple_opt_pass along with...
2685 (pass_data_lower_tm): ...new pass_data instance and...
2686 (make_pass_lower_tm): ...new function.
2687 (pass_tm_init): Convert from a global struct to a subclass of
2688 gimple_opt_pass along with...
2689 (pass_data_tm_init): ...new pass_data instance and...
2690 (make_pass_tm_init): ...new function.
2691 (pass_tm_mark): Convert from a global struct to a subclass of
2692 gimple_opt_pass along with...
2693 (pass_data_tm_mark): ...new pass_data instance and...
2694 (make_pass_tm_mark): ...new function.
2695 (pass_tm_edges): Convert from a global struct to a subclass of
2696 gimple_opt_pass along with...
2697 (pass_data_tm_edges): ...new pass_data instance and...
2698 (make_pass_tm_edges): ...new function.
2699 (pass_tm_memopt): Convert from a global struct to a subclass of
2700 gimple_opt_pass along with...
2701 (pass_data_tm_memopt): ...new pass_data instance and...
2702 (make_pass_tm_memopt): ...new function.
2703 (pass_ipa_tm): Convert from a global struct to a subclass of
2704 simple_ipa_opt_pass along with...
2705 (pass_data_ipa_tm): ...new pass_data instance and...
2706 (make_pass_ipa_tm): ...new function.
2707 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
2708 subclass of gimple_opt_pass along with...
2709 (pass_data_call_cdce): ...new pass_data instance and...
2710 (make_pass_call_cdce): ...new function.
2711 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
2712 subclass of gimple_opt_pass along with...
2713 (pass_data_build_cfg): ...new pass_data instance and...
2714 (make_pass_build_cfg): ...new function.
2715 (pass_split_crit_edges): Convert from a global struct to a subclass of
2716 gimple_opt_pass along with...
2717 (pass_data_split_crit_edges): ...new pass_data instance and...
2718 (make_pass_split_crit_edges): ...new function.
2719 (pass_warn_function_return): Convert from a global struct to a
2720 subclass of gimple_opt_pass along with...
2721 (pass_data_warn_function_return): ...new pass_data instance and...
2722 (make_pass_warn_function_return): ...new function.
2723 (pass_warn_function_noreturn): Convert from a global struct to a
2724 subclass of gimple_opt_pass along with...
2725 (pass_data_warn_function_noreturn): ...new pass_data instance and...
2726 (make_pass_warn_function_noreturn): ...new function.
2727 (pass_warn_unused_result): Convert from a global struct to a subclass
2728 of gimple_opt_pass along with...
2729 (pass_data_warn_unused_result): ...new pass_data instance and...
2730 (make_pass_warn_unused_result): ...new function.
2731 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
2732 a subclass of gimple_opt_pass along with...
2733 (pass_data_merge_phi): ...new pass_data instance and...
2734 (make_pass_merge_phi): ...new function.
2735 * tree-complex.c (pass_lower_complex): Convert from a global struct to
2736 a subclass of gimple_opt_pass along with...
2737 (pass_data_lower_complex): ...new pass_data instance and...
2738 (make_pass_lower_complex): ...new function.
2739 (pass_lower_complex_O0): Convert from a global struct to a subclass of
2740 gimple_opt_pass along with...
2741 (pass_data_lower_complex_O0): ...new pass_data instance and...
2742 (make_pass_lower_complex_O0): ...new function.
2743 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
2744 subclass of gimple_opt_pass along with...
2745 (pass_data_lower_eh): ...new pass_data instance and...
2746 (make_pass_lower_eh): ...new function.
2747 (pass_refactor_eh): Convert from a global struct to a subclass of
2748 gimple_opt_pass along with...
2749 (pass_data_refactor_eh): ...new pass_data instance and...
2750 (make_pass_refactor_eh): ...new function.
2751 (pass_lower_resx): Convert from a global struct to a subclass of
2752 gimple_opt_pass along with...
2753 (pass_data_lower_resx): ...new pass_data instance and...
2754 (make_pass_lower_resx): ...new function.
2755 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
2756 of gimple_opt_pass along with...
2757 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
2758 (make_pass_lower_eh_dispatch): ...new function.
2759 (pass_cleanup_eh): Convert from a global struct to a subclass of
2760 gimple_opt_pass along with...
2761 (pass_data_cleanup_eh): ...new pass_data instance and...
2762 (make_pass_cleanup_eh): ...new function.
2763 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
2764 to a subclass of simple_ipa_opt_pass along with...
2765 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
2766 (make_pass_ipa_lower_emutls): ...new function.
2767 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
2768 a subclass of gimple_opt_pass along with...
2769 (pass_data_if_conversion): ...new pass_data instance and...
2770 (make_pass_if_conversion): ...new function.
2771 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
2772 subclass of gimple_opt_pass along with...
2773 (pass_data_build_ssa): ...new pass_data instance and...
2774 (make_pass_build_ssa): ...new function.
2775 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
2776 global struct to a subclass of gimple_opt_pass along with...
2777 (pass_data_loop_distribution): ...new pass_data instance and...
2778 (make_pass_loop_distribution): ...new function.
2779 * tree-mudflap.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-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
2788 subclass of gimple_opt_pass along with...
2789 (pass_data_mudflap_1): ...new pass_data instance and...
2790 (make_pass_mudflap_1): ...new function.
2791 (pass_mudflap_2): Convert from a global struct to a subclass of
2792 gimple_opt_pass along with...
2793 (pass_data_mudflap_2): ...new pass_data instance and...
2794 (make_pass_mudflap_2): ...new function.
2795 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
2796 gimple_opt_pass along with...
2797 (pass_data_nrv): ...new pass_data instance and...
2798 (make_pass_nrv): ...new function.
2799 (pass_return_slot): Convert from a global struct to a subclass of
2800 gimple_opt_pass along with...
2801 (pass_data_return_slot): ...new pass_data instance and...
2802 (make_pass_return_slot): ...new function.
2803 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2804 to a subclass of gimple_opt_pass along with...
2805 (pass_data_object_sizes): ...new pass_data instance and...
2806 (make_pass_object_sizes): ...new function.
2807 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2808 global struct to a subclass of gimple_opt_pass along with...
2809 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2810 and...
2811 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2812 (pass_fixup_cfg): Convert from a global struct to a subclass of
2813 gimple_opt_pass along with...
2814 (pass_data_fixup_cfg): ...new pass_data instance and...
2815 (make_pass_fixup_cfg): ...new function.
2816 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2817 (make_pass_mudflap_1): ...new function.
2818 (pass_mudflap_2): Replace declaration with that of...
2819 (make_pass_mudflap_2): ...new function.
2820 (pass_asan): Replace declaration with that of...
2821 (make_pass_asan): ...new function.
2822 (pass_asan_O0): Replace declaration with that of...
2823 (make_pass_asan_O0): ...new function.
2824 (pass_tsan): Replace declaration with that of...
2825 (make_pass_tsan): ...new function.
2826 (pass_tsan_O0): Replace declaration with that of...
2827 (make_pass_tsan_O0): ...new function.
2828 (pass_lower_cf): Replace declaration with that of...
2829 (make_pass_lower_cf): ...new function.
2830 (pass_refactor_eh): Replace declaration with that of...
2831 (make_pass_refactor_eh): ...new function.
2832 (pass_lower_eh): Replace declaration with that of...
2833 (make_pass_lower_eh): ...new function.
2834 (pass_lower_eh_dispatch): Replace declaration with that of...
2835 (make_pass_lower_eh_dispatch): ...new function.
2836 (pass_lower_resx): Replace declaration with that of...
2837 (make_pass_lower_resx): ...new function.
2838 (pass_build_cfg): Replace declaration with that of...
2839 (make_pass_build_cfg): ...new function.
2840 (pass_early_tree_profile): Replace declaration with that of...
2841 (make_pass_early_tree_profile): ...new function.
2842 (pass_cleanup_eh): Replace declaration with that of...
2843 (make_pass_cleanup_eh): ...new function.
2844 (pass_sra): Replace declaration with that of...
2845 (make_pass_sra): ...new function.
2846 (pass_sra_early): Replace declaration with that of...
2847 (make_pass_sra_early): ...new function.
2848 (pass_early_ipa_sra): Replace declaration with that of...
2849 (make_pass_early_ipa_sra): ...new function.
2850 (pass_tail_recursion): Replace declaration with that of...
2851 (make_pass_tail_recursion): ...new function.
2852 (pass_tail_calls): Replace declaration with that of...
2853 (make_pass_tail_calls): ...new function.
2854 (pass_tree_loop): Replace declaration with that of...
2855 (make_pass_tree_loop): ...new function.
2856 (pass_tree_loop_init): Replace declaration with that of...
2857 (make_pass_tree_loop_init): ...new function.
2858 (pass_lim): Replace declaration with that of...
2859 (make_pass_lim): ...new function.
2860 (pass_tree_unswitch): Replace declaration with that of...
2861 (make_pass_tree_unswitch): ...new function.
2862 (pass_predcom): Replace declaration with that of...
2863 (make_pass_predcom): ...new function.
2864 (pass_iv_canon): Replace declaration with that of...
2865 (make_pass_iv_canon): ...new function.
2866 (pass_scev_cprop): Replace declaration with that of...
2867 (make_pass_scev_cprop): ...new function.
2868 (pass_empty_loop): Replace declaration with that of...
2869 (make_pass_empty_loop): ...new function.
2870 (pass_record_bounds): Replace declaration with that of...
2871 (make_pass_record_bounds): ...new function.
2872 (pass_graphite): Replace declaration with that of...
2873 (make_pass_graphite): ...new function.
2874 (pass_graphite_transforms): Replace declaration with that of...
2875 (make_pass_graphite_transforms): ...new function.
2876 (pass_if_conversion): Replace declaration with that of...
2877 (make_pass_if_conversion): ...new function.
2878 (pass_loop_distribution): Replace declaration with that of...
2879 (make_pass_loop_distribution): ...new function.
2880 (pass_vectorize): Replace declaration with that of...
2881 (make_pass_vectorize): ...new function.
2882 (pass_slp_vectorize): Replace declaration with that of...
2883 (make_pass_slp_vectorize): ...new function.
2884 (pass_complete_unroll): Replace declaration with that of...
2885 (make_pass_complete_unroll): ...new function.
2886 (pass_complete_unrolli): Replace declaration with that of...
2887 (make_pass_complete_unrolli): ...new function.
2888 (pass_parallelize_loops): Replace declaration with that of...
2889 (make_pass_parallelize_loops): ...new function.
2890 (pass_loop_prefetch): Replace declaration with that of...
2891 (make_pass_loop_prefetch): ...new function.
2892 (pass_iv_optimize): Replace declaration with that of...
2893 (make_pass_iv_optimize): ...new function.
2894 (pass_tree_loop_done): Replace declaration with that of...
2895 (make_pass_tree_loop_done): ...new function.
2896 (pass_ch): Replace declaration with that of...
2897 (make_pass_ch): ...new function.
2898 (pass_ccp): Replace declaration with that of...
2899 (make_pass_ccp): ...new function.
2900 (pass_phi_only_cprop): Replace declaration with that of...
2901 (make_pass_phi_only_cprop): ...new function.
2902 (pass_build_ssa): Replace declaration with that of...
2903 (make_pass_build_ssa): ...new function.
2904 (pass_build_alias): Replace declaration with that of...
2905 (make_pass_build_alias): ...new function.
2906 (pass_build_ealias): Replace declaration with that of...
2907 (make_pass_build_ealias): ...new function.
2908 (pass_dominator): Replace declaration with that of...
2909 (make_pass_dominator): ...new function.
2910 (pass_dce): Replace declaration with that of...
2911 (make_pass_dce): ...new function.
2912 (pass_dce_loop): Replace declaration with that of...
2913 (make_pass_dce_loop): ...new function.
2914 (pass_cd_dce): Replace declaration with that of...
2915 (make_pass_cd_dce): ...new function.
2916 (pass_call_cdce): Replace declaration with that of...
2917 (make_pass_call_cdce): ...new function.
2918 (pass_merge_phi): Replace declaration with that of...
2919 (make_pass_merge_phi): ...new function.
2920 (pass_split_crit_edges): Replace declaration with that of...
2921 (make_pass_split_crit_edges): ...new function.
2922 (pass_pre): Replace declaration with that of...
2923 (make_pass_pre): ...new function.
2924 (pass_profile): Replace declaration with that of...
2925 (make_pass_profile): ...new function.
2926 (pass_strip_predict_hints): Replace declaration with that of...
2927 (make_pass_strip_predict_hints): ...new function.
2928 (pass_lower_complex_O0): Replace declaration with that of...
2929 (make_pass_lower_complex_O0): ...new function.
2930 (pass_lower_complex): Replace declaration with that of...
2931 (make_pass_lower_complex): ...new function.
2932 (pass_lower_vector): Replace declaration with that of...
2933 (make_pass_lower_vector): ...new function.
2934 (pass_lower_vector_ssa): Replace declaration with that of...
2935 (make_pass_lower_vector_ssa): ...new function.
2936 (pass_lower_omp): Replace declaration with that of...
2937 (make_pass_lower_omp): ...new function.
2938 (pass_diagnose_omp_blocks): Replace declaration with that of...
2939 (make_pass_diagnose_omp_blocks): ...new function.
2940 (pass_expand_omp): Replace declaration with that of...
2941 (make_pass_expand_omp): ...new function.
2942 (pass_expand_omp_ssa): Replace declaration with that of...
2943 (make_pass_expand_omp_ssa): ...new function.
2944 (pass_object_sizes): Replace declaration with that of...
2945 (make_pass_object_sizes): ...new function.
2946 (pass_strlen): Replace declaration with that of...
2947 (make_pass_strlen): ...new function.
2948 (pass_fold_builtins): Replace declaration with that of...
2949 (make_pass_fold_builtins): ...new function.
2950 (pass_stdarg): Replace declaration with that of...
2951 (make_pass_stdarg): ...new function.
2952 (pass_early_warn_uninitialized): Replace declaration with that of...
2953 (make_pass_early_warn_uninitialized): ...new function.
2954 (pass_late_warn_uninitialized): Replace declaration with that of...
2955 (make_pass_late_warn_uninitialized): ...new function.
2956 (pass_cse_reciprocals): Replace declaration with that of...
2957 (make_pass_cse_reciprocals): ...new function.
2958 (pass_cse_sincos): Replace declaration with that of...
2959 (make_pass_cse_sincos): ...new function.
2960 (pass_optimize_bswap): Replace declaration with that of...
2961 (make_pass_optimize_bswap): ...new function.
2962 (pass_optimize_widening_mul): Replace declaration with that of...
2963 (make_pass_optimize_widening_mul): ...new function.
2964 (pass_warn_function_return): Replace declaration with that of...
2965 (make_pass_warn_function_return): ...new function.
2966 (pass_warn_function_noreturn): Replace declaration with that of...
2967 (make_pass_warn_function_noreturn): ...new function.
2968 (pass_cselim): Replace declaration with that of...
2969 (make_pass_cselim): ...new function.
2970 (pass_phiopt): Replace declaration with that of...
2971 (make_pass_phiopt): ...new function.
2972 (pass_forwprop): Replace declaration with that of...
2973 (make_pass_forwprop): ...new function.
2974 (pass_phiprop): Replace declaration with that of...
2975 (make_pass_phiprop): ...new function.
2976 (pass_tree_ifcombine): Replace declaration with that of...
2977 (make_pass_tree_ifcombine): ...new function.
2978 (pass_dse): Replace declaration with that of...
2979 (make_pass_dse): ...new function.
2980 (pass_nrv): Replace declaration with that of...
2981 (make_pass_nrv): ...new function.
2982 (pass_rename_ssa_copies): Replace declaration with that of...
2983 (make_pass_rename_ssa_copies): ...new function.
2984 (pass_sink_code): Replace declaration with that of...
2985 (make_pass_sink_code): ...new function.
2986 (pass_fre): Replace declaration with that of...
2987 (make_pass_fre): ...new function.
2988 (pass_check_data_deps): Replace declaration with that of...
2989 (make_pass_check_data_deps): ...new function.
2990 (pass_copy_prop): Replace declaration with that of...
2991 (make_pass_copy_prop): ...new function.
2992 (pass_vrp): Replace declaration with that of...
2993 (make_pass_vrp): ...new function.
2994 (pass_uncprop): Replace declaration with that of...
2995 (make_pass_uncprop): ...new function.
2996 (pass_return_slot): Replace declaration with that of...
2997 (make_pass_return_slot): ...new function.
2998 (pass_reassoc): Replace declaration with that of...
2999 (make_pass_reassoc): ...new function.
3000 (pass_rebuild_cgraph_edges): Replace declaration with that of...
3001 (make_pass_rebuild_cgraph_edges): ...new function.
3002 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
3003 (make_pass_remove_cgraph_callee_edges): ...new function.
3004 (pass_build_cgraph_edges): Replace declaration with that of...
3005 (make_pass_build_cgraph_edges): ...new function.
3006 (pass_local_pure_const): Replace declaration with that of...
3007 (make_pass_local_pure_const): ...new function.
3008 (pass_tracer): Replace declaration with that of...
3009 (make_pass_tracer): ...new function.
3010 (pass_warn_unused_result): Replace declaration with that of...
3011 (make_pass_warn_unused_result): ...new function.
3012 (pass_diagnose_tm_blocks): Replace declaration with that of...
3013 (make_pass_diagnose_tm_blocks): ...new function.
3014 (pass_lower_tm): Replace declaration with that of...
3015 (make_pass_lower_tm): ...new function.
3016 (pass_tm_init): Replace declaration with that of...
3017 (make_pass_tm_init): ...new function.
3018 (pass_tm_mark): Replace declaration with that of...
3019 (make_pass_tm_mark): ...new function.
3020 (pass_tm_memopt): Replace declaration with that of...
3021 (make_pass_tm_memopt): ...new function.
3022 (pass_tm_edges): Replace declaration with that of...
3023 (make_pass_tm_edges): ...new function.
3024 (pass_split_functions): Replace declaration with that of...
3025 (make_pass_split_functions): ...new function.
3026 (pass_feedback_split_functions): Replace declaration with that of...
3027 (make_pass_feedback_split_functions): ...new function.
3028 (pass_strength_reduction): Replace declaration with that of...
3029 (make_pass_strength_reduction): ...new function.
3030 (pass_ipa_lower_emutls): Replace declaration with that of...
3031 (make_pass_ipa_lower_emutls): ...new function.
3032 (pass_ipa_function_and_variable_visibility): Replace declaration with
3033 that of...
3034 (make_pass_ipa_function_and_variable_visibility): ...new function.
3035 (pass_ipa_tree_profile): Replace declaration with that of...
3036 (make_pass_ipa_tree_profile): ...new function.
3037 (pass_early_local_passes): Replace declaration with that of...
3038 (make_pass_early_local_passes): ...new function.
3039 (pass_ipa_whole_program_visibility): Replace declaration with that
3040 of...
3041 (make_pass_ipa_whole_program_visibility): ...new function.
3042 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3043 (make_pass_ipa_lto_gimple_out): ...new function.
3044 (pass_ipa_increase_alignment): Replace declaration with that of...
3045 (make_pass_ipa_increase_alignment): ...new function.
3046 (pass_ipa_inline): Replace declaration with that of...
3047 (make_pass_ipa_inline): ...new function.
3048 (pass_ipa_free_lang_data): Replace declaration with that of...
3049 (make_pass_ipa_free_lang_data): ...new function.
3050 (pass_ipa_free_inline_summary): Replace declaration with that of...
3051 (make_pass_ipa_free_inline_summary): ...new function.
3052 (pass_ipa_cp): Replace declaration with that of...
3053 (make_pass_ipa_cp): ...new function.
3054 (pass_ipa_reference): Replace declaration with that of...
3055 (make_pass_ipa_reference): ...new function.
3056 (pass_ipa_pure_const): Replace declaration with that of...
3057 (make_pass_ipa_pure_const): ...new function.
3058 (pass_ipa_pta): Replace declaration with that of...
3059 (make_pass_ipa_pta): ...new function.
3060 (pass_ipa_lto_finish_out): Replace declaration with that of...
3061 (make_pass_ipa_lto_finish_out): ...new function.
3062 (pass_ipa_tm): Replace declaration with that of...
3063 (make_pass_ipa_tm): ...new function.
3064 (pass_ipa_profile): Replace declaration with that of...
3065 (make_pass_ipa_profile): ...new function.
3066 (pass_ipa_cdtor_merge): Replace declaration with that of...
3067 (make_pass_ipa_cdtor_merge): ...new function.
3068 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3069 of...
3070 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3071 (pass_init_datastructures): Replace declaration with that of...
3072 (make_pass_init_datastructures): ...new function.
3073 (pass_fixup_cfg): Replace declaration with that of...
3074 (make_pass_fixup_cfg): ...new function.
3075 (pass_expand): Replace declaration with that of...
3076 (make_pass_expand): ...new function.
3077 (pass_instantiate_virtual_regs): Replace declaration with that of...
3078 (make_pass_instantiate_virtual_regs): ...new function.
3079 (pass_rtl_fwprop): Replace declaration with that of...
3080 (make_pass_rtl_fwprop): ...new function.
3081 (pass_rtl_fwprop_addr): Replace declaration with that of...
3082 (make_pass_rtl_fwprop_addr): ...new function.
3083 (pass_jump): Replace declaration with that of...
3084 (make_pass_jump): ...new function.
3085 (pass_jump2): Replace declaration with that of...
3086 (make_pass_jump2): ...new function.
3087 (pass_lower_subreg): Replace declaration with that of...
3088 (make_pass_lower_subreg): ...new function.
3089 (pass_cse): Replace declaration with that of...
3090 (make_pass_cse): ...new function.
3091 (pass_fast_rtl_dce): Replace declaration with that of...
3092 (make_pass_fast_rtl_dce): ...new function.
3093 (pass_ud_rtl_dce): Replace declaration with that of...
3094 (make_pass_ud_rtl_dce): ...new function.
3095 (pass_rtl_dce): Replace declaration with that of...
3096 (make_pass_rtl_dce): ...new function.
3097 (pass_rtl_dse1): Replace declaration with that of...
3098 (make_pass_rtl_dse1): ...new function.
3099 (pass_rtl_dse2): Replace declaration with that of...
3100 (make_pass_rtl_dse2): ...new function.
3101 (pass_rtl_dse3): Replace declaration with that of...
3102 (make_pass_rtl_dse3): ...new function.
3103 (pass_rtl_cprop): Replace declaration with that of...
3104 (make_pass_rtl_cprop): ...new function.
3105 (pass_rtl_pre): Replace declaration with that of...
3106 (make_pass_rtl_pre): ...new function.
3107 (pass_rtl_hoist): Replace declaration with that of...
3108 (make_pass_rtl_hoist): ...new function.
3109 (pass_rtl_store_motion): Replace declaration with that of...
3110 (make_pass_rtl_store_motion): ...new function.
3111 (pass_cse_after_global_opts): Replace declaration with that of...
3112 (make_pass_cse_after_global_opts): ...new function.
3113 (pass_rtl_ifcvt): Replace declaration with that of...
3114 (make_pass_rtl_ifcvt): ...new function.
3115 (pass_into_cfg_layout_mode): Replace declaration with that of...
3116 (make_pass_into_cfg_layout_mode): ...new function.
3117 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3118 (make_pass_outof_cfg_layout_mode): ...new function.
3119 (pass_loop2): Replace declaration with that of...
3120 (make_pass_loop2): ...new function.
3121 (pass_rtl_loop_init): Replace declaration with that of...
3122 (make_pass_rtl_loop_init): ...new function.
3123 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3124 (make_pass_rtl_move_loop_invariants): ...new function.
3125 (pass_rtl_unswitch): Replace declaration with that of...
3126 (make_pass_rtl_unswitch): ...new function.
3127 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3128 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3129 (pass_rtl_doloop): Replace declaration with that of...
3130 (make_pass_rtl_doloop): ...new function.
3131 (pass_rtl_loop_done): Replace declaration with that of...
3132 (make_pass_rtl_loop_done): ...new function.
3133 (pass_web): Replace declaration with that of...
3134 (make_pass_web): ...new function.
3135 (pass_cse2): Replace declaration with that of...
3136 (make_pass_cse2): ...new function.
3137 (pass_df_initialize_opt): Replace declaration with that of...
3138 (make_pass_df_initialize_opt): ...new function.
3139 (pass_df_initialize_no_opt): Replace declaration with that of...
3140 (make_pass_df_initialize_no_opt): ...new function.
3141 (pass_reginfo_init): Replace declaration with that of...
3142 (make_pass_reginfo_init): ...new function.
3143 (pass_inc_dec): Replace declaration with that of...
3144 (make_pass_inc_dec): ...new function.
3145 (pass_stack_ptr_mod): Replace declaration with that of...
3146 (make_pass_stack_ptr_mod): ...new function.
3147 (pass_initialize_regs): Replace declaration with that of...
3148 (make_pass_initialize_regs): ...new function.
3149 (pass_combine): Replace declaration with that of...
3150 (make_pass_combine): ...new function.
3151 (pass_if_after_combine): Replace declaration with that of...
3152 (make_pass_if_after_combine): ...new function.
3153 (pass_ree): Replace declaration with that of...
3154 (make_pass_ree): ...new function.
3155 (pass_partition_blocks): Replace declaration with that of...
3156 (make_pass_partition_blocks): ...new function.
3157 (pass_match_asm_constraints): Replace declaration with that of...
3158 (make_pass_match_asm_constraints): ...new function.
3159 (pass_regmove): Replace declaration with that of...
3160 (make_pass_regmove): ...new function.
3161 (pass_split_all_insns): Replace declaration with that of...
3162 (make_pass_split_all_insns): ...new function.
3163 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3164 (make_pass_fast_rtl_byte_dce): ...new function.
3165 (pass_lower_subreg2): Replace declaration with that of...
3166 (make_pass_lower_subreg2): ...new function.
3167 (pass_mode_switching): Replace declaration with that of...
3168 (make_pass_mode_switching): ...new function.
3169 (pass_sms): Replace declaration with that of...
3170 (make_pass_sms): ...new function.
3171 (pass_sched): Replace declaration with that of...
3172 (make_pass_sched): ...new function.
3173 (pass_ira): Replace declaration with that of...
3174 (make_pass_ira): ...new function.
3175 (pass_reload): Replace declaration with that of...
3176 (make_pass_reload): ...new function.
3177 (pass_clean_state): Replace declaration with that of...
3178 (make_pass_clean_state): ...new function.
3179 (pass_branch_prob): Replace declaration with that of...
3180 (make_pass_branch_prob): ...new function.
3181 (pass_value_profile_transformations): Replace declaration with that
3182 of...
3183 (make_pass_value_profile_transformations): ...new function.
3184 (pass_postreload_cse): Replace declaration with that of...
3185 (make_pass_postreload_cse): ...new function.
3186 (pass_gcse2): Replace declaration with that of...
3187 (make_pass_gcse2): ...new function.
3188 (pass_split_after_reload): Replace declaration with that of...
3189 (make_pass_split_after_reload): ...new function.
3190 (pass_branch_target_load_optimize1): Replace declaration with that
3191 of...
3192 (make_pass_branch_target_load_optimize1): ...new function.
3193 (pass_thread_prologue_and_epilogue): Replace declaration with that
3194 of...
3195 (make_pass_thread_prologue_and_epilogue): ...new function.
3196 (pass_stack_adjustments): Replace declaration with that of...
3197 (make_pass_stack_adjustments): ...new function.
3198 (pass_peephole2): Replace declaration with that of...
3199 (make_pass_peephole2): ...new function.
3200 (pass_if_after_reload): Replace declaration with that of...
3201 (make_pass_if_after_reload): ...new function.
3202 (pass_regrename): Replace declaration with that of...
3203 (make_pass_regrename): ...new function.
3204 (pass_cprop_hardreg): Replace declaration with that of...
3205 (make_pass_cprop_hardreg): ...new function.
3206 (pass_reorder_blocks): Replace declaration with that of...
3207 (make_pass_reorder_blocks): ...new function.
3208 (pass_branch_target_load_optimize2): Replace declaration with that
3209 of...
3210 (make_pass_branch_target_load_optimize2): ...new function.
3211 (pass_leaf_regs): Replace declaration with that of...
3212 (make_pass_leaf_regs): ...new function.
3213 (pass_split_before_sched2): Replace declaration with that of...
3214 (make_pass_split_before_sched2): ...new function.
3215 (pass_compare_elim_after_reload): Replace declaration with that of...
3216 (make_pass_compare_elim_after_reload): ...new function.
3217 (pass_sched2): Replace declaration with that of...
3218 (make_pass_sched2): ...new function.
3219 (pass_stack_regs): Replace declaration with that of...
3220 (make_pass_stack_regs): ...new function.
3221 (pass_stack_regs_run): Replace declaration with that of...
3222 (make_pass_stack_regs_run): ...new function.
3223 (pass_df_finish): Replace declaration with that of...
3224 (make_pass_df_finish): ...new function.
3225 (pass_compute_alignments): Replace declaration with that of...
3226 (make_pass_compute_alignments): ...new function.
3227 (pass_duplicate_computed_gotos): Replace declaration with that of...
3228 (make_pass_duplicate_computed_gotos): ...new function.
3229 (pass_variable_tracking): Replace declaration with that of...
3230 (make_pass_variable_tracking): ...new function.
3231 (pass_free_cfg): Replace declaration with that of...
3232 (make_pass_free_cfg): ...new function.
3233 (pass_machine_reorg): Replace declaration with that of...
3234 (make_pass_machine_reorg): ...new function.
3235 (pass_cleanup_barriers): Replace declaration with that of...
3236 (make_pass_cleanup_barriers): ...new function.
3237 (pass_delay_slots): Replace declaration with that of...
3238 (make_pass_delay_slots): ...new function.
3239 (pass_split_for_shorten_branches): Replace declaration with that of...
3240 (make_pass_split_for_shorten_branches): ...new function.
3241 (pass_split_before_regstack): Replace declaration with that of...
3242 (make_pass_split_before_regstack): ...new function.
3243 (pass_convert_to_eh_region_ranges): Replace declaration with that
3244 of...
3245 (make_pass_convert_to_eh_region_ranges): ...new function.
3246 (pass_shorten_branches): Replace declaration with that of...
3247 (make_pass_shorten_branches): ...new function.
3248 (pass_set_nothrow_function_flags): Replace declaration with that of...
3249 (make_pass_set_nothrow_function_flags): ...new function.
3250 (pass_dwarf2_frame): Replace declaration with that of...
3251 (make_pass_dwarf2_frame): ...new function.
3252 (pass_final): Replace declaration with that of...
3253 (make_pass_final): ...new function.
3254 (pass_rtl_seqabstr): Replace declaration with that of...
3255 (make_pass_rtl_seqabstr): ...new function.
3256 (pass_release_ssa_names): Replace declaration with that of...
3257 (make_pass_release_ssa_names): ...new function.
3258 (pass_early_inline): Replace declaration with that of...
3259 (make_pass_early_inline): ...new function.
3260 (pass_inline_parameters): Replace declaration with that of...
3261 (make_pass_inline_parameters): ...new function.
3262 (pass_update_address_taken): Replace declaration with that of...
3263 (make_pass_update_address_taken): ...new function.
3264 (pass_convert_switch): Replace declaration with that of...
3265 (make_pass_convert_switch): ...new function.
3266 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
3267 to a subclass of simple_ipa_opt_pass along with...
3268 (pass_data_ipa_tree_profile): ...new pass_data instance and...
3269 (make_pass_ipa_tree_profile): ...new function.
3270 * tree-sra.c (pass_sra_early): Convert from a global struct to a
3271 subclass of gimple_opt_pass along with...
3272 (pass_data_sra_early): ...new pass_data instance and...
3273 (make_pass_sra_early): ...new function.
3274 (pass_sra): Convert from a global struct to a subclass of
3275 gimple_opt_pass along with...
3276 (pass_data_sra): ...new pass_data instance and...
3277 (make_pass_sra): ...new function.
3278 (pass_early_ipa_sra): Convert from a global struct to a subclass of
3279 gimple_opt_pass along with...
3280 (pass_data_early_ipa_sra): ...new pass_data instance and...
3281 (make_pass_early_ipa_sra): ...new function.
3282 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
3283 subclass of gimple_opt_pass along with...
3284 (pass_data_ccp): ...new pass_data instance and...
3285 (make_pass_ccp): ...new function.
3286 (pass_fold_builtins): Convert from a global struct to a subclass of
3287 gimple_opt_pass along with...
3288 (pass_data_fold_builtins): ...new pass_data instance and...
3289 (make_pass_fold_builtins): ...new function.
3290 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
3291 subclass of gimple_opt_pass along with...
3292 (pass_data_copy_prop): ...new pass_data instance and...
3293 (make_pass_copy_prop): ...new function.
3294 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
3295 global struct to a subclass of gimple_opt_pass along with...
3296 (pass_data_rename_ssa_copies): ...new pass_data instance and...
3297 (make_pass_rename_ssa_copies): ...new function.
3298 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
3299 subclass of gimple_opt_pass along with...
3300 (pass_data_dce): ...new pass_data instance and...
3301 (make_pass_dce): ...new function.
3302 (pass_dce_loop): Convert from a global struct to a subclass of
3303 gimple_opt_pass along with...
3304 (pass_data_dce_loop): ...new pass_data instance and...
3305 (make_pass_dce_loop): ...new function.
3306 (pass_cd_dce): Convert from a global struct to a subclass of
3307 gimple_opt_pass along with...
3308 (pass_data_cd_dce): ...new pass_data instance and...
3309 (make_pass_cd_dce): ...new function.
3310 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
3311 subclass of gimple_opt_pass along with...
3312 (pass_data_dominator): ...new pass_data instance and...
3313 (make_pass_dominator): ...new function.
3314 (pass_phi_only_cprop): Convert from a global struct to a subclass of
3315 gimple_opt_pass along with...
3316 (pass_data_phi_only_cprop): ...new pass_data instance and...
3317 (make_pass_phi_only_cprop): ...new function.
3318 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
3319 subclass of gimple_opt_pass along with...
3320 (pass_data_dse): ...new pass_data instance and...
3321 (make_pass_dse): ...new function.
3322 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
3323 a subclass of gimple_opt_pass along with...
3324 (pass_data_forwprop): ...new pass_data instance and...
3325 (make_pass_forwprop): ...new function.
3326 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
3327 struct to a subclass of gimple_opt_pass along with...
3328 (pass_data_tree_ifcombine): ...new pass_data instance and...
3329 (make_pass_tree_ifcombine): ...new function.
3330 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
3331 subclass of gimple_opt_pass along with...
3332 (pass_data_ch): ...new pass_data instance and...
3333 (make_pass_ch): ...new function.
3334 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
3335 subclass of gimple_opt_pass along with...
3336 (pass_data_tree_loop): ...new pass_data instance and...
3337 (make_pass_tree_loop): ...new function.
3338 (pass_tree_loop_init): Convert from a global struct to a subclass of
3339 gimple_opt_pass along with...
3340 (pass_data_tree_loop_init): ...new pass_data instance and...
3341 (make_pass_tree_loop_init): ...new function.
3342 (pass_lim): Convert from a global struct to a subclass of
3343 gimple_opt_pass along with...
3344 (pass_data_lim): ...new pass_data instance and...
3345 (make_pass_lim): ...new function.
3346 (pass_tree_unswitch): Convert from a global struct to a subclass of
3347 gimple_opt_pass along with...
3348 (pass_data_tree_unswitch): ...new pass_data instance and...
3349 (make_pass_tree_unswitch): ...new function.
3350 (pass_predcom): Convert from a global struct to a subclass of
3351 gimple_opt_pass along with...
3352 (pass_data_predcom): ...new pass_data instance and...
3353 (make_pass_predcom): ...new function.
3354 (pass_vectorize): Convert from a global struct to a subclass of
3355 gimple_opt_pass along with...
3356 (pass_data_vectorize): ...new pass_data instance and...
3357 (make_pass_vectorize): ...new function.
3358 (pass_graphite): Convert from a global struct to a subclass of
3359 gimple_opt_pass along with...
3360 (pass_data_graphite): ...new pass_data instance and...
3361 (make_pass_graphite): ...new function.
3362 (pass_graphite_transforms): Convert from a global struct to a subclass
3363 of gimple_opt_pass along with...
3364 (pass_data_graphite_transforms): ...new pass_data instance and...
3365 (make_pass_graphite_transforms): ...new function.
3366 (pass_check_data_deps): Convert from a global struct to a subclass of
3367 gimple_opt_pass along with...
3368 (pass_data_check_data_deps): ...new pass_data instance and...
3369 (make_pass_check_data_deps): ...new function.
3370 (pass_iv_canon): Convert from a global struct to a subclass of
3371 gimple_opt_pass along with...
3372 (pass_data_iv_canon): ...new pass_data instance and...
3373 (make_pass_iv_canon): ...new function.
3374 (pass_scev_cprop): Convert from a global struct to a subclass of
3375 gimple_opt_pass along with...
3376 (pass_data_scev_cprop): ...new pass_data instance and...
3377 (make_pass_scev_cprop): ...new function.
3378 (pass_record_bounds): Convert from a global struct to a subclass of
3379 gimple_opt_pass along with...
3380 (pass_data_record_bounds): ...new pass_data instance and...
3381 (make_pass_record_bounds): ...new function.
3382 (pass_complete_unroll): Convert from a global struct to a subclass of
3383 gimple_opt_pass along with...
3384 (pass_data_complete_unroll): ...new pass_data instance and...
3385 (make_pass_complete_unroll): ...new function.
3386 (pass_complete_unrolli): Convert from a global struct to a subclass of
3387 gimple_opt_pass along with...
3388 (pass_data_complete_unrolli): ...new pass_data instance and...
3389 (make_pass_complete_unrolli): ...new function.
3390 (pass_parallelize_loops): Convert from a global struct to a subclass
3391 of gimple_opt_pass along with...
3392 (pass_data_parallelize_loops): ...new pass_data instance and...
3393 (make_pass_parallelize_loops): ...new function.
3394 (pass_loop_prefetch): Convert from a global struct to a subclass of
3395 gimple_opt_pass along with...
3396 (pass_data_loop_prefetch): ...new pass_data instance and...
3397 (make_pass_loop_prefetch): ...new function.
3398 (pass_iv_optimize): Convert from a global struct to a subclass of
3399 gimple_opt_pass along with...
3400 (pass_data_iv_optimize): ...new pass_data instance and...
3401 (make_pass_iv_optimize): ...new function.
3402 (pass_tree_loop_done): Convert from a global struct to a subclass of
3403 gimple_opt_pass along with...
3404 (pass_data_tree_loop_done): ...new pass_data instance and...
3405 (make_pass_tree_loop_done): ...new function.
3406 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
3407 struct to a subclass of gimple_opt_pass along with...
3408 (pass_data_cse_reciprocals): ...new pass_data instance and...
3409 (make_pass_cse_reciprocals): ...new function.
3410 (pass_cse_sincos): Convert from a global struct to a subclass of
3411 gimple_opt_pass along with...
3412 (pass_data_cse_sincos): ...new pass_data instance and...
3413 (make_pass_cse_sincos): ...new function.
3414 (pass_optimize_bswap): Convert from a global struct to a subclass of
3415 gimple_opt_pass along with...
3416 (pass_data_optimize_bswap): ...new pass_data instance and...
3417 (make_pass_optimize_bswap): ...new function.
3418 (pass_optimize_widening_mul): Convert from a global struct to a
3419 subclass of gimple_opt_pass along with...
3420 (pass_data_optimize_widening_mul): ...new pass_data instance and...
3421 (make_pass_optimize_widening_mul): ...new function.
3422 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
3423 subclass of gimple_opt_pass along with...
3424 (pass_data_phiopt): ...new pass_data instance and...
3425 (make_pass_phiopt): ...new function.
3426 (pass_cselim): Convert from a global struct to a subclass of
3427 gimple_opt_pass along with...
3428 (pass_data_cselim): ...new pass_data instance and...
3429 (make_pass_cselim): ...new function.
3430 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
3431 subclass of gimple_opt_pass along with...
3432 (pass_data_phiprop): ...new pass_data instance and...
3433 (make_pass_phiprop): ...new function.
3434 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
3435 subclass of gimple_opt_pass along with...
3436 (pass_data_pre): ...new pass_data instance and...
3437 (make_pass_pre): ...new function.
3438 (pass_fre): Convert from a global struct to a subclass of
3439 gimple_opt_pass along with...
3440 (pass_data_fre): ...new pass_data instance and...
3441 (make_pass_fre): ...new function.
3442 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
3443 subclass of gimple_opt_pass along with...
3444 (pass_data_reassoc): ...new pass_data instance and...
3445 (make_pass_reassoc): ...new function.
3446 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
3447 subclass of gimple_opt_pass along with...
3448 (pass_data_sink_code): ...new pass_data instance and...
3449 (make_pass_sink_code): ...new function.
3450 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
3451 subclass of gimple_opt_pass along with...
3452 (pass_data_strlen): ...new pass_data instance and...
3453 (make_pass_strlen): ...new function.
3454 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
3455 struct to a subclass of gimple_opt_pass along with...
3456 (pass_data_build_alias): ...new pass_data instance and...
3457 (make_pass_build_alias): ...new function.
3458 (pass_build_ealias): Convert from a global struct to a subclass of
3459 gimple_opt_pass along with...
3460 (pass_data_build_ealias): ...new pass_data instance and...
3461 (make_pass_build_ealias): ...new function.
3462 (pass_ipa_pta): Convert from a global struct to a subclass of
3463 simple_ipa_opt_pass along with...
3464 (pass_data_ipa_pta): ...new pass_data instance and...
3465 (make_pass_ipa_pta): ...new function.
3466 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
3467 subclass of gimple_opt_pass along with...
3468 (pass_data_uncprop): ...new pass_data instance and...
3469 (make_pass_uncprop): ...new function.
3470 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
3471 global struct to a subclass of gimple_opt_pass along with...
3472 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
3473 (make_pass_late_warn_uninitialized): ...new function.
3474 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
3475 to a subclass of gimple_opt_pass along with...
3476 (pass_data_init_datastructures): ...new pass_data instance and...
3477 (make_pass_init_datastructures): ...new function.
3478 (pass_early_warn_uninitialized): Convert from a global struct to a
3479 subclass of gimple_opt_pass along with...
3480 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
3481 (make_pass_early_warn_uninitialized): ...new function.
3482 (pass_update_address_taken): Convert from a global struct to a
3483 subclass of gimple_opt_pass along with...
3484 (pass_data_update_address_taken): ...new pass_data instance and...
3485 (make_pass_update_address_taken): ...new function.
3486 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
3487 struct to a subclass of gimple_opt_pass along with...
3488 (pass_data_release_ssa_names): ...new pass_data instance and...
3489 (make_pass_release_ssa_names): ...new function.
3490 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
3491 subclass of gimple_opt_pass along with...
3492 (pass_data_stdarg): ...new pass_data instance and...
3493 (make_pass_stdarg): ...new function.
3494 * tree-switch-conversion.c (pass_convert_switch): Convert from a
3495 global struct to a subclass of gimple_opt_pass along with...
3496 (pass_data_convert_switch): ...new pass_data instance and...
3497 (make_pass_convert_switch): ...new function.
3498 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
3499 to a subclass of gimple_opt_pass along with...
3500 (pass_data_tail_recursion): ...new pass_data instance and...
3501 (make_pass_tail_recursion): ...new function.
3502 (pass_tail_calls): Convert from a global struct to a subclass of
3503 gimple_opt_pass along with...
3504 (pass_data_tail_calls): ...new pass_data instance and...
3505 (make_pass_tail_calls): ...new function.
3506 * tree-vect-generic.c (pass_lower_vector): Convert from a global
3507 struct to a subclass of gimple_opt_pass along with...
3508 (pass_data_lower_vector): ...new pass_data instance and...
3509 (make_pass_lower_vector): ...new function.
3510 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
3511 gimple_opt_pass along with...
3512 (pass_data_lower_vector_ssa): ...new pass_data instance and...
3513 (make_pass_lower_vector_ssa): ...new function.
3514 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
3515 to a subclass of gimple_opt_pass along with...
3516 (pass_data_slp_vectorize): ...new pass_data instance and...
3517 (make_pass_slp_vectorize): ...new function.
3518 (pass_ipa_increase_alignment): Convert from a global struct to a
3519 subclass of simple_ipa_opt_pass along with...
3520 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
3521 (make_pass_ipa_increase_alignment): ...new function.
3522 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
3523 gimple_opt_pass along with...
3524 (pass_data_vrp): ...new pass_data instance and...
3525 (make_pass_vrp): ...new function.
3526 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
3527 subclass of simple_ipa_opt_pass along with...
3528 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
3529 (make_pass_ipa_free_lang_data): ...new function.
3530 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
3531 gimple_opt_pass along with...
3532 (pass_data_tsan): ...new pass_data instance and...
3533 (make_pass_tsan): ...new function.
3534 (pass_tsan_O0): Convert from a global struct to a subclass of
3535 gimple_opt_pass along with...
3536 (pass_data_tsan_O0): ...new pass_data instance and...
3537 (make_pass_tsan_O0): ...new function.
3538 * var-tracking.c (pass_variable_tracking): Convert from a global
3539 struct to a subclass of rtl_opt_pass along with...
3540 (pass_data_variable_tracking): ...new pass_data instance and...
3541 (make_pass_variable_tracking): ...new function.
3542 * web.c (pass_web): Convert from a global struct to a subclass of
3543 rtl_opt_pass along with...
3544 (pass_data_web): ...new pass_data instance and...
3545 (make_pass_web): ...new function.
3546 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
3547 declaration with that of...
3548 (make_pass_mode_switch_use): ...new function.
3549 (pass_resolve_sw_modes): Replace declaration with that of...
3550 (make_pass_resolve_sw_modes): ...new function.
3551 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
3552 from a global struct to a subclass of rtl_opt_pass along with...
3553 (pass_data_mode_switch_use): ...new pass_data instance and...
3554 (make_pass_mode_switch_use): ...new function.
3555 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
3556 from a global struct to a subclass of rtl_opt_pass along with...
3557 (pass_data_resolve_sw_modes): ...new pass_data instance and...
3558 (make_pass_resolve_sw_modes): ...new function.
3559 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
3560 struct to a subclass of rtl_opt_pass along with...
3561 (pass_data_insert_vzeroupper): ...new pass_data instance and...
3562 (make_pass_insert_vzeroupper): ...new function.
3563 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
3564 global struct to a subclass of rtl_opt_pass along with...
3565 (pass_data_work_around_errata): ...new pass_data instance and...
3566 (make_pass_work_around_errata): ...new function.
3567 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
3568 struct to a subclass of rtl_opt_pass along with...
3569 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
3570 (make_pass_mips_machine_reorg2): ...new function.
3571
3572 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3573
3574 * passes.c (pass_manager::operator new): New.
3575
3576 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3577
3578 Handwritten part of conversion of passes to C++ classes.
3579
3580 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
3581 (toplev.o): Add dep on PASS_MANAGER_H.
3582 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
3583 of early local pases to reflect this moving from a global to a
3584 member of gcc::pass_manager.
3585 (cgraph_add_new_function): Likewise.
3586 * lto-cgraph.c (lto_output_node): Update for conversion of
3587 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
3588 * passes.c (opt_pass::clone): New.
3589 (opt_pass::gate): New.
3590 (opt_pass::execute): New.
3591 (opt_pass::opt_pass): New.
3592 (pass_manager::execute_early_local_passes): New.
3593 (pass_manager::execute_pass_mode_switching): new.
3594 (finish_optimization_passes): Convert to...
3595 (pass_manager::finish_optimization_passes): ...this.
3596 (finish_optimization_passes): Update for conversion of passes to
3597 C++ classes.
3598 (register_dump_files_1): Use has_gate since we cannot portably
3599 check a vtable entry against NULL.
3600 (dump_one_pass): Likewise.
3601 (ipa_write_summaries_2): Likewise.
3602 (ipa_write_optimization_summaries_1): Likewise.
3603 (ipa_read_summaries_1): Likewise.
3604 (ipa_read_optimization_summaries_1): Likewise.
3605 (execute_ipa_stmt_fixups): Likewise.
3606 (pass_manager::pass_manager): Rewrite pass-creation, invoking
3607 pass-creation functions rather than wiring up globals, and
3608 storing the results in fields of pass_manager generated using
3609 pass-instances.def.
3610 (pass_manager::dump_profile_report): Update for conversion of
3611 passes to C++ classes.
3612 (pass_manager::execute_ipa_summary_passes): Likewise.
3613 (execute_one_ipa_transform_pass): Likewise.
3614 (execute_one_pass): Use has_gate and has_execute since we cannot
3615 portably check a vtable entry against NULL.
3616 * pass_manager.h (pass_manager::finish_optimization_passes): New.
3617 (pass_manager): Use pass-instances.def to add fields for the
3618 various pass instances.
3619 * toplev.c (finalize): Update for move of
3620 finish_optimization_passes to a method of gcc::pass_manager.
3621 * toplev.h (finish_optimization_passes): Move to method of class
3622 pass_manager.
3623 * tree-pass.h (struct pass_data): New.
3624 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
3625 (opt_pass::gate): Convert to virtual function.
3626 (opt_pass::~opt_pass): New.
3627 (opt_pass::clone): New.
3628 (opt_pass::execute): Convert to virtual function.
3629 (opt_pass::opt_pass): New.
3630 (opt_pass::ctxt_): new.
3631 (gimple_opt_pass): Convert to subclass of opt_pass.
3632 (gimple_opt_pass::gimple_opt_pass): New.
3633 (rtl_opt_pass): Convert to subclass of opt_pass.
3634 (rtl_opt_pass::rtl_opt_pass): New.
3635 (ipa_opt_pass_d): Convert to subclass of opt_pass.
3636 (ipa_opt_pass_d::ipa_opt_pass_d): New.
3637 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
3638 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
3639 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
3640 invocation of pass_mode_switching to reflect this moving from a
3641 global to a member of gcc::pass_manager.
3642 (ix86_option_override): Rework how pass_insert_vzeroupper is
3643 added to the pass_manager to reflect autogenerated changes.
3644 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
3645
3646 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
3647
3648 PR rtl-optimization/57708
3649 * recog.c (peep2_find_free_register): Validate all regs in a
3650 multi-reg mode.
3651
3652 2013-08-05 Jan Hubicka <jh@suse.cz>
3653
3654 PR lto/57602
3655 * cgraph.c (verify_cgraph_node): Accept local flags from other
3656 partitions.
3657 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
3658 (function_and_variable_visibility): Likewise.
3659 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
3660
3661 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3662
3663 * graph.c (init_graph_slim_pretty_print): Remove.
3664 (print_graph_cfg): Do not call it. Use local pretty printer.
3665 (start_graph_dump): Likewise.
3666
3667 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3668
3669 * gimple-pretty-print.c (buffer): Remove.
3670 (initialized): Likewise.
3671 (maybe_init_pretty_print): Likewise.
3672 (print_gimple_stmt): Do not call it. Use non-static local
3673 pretty_printer variable.
3674 (print_gimple_expr): Likewise.
3675 (print_gimple_seq): Likewise.
3676 (gimple_dump_bb): Likewise.
3677
3678 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3679
3680 * asan.c (asan_pp): Remove.
3681 (asan_pp_initialized): Likewise.
3682 (asan_pp_initialize): Likewise.
3683 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
3684 (asan_emit_stack_protection): Tidy. Use local pretty printer.
3685 (asan_add_global): Likewise.
3686
3687 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
3688
3689 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
3690 * diagnostic.h (diagnostic_flush_buffer): Adjust.
3691 * pretty-print.c (pp_formatted_text_data): Likewise.
3692 (pp_indent): Rename from pp_base_indent.
3693 (pp_format): Rename from pp_base_format.
3694 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
3695 (pp_format_verbatim): Rename from pp_base_format_verbatim.
3696 (pp_flush): Rename from pp_base_flush.
3697 (pp_set_line_maximum_length): Rename from
3698 pp_base_set_line_maximum_length.
3699 (pp_clear_output_area): Rename from pp_base_clear_output_area.
3700 (pp_set_prefix): Rename from pp_base_set_prefix.
3701 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
3702 (pp_emit_prefix): Rename from pp_base_emit_prefix.
3703 (pp_append_text): Rename from pp_base_append_text.
3704 (pp_formatted_text): Rename from pp_base_formatted_text.
3705 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
3706 (pp_remaining_character_count_for_line): Rename from
3707 pp_base_remaining_character_count_for_line.
3708 (pp_newline): Rename from pp_base_newline.
3709 (pp_character): Rename from pp_base_character.
3710 (pp_string): Rename from pp_base_string.
3711 (pp_maybe_space): Rename from pp_base_maybe_space.
3712 * asan.c (asan_pp_string): Adjust.
3713 (asan_emit_stack_protection): Likewise.
3714 (asan_add_global): Likewise.
3715 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
3716 * tree-mudflap.c (mf_varname_tree): Likewise.
3717 * tree-pretty-print.c (pp_tree_identifier): Rename from
3718 pp_base_tree_identifier.
3719 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
3720 Declare as function.
3721
3722 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3723
3724 * pretty-print.h (pp_bar_bar): New.
3725 (pp_ampersand_ampersand): Likewise.
3726 (pp_less_equal): Likewise.
3727 (pp_greater_equal): Likewise.
3728 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
3729 printer functions instead of pp_string or operators and punctuators.
3730 (dump_gimple_call): Likewise.
3731 (dump_gimple_omp_for): Likewise.
3732 (dump_gimple_transaction): Likewise.
3733 (dump_gimple_phi): Likewise.
3734 (pp_gimple_stmt_1): Likewise.
3735 * sched-vis.c (print_insn): Likewise.
3736 * tree-mudflap.c (mf_varname_tree): Likewise.
3737 * tree-pretty-print.c (dump_block_node): Likewise.
3738 (dump_generic_node): Likewise.
3739
3740 2013-08-02 Jan Hubicka <jh@suse.cz>
3741
3742 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
3743 boundaries.
3744 * lto-streamer-out.c (tree_is_indexable): Results decls and
3745 parm decls are not indexable.
3746 (DFS_write_tree_body): Do not follow args and results.
3747 (hash_tree): Likewise.
3748 (output_functions): Rearrange so struct function is needed
3749 only when real body is output; be able to also ouptut abstract
3750 functions; output DECL_ARGUMENTS and DECL_RESULT.
3751 (lto_output): When not in WPA, ale store abstract functions.
3752 (write_symbol): Do not care about RESULT_DECL.
3753 (output_symbol_p): Handle correctly sbtract decls.
3754 * lto-streamer-in.c (input_function): Rearrange so struct
3755 function can be NULL at entry; allow streaming of
3756 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
3757 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
3758 sanity check during LTO.
3759 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
3760 RESULT_DECl and DECL_ARGUMENTS.
3761 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
3762 Likewise.
3763
3764 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3765
3766 * pretty-print.h (pp_underscore): New.
3767 (pp_comma): Tidy.
3768 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
3769 printer functions instead of pp_character.
3770 (dump_binary_rhs): Likewise.
3771 (dump_ternary_rhs): Likewise.
3772 (dump_gimple_call_args): Likewise.
3773 (pp_points_to_solution): Likewise.
3774 (dump_gimple_call): Likewise.
3775 (dump_gimple_switch): Likewise.
3776 (dump_gimple_cond): Likewise.
3777 (dump_gimple_bind): Likewise.
3778 (dump_gimple_try): Likewise.
3779 (dump_gimple_omp_for): Likewise.
3780 (dump_gimple_omp_continue): Likewise.
3781 (dump_gimple_omp_single): Likewise.
3782 (dump_gimple_omp_sections): Likewise.
3783 (dump_gimple_omp_block): Likewise.
3784 (dump_gimple_omp_critical): Likewise.
3785 (dump_gimple_transaction): Likewise.
3786 (dump_gimple_asm): Likewise.
3787 (dump_gimple_phi): Likewise.
3788 (dump_gimple_omp_parallel): Likewise.
3789 (dump_gimple_omp_task): Likewise.
3790 (dump_gimple_omp_atomic_load): Likewise.
3791 (dump_gimple_omp_atomic_store): Likewise.
3792 (dump_gimple_mem_ops): Likewise.
3793 (pp_gimple_stmt_1): Likewise.
3794 (pp_cfg_jump): Likewise.
3795 (dump_implicit_edges): Likewise.
3796 (gimple_dump_bb_for_graph): Likewise.
3797 * graph.c (draw_cfg_node): Likewise.
3798 * langhooks.c (lhd_print_error_function): Likewise.
3799 * sched-vis.c (print_exp): Likewise.
3800 (print_value): Likewise.
3801 (print_pattern): Likewise.
3802 (print_insn): Likewise.
3803 (rtl_dump_bb_for_graph): Likewise.
3804 * tree-pretty-print.c (dump_function_declaration): Likewise.
3805 (dump_array_domain): Likewise.
3806 (dump_omp_clause): Likewise.
3807 (dump_location): Likewise.
3808 (dump_generic_node): Likewise.
3809 (print_struct_decl): Likewise.
3810 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3811
3812 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3813
3814 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3815 candidate table when replacing a candidate statement.
3816 (replace_rhs_if_not_dup): Likewise.
3817 (replace_one_candidate): Likewise.
3818
3819 2013-08-02 Jan Hubicka <jh@suse.cz>
3820 Martin Liska <marxin.liska@gmail.com>
3821
3822 * cgraphunit.c (add_new_function): Fix logic when adding from
3823 late IPA pass.
3824 (assemble_thunk): Rename to ...
3825 (expand_thunk); .. this one; export; get it working with
3826 general functions; make produced gimple valid.
3827 * cgraph.h (expand_thunk): Declare.
3828
3829 2013-08-02 Jan Hubicka <jh@suse.cz>
3830
3831 * ipa-cp.c (gather_context_independent_values): Use
3832 ipa_get_param_move_cost.
3833 (get_replacement_map): Remove PARAM; move parameter folding
3834 into tree-inline.c
3835 (create_specialized_node): Update.
3836 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3837 assert that we have gimple body; update move_cost.
3838 (count_formal_params): Assert that we have gimple body.
3839 (ipa_dump_param): New function.
3840 (ipa_alloc_node_params): Break out from ...
3841 (ipa_initialize_node_params): ... here.
3842 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3843 (ipa_write_node_info): Stream move costs.
3844 (ipa_read_node_info): Read move costs.
3845 (ipa_update_after_lto_read): Do not recompute node params.
3846 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3847 (ipa_get_param): Check we are not in WPA.
3848 (ipa_get_param_move_cost): New.
3849 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3850 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3851 parm numbers to be present.
3852
3853 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3854
3855 PR rtl-optimization/58048
3856 * lra-constraints.c (process_alt_operands): Don't check asm
3857 operand on register.
3858
3859 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3860
3861 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3862 the implied StoreLoad barrier for atomic operations if before.
3863
3864 2013-08-02 Jan Hubicka <jh@suse.cz>
3865 Martin Liska <marxin.liska@gmail.com>
3866
3867 * cgraph.c (cgraph_function_body_availability): Do not check
3868 cgraph flags.
3869 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3870 symtab_node_availability): Declare.
3871 * ipa.c (can_replace_by_local_alias): New.
3872 (function_and_variable_visibility): Use it.
3873 * symtab.c (symtab_for_node_and_aliases,
3874 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3875
3876 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3877
3878 PR rtl-optimization/57963
3879 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3880 (lra_constraints): Use them.
3881
3882 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3883
3884 * config/arm/types.md (define_attr "type"): Add "load_acq"
3885 and "store_rel".
3886 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3887 changes.
3888 (cortex_a53_store1): Likewise.
3889
3890 2013-08-01 Jan Hubicka <jh@suse.cz>
3891
3892 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3893 partitions are not needed.
3894
3895 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3896
3897 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3898 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3899 MAYBE_NON_Q_CLASS_P where appropriate.
3900
3901 2013-08-01 Jan Hubicka <jh@suse.cz>
3902
3903 * cgraph.h (release_function_body): Declare.
3904 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3905 of unused delcarations.
3906
3907 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3908
3909 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3910 RTL form when subtracting a constant.
3911
3912 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3913
3914 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3915 Generate canonical plus rtx with negated immediate instead of minus
3916 where appropriate.
3917 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3918
3919 2013-08-01 Jan Hubicka <jh@suse.cz>
3920
3921 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3922 (cgraph_release_function_body): Likewise.
3923 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3924 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3925 to used_as_abstract_origin.
3926 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3927 symbols used as abstract origins.
3928 * cgraphunit.c (analyze_functions): Update.
3929 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3930 used_as_abstract_origin.
3931 * tree-inline.c (tree_function_versioning): Update
3932 used_as_abstract_origin; be ready for DECL_RESULT and
3933 DECL_ARGUMENTS to be NULL.
3934
3935 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3936 for abstract functions.
3937 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3938 real symbols.
3939
3940 2013-08-01 Jan Hubicka <jh@suse.cz>
3941
3942 * profile.c (compute_value_histograms): Fix thinko.
3943
3944 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3945
3946 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3947 aarch-common-protos.h to extra_headers.
3948 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3949 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3950 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3951
3952 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3953
3954 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3955 Include "../arm/types.md". Define "type" attribute for all patterns.
3956 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3957 attribute changes.
3958
3959 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3960
3961 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3962 to support power8 load fusion.
3963 (fusion_gpr_mem_load): Likewise.
3964
3965 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3966
3967 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3968 declarations for power8 load fusion.
3969 (emit_fusion_gpr_load): Likewise.
3970
3971 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3972 tuning for power8, turn on fusion mode by default. Turn on sign
3973 extending fusion mode if normal fusion mode is on, and we are at
3974 -O2 or -O3.
3975 (fusion_gpr_load_p): New function, return true if we can fuse an
3976 addis instruction with a dependent load to a GPR.
3977 (emit_fusion_gpr_load): Emit the instructions for power8 load
3978 fusion to GPRs.
3979
3980 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3981 (VSX load fusion peepholes): New peepholes to fuse together an
3982 addi instruction with a VSX load instruction.
3983
3984 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3985 peepholes to fuse an addis instruction with a load to a GPR base
3986 register. If we are supporting sign extending fusions, convert
3987 sign extending loads to zero extending loads and add an explicit
3988 sign extension.
3989
3990 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3991
3992 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3993 aarch-common-protos.h to extra_headers.
3994 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3995 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3996 (arm_early_store_addr_dep): Likewise.
3997 (arm_no_early_alu_shift_dep): Likewise.
3998 (arm_no_early_alu_shift_value_dep): Likewise.
3999 (arm_no_early_mul_dep): Likewise.
4000 (arm_no_early_store_addr_dep): Likewise.
4001 (arm_mac_accumulator_is_mul_result): Likewise.
4002 (arm_mac_accumulator_is_result): Likewise.
4003 * config/arm/aarch-common.c: ... here. New file.
4004 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
4005 here to ...
4006 (arm_early_store_addr_dep): Likewise.
4007 (arm_no_early_alu_shift_dep): Likewise.
4008 (arm_no_early_alu_shift_value_dep): Likewise.
4009 (arm_no_early_mul_dep): Likewise.
4010 (arm_no_early_store_addr_dep): Likewise.
4011 (arm_mac_accumulator_is_mul_result): Likewise.
4012 (arm_mac_accumulator_is_result): Likewise.
4013 * config/arm/aarch-common-protos.h: ... here. New file.
4014 * config/arm/t-arm (aarch-common.o): Define.
4015
4016 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
4017
4018 * config/arm/arm.md: Include new file "types.md".
4019 (define_attr "type"): Move from here to ...
4020 (define_attr "mul32"): Likewise.
4021 (define_attr "mul64"): Likewise.
4022 * config/arm/types.md: ... here. New file.
4023
4024 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4025
4026 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4027 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4028
4029 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4030
4031 * gen-pass-instances.awk: Fix offset of substr().
4032
4033 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4034
4035 * Makefile.in (pass-instances.def): New.
4036 (passes.o): Replace dependency on passes.def with one on
4037 pass-instances.def
4038
4039 * gen-pass-instances.awk: New.
4040
4041 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4042 rather than passes.def, updating local definition of NEXT_PASS
4043 macro to add an extra NUM parameter (currently unused).
4044
4045 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4046
4047 * Makefile.in (PASS_MANAGER_H): New.
4048 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4049 (passes.o): Likewise.
4050 (statistics.o): Likewise.
4051 (cgraphunit.o): Likewise.
4052 (context.o): Depend on PASS_MANAGER_H.
4053
4054 * pass_manager.h: New.
4055
4056 * cgraphunit.c (cgraph_add_new_function): Update for moves
4057 of globals to fields of pass_manager.
4058 (analyze_function): Likewise.
4059 (expand_function): Likewise.
4060 (ipa_passes): Likewise.
4061 (compile): Likewise.
4062
4063 * context.c (context::context): New.
4064 * context.h (context::context): New.
4065 (context::get_passes): New.
4066 (context::passes_): New.
4067
4068 * lto-cgraph.c (input_node): Update for moves of globals to
4069 fields of pass_manager.
4070
4071 * passes.c (all_passes): Remove, in favor of a field of the
4072 same name within the new class pass_manager.
4073 (all_small_ipa_passes): Likewise.
4074 (all_lowering_passes): Likewise.
4075 (all_regular_ipa_passes): Likewise.
4076 (all_late_ipa_passes): Likewise.
4077 (all_lto_gen_passes): Likewise.
4078 (passes_by_id): Likewise.
4079 (passes_by_id_size): Likewise.
4080 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4081 the new class pass_manager.
4082 (set_pass_for_id): Convert to...
4083 (pass_manager::set_pass_for_id): ...method.
4084 (get_pass_for_id): Convert to...
4085 (pass_manager::get_pass_for_id): ...method.
4086 (register_one_dump_file): Move body of implementation into...
4087 (pass_manager::register_one_dump_file): ...here.
4088 (register_dump_files_1): Convert to...
4089 (pass_manager::register_dump_files_1): ...method.
4090 (register_dump_files): Convert to...
4091 (pass_manager::register_dump_files): ...method.
4092 (create_pass_tab): Update for moves of globals to fields of
4093 pass_manager.
4094 (dump_passes): Move body of implementation into...
4095 (pass_manager::dump_passes): ...here.
4096 (register_pass): Move body of implementation into...
4097 (pass_manager::register_pass): ...here.
4098 (init_optimization_passes): Convert into...
4099 (pass_manager::pass_manager): ...constructor for new
4100 pass_manager class, and initialize the pass_lists array.
4101 (check_profile_consistency): Update for moves of globals to
4102 fields of pass_manager.
4103 (dump_profile_report): Move body of implementation into...
4104 (pass_manager::dump_profile_report): ...here.
4105 (ipa_write_summaries_1): Update for moves of pass lists from
4106 being globals to fields of pass_manager.
4107 (ipa_write_optimization_summaries): Likewise.
4108 (ipa_read_summaries): Likewise.
4109 (ipa_read_optimization_summaries): Likewise.
4110 (execute_all_ipa_stmt_fixups): Likewise.
4111
4112 * statistics.c (statistics_fini): Update for moves of globals to
4113 fields of pass_manager.
4114
4115 * toplev.c (general_init): Replace call to
4116 init_optimization_passes with construction of the pass_manager
4117 instance.
4118
4119 * tree-pass.h (all_passes): Remove, in favor of a field of the
4120 same name within the new class pass_manager.
4121 (all_small_ipa_passes): Likewise.
4122 (all_lowering_passes): Likewise.
4123 (all_regular_ipa_passes): Likewise.
4124 (all_lto_gen_passes): Likewise.
4125 (all_late_ipa_passes): Likewise.
4126 (passes_by_id): Likewise.
4127 (passes_by_id_size): Likewise.
4128 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4129 the new class pass_manager.
4130 (get_pass_for_id): Remove.
4131
4132 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4133
4134 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4135 configs.
4136
4137 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4138
4139 * arm.md (mulhi3): New expand pattern.
4140
4141 2013-07-30 Jan Hubicka <jh@suse.cz>
4142 Martin Liska <marxin.liska@gmail.com>
4143
4144 * profile.c (compute_value_histograms): Do not ICE when
4145 there is mismatch only on some counters.
4146
4147 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4148
4149 PR rtl-optimization/57637
4150 * function.c (move_insn_for_shrink_wrap): Also check the
4151 GEN set of the LIVE problem for the liveness analysis
4152 if it exists, otherwise give up.
4153
4154 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4155
4156 PR tree-optimization/57993
4157 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4158 replaced statement in the candidate table.
4159 (phi_add_costs): Return infinite cost when the hidden basis does
4160 not dominate all phis on which the candidate is dependent.
4161 (replace_one_candidate): Record replaced statement in the
4162 candidate table.
4163
4164 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4165
4166 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4167 (ashlv2si3): New expander.
4168 (*ashlv2si3_i): New define_insn_and_split.
4169 * predicates.md (float_operation): Allow patterns with three
4170 basic sub-patterns.
4171
4172 PR rtl-optimization/58021
4173 * mode-switching.c (create_pre_exit): Always split off preceding
4174 insns if we are not at the basic block head.
4175
4176 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4177
4178 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4179 (UCLIBC_DYNAMIC_LINKER): New macro.
4180 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4181 `-mnan=2008'.
4182 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4183 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4184 `-mnan=2008'.
4185 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4186 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4187 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4188 for SF and DF modes. Use ieee_quad_format for TF mode.
4189 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
4190 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
4191 (mips_option_override): Handle `-mnan=legacy'.
4192 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
4193 `-mabs=2008' and `-mnan=2008'.
4194 (OPTION_DEFAULT_SPECS): Add "nan" default.
4195 (ASM_SPEC): Handle `-mnan='.
4196 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
4197 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
4198 comment accordingly.
4199 (neg<mode>2): Likewise.
4200 * config/mips/mips.opt (mabs, mnan): New options.
4201 * doc/install.texi (Configuration): Document `--with-nan=' option.
4202 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
4203 `-mnan=' options.
4204 (MIPS Options): Document them.
4205 * config.gcc <mips*-*-*>: Handle `--with-nan='.
4206 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
4207 * configure: Regenerate.
4208 * config.in: Regenerate.
4209
4210 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4211
4212 * config/i386/i386.md (float post-reload splitters): Do not check
4213 for subregs of SSE registers.
4214
4215 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4216 H.J. Lu <hongjiu.lu@intel.com>
4217
4218 PR target/57954
4219 PR target/57988
4220 * config/i386/i386.md (post-reload splitter
4221 to avoid partial SSE reg dependency stalls): New pattern.
4222
4223 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
4224
4225 * config/s390/s390.md ("movcc"): Swap load and store instructions.
4226
4227 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
4228
4229 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
4230 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
4231
4232 2013-07-26 Cary Coutant <ccoutant@google.com>
4233
4234 * dwarf2out.c (die_checksum_ordered): Don't include template
4235 instantiations in signature.
4236 (is_template_parameter): New function.
4237 (is_template_instantiation): New function.
4238 (generate_skeleton_bottom_up): Don't include template instantiations
4239 in type unit DIE.
4240 (generate_skeleton): Likewise.
4241 (break_out_comdat_types): Move recursive call to break out nested
4242 types earlier.
4243 (prune_unused_types_mark_generic_parms_dies): Call
4244 is_template_parameter.
4245
4246 2013-07-26 Ian Bolton <ian.bolton@arm.com>
4247
4248 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
4249 uses vector registers.
4250 * config/aarch64/iterators.md: Add attributes rtn and vas.
4251
4252 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4253 Richard Earnshaw <richard.earnshaw@arm.com>
4254
4255 * combine.c (simplify_comparison): Re-canonicalize operands
4256 where appropriate.
4257 * config/arm/arm.md (movcond_addsi): New splitter.
4258
4259 2013-07-25 Sterling Augustine <saugustine@google.com>
4260
4261 * dwarf2out.c (size_of_pubnames): Move code to...
4262 (include_pubname_in_output): ...here. New.
4263 (want_pubnames): Rearrange.
4264 (output_pubnames): Call include_pubname_in_output. Move assertion.
4265
4266 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4267
4268 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
4269
4270 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4271
4272 PR target/38836
4273 * doc/extend.texi: Remove obsolete builtins. Fix
4274 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
4275
4276 2013-07-25 Jan Hubicka <jh@suse.cz>
4277
4278 * cgraph.c (release_function_body): Break out from ...
4279 (cgraph_release_function_body): ... this one; also release DECL_RESULT
4280 and DECL_ARGUMENTS.
4281 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
4282 old_tree in the map.
4283 (create_specialized_node): Update.
4284 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
4285 into index.
4286 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
4287 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
4288 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
4289 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
4290 DECL_RESULT.
4291
4292 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4293
4294 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
4295 addsi3_carryin_alt2_<optab>): Correct output template.
4296
4297 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4298
4299 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
4300 Adjust for arm_restrict_it.
4301 Remove trailing whitespace.
4302
4303 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
4304
4305 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
4306 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
4307
4308 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
4309
4310 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
4311
4312 PR rtl-optimization/57960
4313 * lra-constraints.c (process_alt_operands): Use the right mode
4314 when checking strict_low.
4315
4316 2013-07-25 Jan Hubicka <jh@suse.cz>
4317
4318 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
4319 * cgraph.c (cgraph_remove_node): Do not release function body
4320 when in cgraph streaming.
4321 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
4322 in other partitions are not considered reachable; fix handling of
4323 clones.
4324
4325 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4326
4327 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
4328
4329 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4330
4331 PR target/19599
4332 PR target/57731
4333 PR target/57837
4334 * config/arm/arm.md ("*sibcall_insn): Replace use of
4335 Ss with US. Adjust output for v5 and v4t.
4336 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
4337
4338 * config/arm/constraints.md ("Ss"): Rename to US.
4339
4340 2013-07-25 Terry Guo <terry.guo@arm.com>
4341
4342 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
4343 shift_add/shift_sub0/shift_sub1 RTXs.
4344
4345 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
4346 Anton Blanchard <anton@au1.ibm.com>
4347
4348 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
4349 (altivec_vpks<VI_char>ss): Likewise.
4350 (altivec_vpks<VI_char>us): Likewise.
4351 (altivec_vpku<VI_char>us): Likewise.
4352 (altivec_vpku<VI_char>um): Likewise.
4353
4354 2013-07-24 David Malcolm <dmalcolm@redhat.com>
4355
4356 Introduce context class.
4357
4358 * Makefile.in (CONTEXT_H): New.
4359 (OBJS): Add context.o.
4360 (toplev.o): Add CONTEXT_H to dependencies.
4361 (context.o): New.
4362
4363 * toplev.c (general_init): Create the singleton gcc::context instance.
4364
4365 * context.c: New.
4366
4367 * context.h: New.
4368
4369 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
4370
4371 PR rtl-optimization/57968
4372 * mode-switching.c (create_pre_exit): Allow instructions that
4373 don't set a return register to need a non-exit mode.
4374
4375 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4376 Anton Blanchard <anton@au1.ibm.com>
4377
4378 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
4379 operands to vperm for little endian.
4380 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
4381 of lvsl to create the control mask for a vperm for little endian.
4382
4383 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4384 Anton Blanchard <anton@au1.ibm.com>
4385
4386 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4387 two operands for little-endian.
4388
4389 2013-07-23 Steve Ellcey <sellcey@mips.com>
4390
4391 * config/mips/mips.c (mips_case_values_threshold): New.
4392 (TARGET_CASE_VALUES_THRESHOLD): Define.
4393
4394 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4395 Anton Blanchard <anton@au1.ibm.com>
4396
4397 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
4398 selection of field for vector splat in little endian mode.
4399
4400 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4401
4402 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
4403 expanders to rs6000.md.
4404 (ior<mode>3): Likewise.
4405 (and<mode>3): Likewise.
4406 (one_cmpl<mode>2): Likewise.
4407 (nor<mode>3): Likewise.
4408 (andc<mode>3): Likewise.
4409 (eqv<mode>3): Likewise.
4410 (nand<mode>3): Likewise.
4411 (orc<mode>3): Likewise.
4412
4413 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
4414 declaration.
4415
4416 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
4417 to split multi-word logical operations.
4418 (rs6000_split_logical_di): Likewise.
4419 (rs6000_split_logical): Likewise.
4420
4421 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
4422 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
4423 and allow TImode operations in 32-bit.
4424 (vsx_and<mode>3_64bit): Likewise.
4425 (vsx_ior<mode>3_32bit): Likewise.
4426 (vsx_ior<mode>3_64bit): Likewise.
4427 (vsx_xor<mode>3_32bit): Likewise.
4428 (vsx_xor<mode>3_64bit): Likewise.
4429 (vsx_one_cmpl<mode>2_32bit): Likewise.
4430 (vsx_one_cmpl<mode>2_64bit): Likewise.
4431 (vsx_nor<mode>3_32bit): Likewise.
4432 (vsx_nor<mode>3_64bit): Likewise.
4433 (vsx_andc<mode>3_32bit): Likewise.
4434 (vsx_andc<mode>3_64bit): Likewise.
4435 (vsx_eqv<mode>3_32bit): Likewise.
4436 (vsx_eqv<mode>3_64bit): Likewise.
4437 (vsx_nand<mode>3_32bit): Likewise.
4438 (vsx_nand<mode>3_64bit): Likewise.
4439 (vsx_orc<mode>3_32bit): Likewise.
4440 (vsx_orc<mode>3_64bit): Likewise.
4441
4442 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
4443 logical types in GPRs.
4444
4445 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
4446 logical insns to rs6000.md, and allow TImode operations in
4447 32-bit.
4448 (altivec_ior<mode>3): Likewise.
4449 (altivec_xor<mode>3): Likewise.
4450 (altivec_one_cmpl<mode>2): Likewise.
4451 (altivec_nor<mode>3): Likewise.
4452 (altivec_andc<mode>3): Likewise.
4453
4454 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
4455 attributes for moving the 128-bit logical operations into
4456 rs6000.md.
4457 (BOOL_REGS_OUTPUT): Likewise.
4458 (BOOL_REGS_OP1): Likewise.
4459 (BOOL_REGS_OP2): Likewise.
4460 (BOOL_REGS_UNARY): Likewise.
4461 (BOOL_REGS_AND_CR0): Likewise.
4462 (one_cmpl<mode>2): Add support for DI logical operations on
4463 32-bit, splitting the operations to 32-bit.
4464 (anddi3): Likewise.
4465 (iordi3): Likewise.
4466 (xordi3): Likewise.
4467 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
4468 changes to combine the 32/64-bit code, allow logical operations on
4469 TI mode in 32-bit, and to use similar match_operator patterns like
4470 scalar mode uses. Combine the Altivec and VSX code for logical
4471 operations, and move it here.
4472 (ior<mode>3, 128-bit types): Likewise.
4473 (xor<mode>3, 128-bit types): Likewise.
4474 (one_cmpl<mode>3, 128-bit types): Likewise.
4475 (nor<mode>3, 128-bit types): Likewise.
4476 (andc<mode>3, 128-bit types): Likewise.
4477 (eqv<mode>3, 128-bit types): Likewise.
4478 (nand<mode>3, 128-bit types): Likewise.
4479 (orc<mode>3, 128-bit types): Likewise.
4480 (and<mode>3_internal): Likewise.
4481 (bool<mode>3_internal): Likewise.
4482 (boolc<mode>3_internal1): Likewise.
4483 (boolc<mode>3_internal2): Likewise.
4484 (boolcc<mode>3_internal1): Likewise.
4485 (boolcc<mode>3_internal2): Likewise.
4486 (eqv<mode>3_internal1): Likewise.
4487 (eqv<mode>3_internal2): Likewise.
4488 (one_cmpl1<mode>3_internal): Likewise.
4489
4490 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4491
4492 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4493 Rename flag_stack_usage to flag_stack_usage_info.
4494
4495 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4496
4497 * config/microblaze/sync.md: New file.
4498 * config/microblaze/microblaze.md: Include sync.md
4499 * config/microblaze/microblaze.c: Add print_operand 'y'.
4500 * config/microblaze/constraints.md: Add memory_contraint
4501 'Q' which is a single register.
4502
4503 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
4504
4505 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
4506
4507 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4508
4509 * reload.c (find_reloads): Exit loop once we find this operand
4510 cannot be reloaded somehow for this alternative.
4511
4512 * reload.c (find_reloads): Exit loop once we find a hard register.
4513
4514 * rtlanal.c (computed_jump_p): Exit loop once we find label
4515 reference is used.
4516
4517 * i386.c (ix86_pad_returns): Exit loop after setting replace.
4518
4519 * cfgloopmanip.c (remove_path): Exit loop after setting
4520 irred_invalidated.
4521
4522 * gensupport.c (subst_dup): Avoid loop if code is not
4523 MATCH_DUP nor MATCH_OP_DUP.
4524
4525 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
4526
4527 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
4528
4529 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4530
4531 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
4532 true for SP_REGNUM if mode == ptr_mode.
4533 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
4534 with value R0_REGNUM + 31.
4535
4536 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4537
4538 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
4539 pad pointer-typed argument downward.
4540
4541 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4542
4543 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
4544 and __ILP32__ when the ILP32 model is in use.
4545
4546 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4547
4548 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
4549 (aarch64_load_symref_appropriately): In the case of
4550 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
4551 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
4552 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
4553 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
4554 if 'mode' doesn't equal to 'ptr_mode'.
4555 (aarch64_output_mi_thunk): Add an assertion on the alignment of
4556 'vcall_offset'; change to call aarch64_emit_move differently depending
4557 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
4558 to calculate the upper bound of 'vcall_offset'.
4559 (aarch64_cannot_force_const_mem): Change to also return true if
4560 mode != ptr_mode.
4561 (aarch64_legitimize_reload_address): In the case of large
4562 displacements, add new local variable 'xmode' and an assertion
4563 based on it; change to use 'xmode' to generate the new rtx and
4564 reload.
4565 (aarch64_asm_trampoline_template): Change to generate the template
4566 differently depending on TARGET_ILP32 or not; change to use
4567 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
4568 (aarch64_trampoline_size): Removed.
4569 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
4570 and replace immediate literals with it. Change to use 'ptr_mode'
4571 instead of 'DImode' and call convert_memory_address if the mode
4572 of 'fnaddr' doesn't equal to 'ptr_mode'.
4573 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
4574 to output symbol.
4575 (aarch64_elf_asm_destructor): Likewise.
4576 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
4577 on TARGET_ILP32 instead of aarch64_trampoline_size.
4578 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
4579 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
4580 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
4581 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
4582 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
4583 (storewb_pair<GPI:mode>_<P:mode>): ... this.
4584 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
4585 depending on the value of 'mode'.
4586 (add_losym_<mode>): New.
4587 (ldr_got_small_<mode>): New, based on ldr_got_small.
4588 (ldr_got_small): Remove.
4589 (ldr_got_small_sidi): New.
4590 * config/aarch64/iterators.md (P): New.
4591 (PTR): Change to 'ptr_mode' in the condition.
4592
4593 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4594
4595 * config.gcc (aarch64*-*-*): Support --with-abi.
4596 (aarch64*-*-elf): Support --with-multilib-list.
4597 (aarch64*-*-linux*): Likewise.
4598 (supported_defaults): Add abi to aarch64*-*-*.
4599 * configure.ac: Mention AArch64 for --with-multilib-list.
4600 * configure: Re-generated.
4601 * config/aarch64/biarchilp32.h: New file.
4602 * config/aarch64/biarchlp64.h: New file.
4603 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
4604 (ABI_SPEC): Ditto.
4605 (MULTILIB_DEFAULTS): Ditto.
4606 (DRIVER_SELF_SPECS): Ditto.
4607 (ASM_SPEC): Update to also substitute -mabi.
4608 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
4609 file whose name depends on -mabi= and -mbig-endian.
4610 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
4611 TARGET_ILP32.
4612 (POINTER_SIZE): New define.
4613 (POINTERS_EXTEND_UNSIGNED): Ditto.
4614 (enum aarch64_abi_type): New enumeration tag.
4615 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
4616 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
4617 (TARGET_ILP32): New define.
4618 * config/aarch64/aarch64.opt (mabi): New.
4619 (aarch64_abi): New.
4620 (ilp32, lp64): New values for -mabi.
4621 * config/aarch64/t-aarch64 (comma): New define.
4622 (MULTILIB_OPTIONS): Ditto.
4623 (MULTILIB_DIRNAMES): Ditto.
4624 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
4625 * doc/invoke.texi: Document -mabi for AArch64.
4626
4627 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
4628
4629 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
4630
4631 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4632 Anton Blanchard <anton@au1.ibm.com>
4633
4634 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
4635 endianness when selecting field to splat.
4636
4637 2013-07-22 Eric Christopher <echristo@gmail.com>
4638
4639 * dwarf2out.c (die_odr_checksum): New function to use
4640 CHECKSUM_ macros and ULEB128 for DIE tag.
4641 (generate_type_signature): Use.
4642
4643 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
4644
4645 * config.gcc (sparc*-*-*): Accept leon3 processor.
4646 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
4647 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
4648 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
4649 * config/sparc/sparc.opt (enum processor_type): Add leon3.
4650 (mfix-ut699): Adjust comment.
4651 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
4652 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
4653 (CPP_CPU_SPEC): Likewise.
4654 (ASM_CPU_SPEC): Likewise.
4655 * config/sparc/sparc.c (leon3_cost): New constant.
4656 (sparc_option_override): Add leon3 support.
4657 (mem_ref): New function.
4658 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
4659 (sparc_do_work_around_errata): Look into the instruction in the delay
4660 slot and adjust accordingly. Add fix for the data cache nullify issues
4661 of the UT699. Change insertion position for the NOP.
4662 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
4663 (leon3_load): New reservation.
4664 (leon_store): Bump latency to 2.
4665 (grfpu): New automaton.
4666 (grfpu_alu): New unit.
4667 (grfpu_ds): Likewise.
4668 (leon_fp_alu): Adjust.
4669 (leon_fp_mult): Delete.
4670 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
4671 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
4672 * config/sparc/sparc.md (cpu): Add leon3.
4673 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
4674 (swapsi): Likewise.
4675 (atomic_test_and_set): Likewise.
4676 (ldstub): Likewise.
4677
4678 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
4679
4680 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
4681 default for R5900 targets.
4682 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
4683 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
4684 * config/mips/mips.c (mips_option_override): Report an error for
4685 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
4686 for -march=r5900 -mhard-float.
4687
4688 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4689
4690 * df-problems.c (can_move_insns_across): Exit loop once we
4691 find a non-fixed, non-global register.
4692
4693 * ipa-pure-const.c (propagate_nothrow): Exit loop after
4694 setting can_throw.
4695
4696 * omega.c (omega_eliminate_red): Break after setting red_found.
4697 (omega_problem_has_red_equations): Similarly after setting found.
4698 (omega_query_variable): Similarly after setting coupled.
4699
4700 2013-07-22 Marek Polacek <polacek@redhat.com>
4701
4702 * gimplify.c: Don't include gimple.h twice.
4703
4704 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4705
4706 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
4707 instead of TARGET_THUMB1.
4708 (Pz): New constraint.
4709 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
4710 encodings.
4711 (compare_negsi_si): Likewise.
4712 (compare_addsi2_op0): Likewise.
4713 (compare_addsi2_op1): Likewise.
4714 (addsi3_carryin_<optab>): Likewise.
4715 (addsi3_carryin_alt2_<optab>): Likewise.
4716 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
4717 for arm_restrict_it.
4718 (subsi3_carryin): Likewise.
4719 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
4720 (minmax_arithsi): Disable for arm_restrict_it.
4721 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
4722 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
4723 (satsi_<SAT:code>_shift): Likewise.
4724 (arm_shiftsi3): Add alternative for 16-bit encoding.
4725 (arm32_movhf): Disable for arm_restrict_it.
4726 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
4727 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
4728
4729 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
4730
4731 * config/arm/arm.md (attribute "insn"): Delete.
4732 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
4733 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
4734 (not_shiftsi): Update for attribute change.
4735 (not_shiftsi_compare0): Likewise.
4736 (not_shiftsi_compare0_scratch): Likewise.
4737 (arm_one_cmplsi2): Likewise.
4738 (thumb1_one_cmplsi2): Likewise.
4739 (notsi_compare0): Likewise.
4740 (notsi_compare0_scratch): Likewise.
4741 (thumb1_movdi_insn): Likewise.
4742 (arm_movsi_insn): Likewise.
4743 (movhi_insn_arch4): Likewise.
4744 (movhi_bytes): Likewise.
4745 (arm_movqi_insn): Likewise.
4746 (thumb1_movqi_insn): Likewise.
4747 (arm32_movhf): Likewise.
4748 (thumb1_movhf): Likewise.
4749 (arm_movsf_soft_insn): Likewise.
4750 (thumb1_movsf_insn): Likewise.
4751 (thumb_movdf_insn): Likewise.
4752 (movsicc_insn): Likewise.
4753 (movsfcc_soft_insn): Likewise.
4754 (and_scc): Likewise.
4755 (cond_move): Likewise.
4756 (if_move_not): Likewise.
4757 (if_not_move): Likewise.
4758 (if_shift_move): Likewise.
4759 (if_move_shift): Likewise.
4760 (if_shift_shift): Likewise.
4761 (if_not_arith): Likewise.
4762 (if_arith_not): Likewise.
4763 (cond_move_not): Likewise.
4764 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
4765 (neon_mov<mode>): Likewise.
4766 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
4767 (thumb2_movsi_vfp): Likewise.
4768 (movsf_vfp): Likewise.
4769 (thumb2_movsf_vfp): Likewise.
4770 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4771 change.
4772 (cortexa7_older_only): Likewise.
4773 (cortexa7_younger): Likewise.
4774 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
4775 (1020alu_shift_op): Likewise.
4776 (1020alu_shift_reg_op): Likewise.
4777 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
4778 (alu_shift_op): Likewise.
4779 (alu_shift_reg_op): Likewise.
4780 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
4781 (11_alu_shift_op): Likewise.
4782 (11_alu_shift_reg_op): Likewise.
4783 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
4784 (9_alu_shift_reg_op): Likewise.
4785 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4786 change.
4787 (cortex_a15_alu_shift): Likewise.
4788 (cortex_a15_alu_shift_reg): Likewise.
4789 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
4790 (cortex_a5_alu_shift): Likewise.
4791 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
4792 change.
4793 (cortex_a53_alu_shift): Likewise.
4794 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4795 change.
4796 (cortex_a7_alu_reg): Likewise.
4797 (cortex_a7_alu_shift): Likewise.
4798 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4799 (cortex_a8_alu_shift): Likewise.
4800 (cortex_a8_alu_shift_reg): Likewise.
4801 (cortex_a8_mov): Likewise.
4802 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4803 (cortex_a9_dp_shift): Likewise.
4804 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4805 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4806 (cortex_r4_mov): Likewise.
4807 (cortex_r4_alu_shift): Likewise.
4808 (cortex_r4_alu_shift_reg): Likewise.
4809 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4810 (526_alu_shift_op): Likewise.
4811 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4812 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4813 (626te_alu_shift_op): Likewise.
4814 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4815 (726te_alu_op): Likewise.
4816 (726te_alu_shift_op): Likewise.
4817 (726te_alu_shift_reg_op): Likewise.
4818 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4819 (mp626_alu_shift_op): Likewise.
4820 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4821 (pj4_alu_e1_conds): Likewise.
4822 (pj4_alu): Likewise.
4823 (pj4_alu_conds): Likewise.
4824 (pj4_shift): Likewise.
4825 (pj4_shift_conds): Likewise.
4826 (pj4_alu_shift): Likewise.
4827 (pj4_alu_shift_conds): Likewise.
4828
4829 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4830
4831 * config/arm/predicates.md (shiftable_operator_strict_it):
4832 New predicate.
4833 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4834 Disable cond_exec version for arm_restrict_it.
4835 (thumb2_smaxsi3): Convert to generate cond_exec.
4836 (thumb2_sminsi3): Likewise.
4837 (thumb32_umaxsi3): Likewise.
4838 (thumb2_uminsi3): Likewise.
4839 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4840 (thumb2_neg_abssi2): Likewise.
4841 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4842 (thumb2_movsicc_insn): Adjust alternatives.
4843 (thumb2_mov_negscc): Disable for arm_restrict_it.
4844 (thumb2_mov_negscc_strict_it): New pattern.
4845 (thumb2_mov_notscc_strict_it): New pattern.
4846 (thumb2_mov_notscc): Disable for arm_restrict_it.
4847 (thumb2_ior_scc): Likewise.
4848 (thumb2_ior_scc_strict_it): New pattern.
4849 (thumb2_cond_move): Adjust for arm_restrict_it.
4850 (thumb2_cond_arith): Disable for arm_restrict_it.
4851 (thumb2_cond_arith_strict_it): New pattern.
4852 (thumb2_cond_sub): Adjust for arm_restrict_it.
4853 (thumb2_movcond): Likewise.
4854 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4855 (thumb2_zero_extendhisi2_v6): Likewise.
4856 (thumb2_zero_extendqisi2_v6): Likewise.
4857 (orsi_notsi_si): Likewise.
4858 (orsi_not_shiftsi_si): Likewise.
4859
4860 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4861
4862 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4863 instruction sequence is 1 byte shorter.
4864
4865 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4866
4867 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4868 it is not needed after split.
4869
4870 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4871
4872 PR target/51784
4873 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4874 second label for nonlocal goto receivers. Don't output pic base labels
4875 unless we're producing PIC; mark that action unreachable().
4876 (ix86_save_reg): If the function contains a nonlocal label, save the
4877 PIC base reg.
4878 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4879 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4880 (update_pic_label_number_if_needed): New.
4881 (machopic_output_function_base_name): Adjust for nonlocal receiver
4882 case.
4883 (machopic_should_output_picbase_label): New.
4884 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4885 (nonlocal_goto_receiver): New insn and split.
4886
4887 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4888
4889 * config/aarch64/aarch64-builtins.c
4890 (aarch64_fold_builtin): Fold abs in all modes.
4891 * config/aarch64/aarch64-simd-builtins.def
4892 (abs): Enable for all modes.
4893 * config/aarch64/arm_neon.h
4894 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4895 (vabs_f64): Add missing intrinsic.
4896
4897 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4898
4899 * config/aarch64/arm_neon.h (vabs_s64): New function
4900
4901 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4902
4903 PR target/57516
4904 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4905 * config/avr/avr.md (adjust_len): Add `round'.
4906 * config/avr/avr-protos.h (avr_out_round): New prototype.
4907 (avr_out_plus): Add `out_label' argument.
4908 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4909 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4910 Handle the case where `insn' is just a pattern.
4911 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4912 (avr_out_round): New function.
4913 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4914
4915 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4916
4917 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4918 Add check for flag_stack_usage to handle -fstack-usage support
4919
4920 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4921
4922 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4923 interaction for new Power8 flags and VSX.
4924
4925 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4926
4927 PR middle-end/57698
4928 * tree-inline.c (expand_call_inline): Emit errors during
4929 early_inlining only if optimization is not turned on.
4930
4931 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4932
4933 * passes.def: New.
4934
4935 * passes.c (init_optimization_passes): Move the construction of
4936 the pass hierarchy into a new passes.def file.
4937
4938 * Makefile.in (passes.o): Add dependency on passes.def.
4939
4940 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4941
4942 * passes.c (init_optimization_passes): Introduce macros for
4943 constructing the tree of passes (INSERT_PASSES_AFTER,
4944 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4945 TERMINATE_PASS_LIST).
4946
4947 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4948 Wei Mi <wmi@google.com>
4949
4950 PR rtl-optimization/57878
4951 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4952 top.
4953 (reload_pseudo_compare_func): Check nregs first for reload
4954 pseudos.
4955
4956 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4957
4958 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4959
4960 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4961
4962 * read-rtl.c (validate_const_int): Once an invalid character is
4963 seen, quit the loop.
4964
4965 * gengtype.c (write_roots): Similarly once we find the "deletable"
4966 or "if_marked" option.
4967
4968 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4969
4970 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4971 "xtab" and "sat". Move value "clz" from here to ...
4972 (attriubte "type"): ... here.
4973 (satsi_<SAT:code>): Delete "insn" attribute.
4974 (satsi_<SAT:code>_shift): Likewise.
4975 (arm_zero_extendqisi2addsi): Likewise.
4976 (arm_extendqisi2addsi): Likewise.
4977 (clzsi2): Update for attribute changes.
4978 (rbitsi2): Likewise.
4979 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4980 attribute.
4981 (arm_usatsihi): Likewise.
4982 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4983
4984 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4985
4986 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4987 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4988 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4989 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4990 in alphabetical order.
4991 (attribute "core_cycles"): Update for attribute changes.
4992 (arm_addsi3): Likewise.
4993 (addsi3_compare0): Likewise.
4994 (addsi3_compare0_scratch): Likewise.
4995 (addsi3_compare_op1): Likewise.
4996 (addsi3_compare_op2): Likewise.
4997 (compare_addsi2_op0): Likewise.
4998 (compare_addsi2_op1): Likewise.
4999 (addsi3_carryin_shift_<optab>): Likewise.
5000 (subsi3_carryin_shift): Likewise.
5001 (rsbsi3_carryin_shift): Likewise.
5002 (arm_subsi3_insn): Likewise.
5003 (subsi3_compare0): Likewise.
5004 (subsi3_compare): Likewise.
5005 (arm_andsi3_insn): Likewise.
5006 (thumb1_andsi3_insn): Likewise.
5007 (andsi3_compare0): Likewise.
5008 (andsi3_compare0_scratch): Likewise.
5009 (zeroextractsi_compare0_scratch
5010 (andsi_not_shiftsi_si): Likewise.
5011 (iorsi3_insn): Likewise.
5012 (iorsi3_compare0): Likewise.
5013 (iorsi3_compare0_scratch): Likewise.
5014 (arm_xorsi3): Likewise.
5015 (thumb1_xorsi3_insn): Likewise.
5016 (xorsi3_compare0): Likewise.
5017 (xorsi3_compare0_scratch): Likewise.
5018 (satsi_<SAT:code>_shift): Likewise.
5019 (rrx): Likewise.
5020 (arm_shiftsi3): Likewise.
5021 (shiftsi3_compare0): Likewise.
5022 (not_shiftsi): Likewise.
5023 (not_shiftsi_compare0): Likewise.
5024 (not_shiftsi_compare0_scratch): Likewise.
5025 (arm_one_cmplsi2): Likewise.
5026 (thumb_one_complsi2): Likewise.
5027 (notsi_compare0): Likewise.
5028 (notsi_compare0_scratch): Likewise.
5029 (thumb1_zero_extendhisi2): Likewise.
5030 (arm_zero_extendhisi2): Likewise.
5031 (arm_zero_extendhisi2_v6): Likewise.
5032 (arm_zero_extendhisi2addsi): Likewise.
5033 (thumb1_zero_extendqisi2): Likewise.
5034 (thumb1_zero_extendqisi2_v6): Likewise.
5035 (arm_zero_extendqisi2): Likewise.
5036 (arm_zero_extendqisi2_v6): Likewise.
5037 (arm_zero_extendqisi2addsi): Likewise.
5038 (thumb1_extendhisi2): Likewise.
5039 (arm_extendhisi2): Likewise.
5040 (arm_extendhisi2_v6): Likewise.
5041 (arm_extendqisi): Likewise.
5042 (arm_extendqisi_v6): Likewise.
5043 (arm_extendqisi2addsi): Likewise.
5044 (thumb1_extendqisi2): Likewise.
5045 (thumb1_movdi_insn): Likewise.
5046 (arm_movsi_insn): Likewise.
5047 (movsi_compare0): Likewise.
5048 (movhi_insn_arch4): Likewise.
5049 (movhi_bytes): Likewise.
5050 (arm_movqi_insn): Likewise.
5051 (thumb1_movqi_insn): Likewise.
5052 (arm32_movhf): Likewise.
5053 (thumb1_movhf): Likewise.
5054 (arm_movsf_soft_insn): Likewise.
5055 (thumb1_movsf_insn): Likewise.
5056 (movdf_soft_insn): Likewise.
5057 (thumb_movdf_insn): Likewise.
5058 (arm_cmpsi_insn): Likewise.
5059 (cmpsi_shiftsi): Likewise.
5060 (cmpsi_shiftsi_swp): Likewise.
5061 (arm_cmpsi_negshiftsi_si): Likewise.
5062 (movsicc_insn): Likewise.
5063 (movsfcc_soft_insn): Likewise.
5064 (arith_shiftsi): Likewise.
5065 (arith_shiftsi_compare0
5066 (arith_shiftsi_compare0_scratch
5067 (sub_shiftsi): Likewise.
5068 (sub_shiftsi_compare0
5069 (sub_shiftsi_compare0_scratch
5070 (and_scc): Likewise.
5071 (cond_move): Likewise.
5072 (if_plus_move): Likewise.
5073 (if_move_plus): Likewise.
5074 (if_move_not): Likewise.
5075 (if_not_move): Likewise.
5076 (if_shift_move): Likewise.
5077 (if_move_shift): Likewise.
5078 (if_shift_shift): Likewise.
5079 (if_not_arith): Likewise.
5080 (if_arith_not): Likewise.
5081 (cond_move_not): Likewise.
5082 (thumb1_ashlsi3): Set type attribute.
5083 (thumb1_ashrsi3): Likewise.
5084 (thumb1_lshrsi3): Likewise.
5085 (thumb1_rotrsi3): Likewise.
5086 (shiftsi3_compare0_scratch): Likewise.
5087 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5088 (neon_mov<mode>): Likewise.
5089 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5090 attribute changes.
5091 (thumb2_movsi_insn): Likewise.
5092 (thumb2_cmpsi_neg_shiftsi): Likewise.
5093 (thumb2_extendqisi_v6): Likewise.
5094 (thumb2_zero_extendhisi2_v6): Likewise.
5095 (thumb2_zero_extendqisi2_v6): Likewise.
5096 (thumb2_shiftsi3_short): Likewise.
5097 (thumb2_addsi3_compare0_scratch): Likewise.
5098 (orsi_not_shiftsi_si): Likewise.
5099 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5100 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5101 changes.
5102 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5103 (1020alu_shift_op): Likewise.
5104 (1020alu_shift_reg_op): Likewise.
5105 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5106 (alu_shift_op): Likewise.
5107 (alu_shift_reg_op): Likewise.
5108 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5109 (11_alu_shift_op): Likewise.
5110 (11_alu_shift_reg_op): Likewise.
5111 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5112 (9_alu_shift_reg_op): Likewise.
5113 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5114 changes.
5115 (cortex_a15_alu_shift): Likewise.
5116 (cortex_a15_alu_shift_reg): Likewise.
5117 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5118 changes.
5119 (cortex_a5_alu_shift): Likewise.
5120 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5121 changes.
5122 (cortex_a53_alu_shift): Likewise.
5123 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5124 changes.
5125 (cortex_a7_alu_reg): Likewise.
5126 (cortex_a7_alu_shift): Likewise.
5127 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5128 changes.
5129 (cortex_a8_alu_shift): Likewise.
5130 (cortex_a8_alu_shift_reg): Likewise.
5131 (cortex_a8_mov): Likewise.
5132 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5133 (cortex_a9_dp_shift): Likewise.
5134 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5135 changes.
5136 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5137 changes.
5138 (cortex_r4_mov): Likewise.
5139 (cortex_r4_alu_shift): Likewise.
5140 (cortex_r4_alu_shift_reg): Likewise.
5141 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5142 (526_alu_shift_op): Likewise.
5143 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5144 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5145 (626te_alu_shift_op): Likewise.
5146 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5147 (726te_alu_op): Likewise.
5148 (726te_alu_shift_op): Likewise.
5149 (726te_alu_shift_reg_op): Likewise.
5150 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5151 (mp626_alu_shift_op): Likewise.
5152 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5153 (pj4_alu_e1_conds): Likewise.
5154 (pj4_alu): Likewise.
5155 (pj4_alu_conds): Likewise.
5156 (pj4_shift): Likewise.
5157 (pj4_shift_conds): Likewise.
5158 (pj4_alu_shift): Likewise.
5159 (pj4_alu_shift_conds): Likewise.
5160 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5161 changes.
5162 (cortexa7_older_only): Likewise.
5163 (cortexa7_younger): Likewise.
5164
5165 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5166
5167 * ipa-pure-const.c (generate_summary): Rename to...
5168 (pure_const_generate_summary): ... this.
5169
5170 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5171
5172 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5173
5174 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5175
5176 PR target/57909
5177 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5178 usage in HI mode.
5179
5180 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5181
5182 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5183 enabled without -march=zEC12.
5184 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5185 flags to be set.
5186
5187 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5188
5189 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
5190 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
5191 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
5192 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
5193 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
5194 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
5195 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
5196 ISA_HAS_FP4.
5197 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
5198 and ISA_HAS_NMADD3_NMSUB3.
5199 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
5200 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
5201 (nmsub4<mode>, nmsub3<mode>): Likewise.
5202 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
5203
5204 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5205
5206 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
5207 TARGET_MIPS5400 checking.
5208
5209 2013-07-16 Jakub Jelinek <jakub@redhat.com>
5210 Peter Bergner <bergner@vnet.ibm.com>
5211
5212 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
5213 registers in the comment.
5214 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
5215 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
5216 rather than FIRST_PSEUDO_REGISTERS.
5217
5218 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
5219
5220 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
5221 enable extra ISA flags with TARGET_HTM.
5222
5223 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5224
5225 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
5226 Fix comment typos.
5227
5228 2013-07-15 Cong Hou <congh@google.com>
5229
5230 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
5231 in compare function for sorting.
5232
5233 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
5234
5235 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
5236 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
5237 * config/rs6000/rs6000.opt: Add -mhtm option.
5238 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
5239 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
5240 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
5241 __HTM__ if the HTM instructions are available.
5242 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
5243 htm_spr_reg_operand): New define_predicates.
5244 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
5245 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
5246 Include htm.md.
5247 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
5248 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
5249 HTM builtin functions.
5250 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
5251 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
5252 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
5253 (rs6000_builtin_mask_calculate): Likewise.
5254 (rs6000_option_override_internal): Likewise.
5255 (bdesc_htm): Add new HTM builtin support.
5256 (htm_spr_num): New function.
5257 (htm_spr_regno): Likewise.
5258 (rs6000_htm_spr_icode): Likewise.
5259 (htm_expand_builtin): Likewise.
5260 (htm_init_builtins): Likewise.
5261 (rs6000_expand_builtin): Add support for HTM builtin functions.
5262 (rs6000_init_builtins): Likewise.
5263 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
5264 option.
5265 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
5266 (TARGET_HTM, MASK_HTM): Define macros.
5267 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
5268 (FIXED_REGISTERS): Likewise.
5269 (CALL_USED_REGISTERS): Likewise.
5270 (CALL_REALLY_USED_REGISTERS): Likewise.
5271 (REG_ALLOC_ORDER): Likewise.
5272 (enum reg_class): Likewise.
5273 (REG_CLASS_NAMES): Likewise.
5274 (REG_CLASS_CONTENTS): Likewise.
5275 (REGISTER_NAMES): Likewise.
5276 (ADDITIONAL_REGISTER_NAMES): Likewise.
5277 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
5278 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
5279 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
5280 * config/rs6000/htm.md: New file.
5281 * config/rs6000/htmintrin.h: New file.
5282 * config/rs6000/htmxlintrin.h: New file.
5283
5284 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
5285
5286 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5287 Define SYMBOL_TINY_GOT, update comment.
5288 * config/aarch64/aarch64.c
5289 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
5290 (aarch64_expand_mov_immediate): Likewise.
5291 (aarch64_print_operand): Likewise.
5292 (aarch64_classify_symbol): Likewise.
5293 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
5294 (ldr_got_tiny): Define.
5295
5296 2013-07-13 Tobias Grosser <tobias@grosser.es>
5297
5298 PR tree-optimization/54094
5299 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
5300 scheduling dimension for the parallelism check from the polyhedral
5301 information in the AST.
5302 * graphite-dependences.c (carries_deps): Do not assume the schedule is
5303 in 2D + 1 form.
5304
5305 2013-07-13 Jason Merrill <jason@redhat.com>
5306
5307 * print-tree.c (debug_vec_tree): Use debug_raw.
5308 (debug_raw (vec<tree, va_gc> &)): New.
5309 (debug_raw (vec<tree, va_gc> *)): New.
5310 * tree.h: Declare them.
5311
5312 2013-07-13 Bin Cheng <bin.cheng@arm.com>
5313
5314 * ifcvt.c (ifcvt_after_combine): New static variable.
5315 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
5316 for size.
5317 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
5318 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
5319 rest_of_handle_if_after_reload): Pass new argument for if_convert.
5320
5321 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
5322
5323 * config/mips/mips.c (mips_expand_call): Remove empty statement.
5324
5325 2013-07-12 Michael Matz <matz@suse.de>
5326
5327 PR middle-end/55771
5328 * convert.c (convert_to_real): Reject non-float inner types.
5329
5330 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
5331
5332 * config/aarch64/aarch64-protos.h
5333 (aarch64_simd_immediate_valid_for_move): Remove.
5334 * config/aarch64/aarch64.c (simd_immediate_info): New member.
5335 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
5336 cases.
5337 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
5338
5339 2013-07-11 Steve Ellcey <sellcey@mips.com>
5340
5341 * config/mips/mips.c (mips_conditional_register_usage): Do not
5342 use t[0-7] registers in MIPS16 mode when optimizing for size.
5343
5344 2013-07-11 Sriraman Tallam <tmsriram@google.com>
5345
5346 * config/i386/i386.c (dispatch_function_versions): Fix array
5347 indexing of function_version_info to match actual_versions.
5348
5349 2013-07-11 Teresa Johnson <tejohnson@google.com>
5350
5351 * vec.h (struct va_gc): Move release out-of-line.
5352 (va_gc::release): Call ggc_free on released vec.
5353
5354 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5355
5356 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
5357 Require GOT register as additional operand in UNSPEC.
5358 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
5359 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5360 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5361 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
5362 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
5363 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
5364 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
5365
5366 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
5367
5368 PR target/57631
5369 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
5370 name seen by assembler/linker rather if available.
5371
5372 2013-07-11 Andreas Schwab <schwab@suse.de>
5373
5374 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
5375
5376 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
5377
5378 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
5379
5380 2013-07-10 Joseph Myers <joseph@codesourcery.com>
5381
5382 * doc/tm.texi.in: Move hook documentation to ....
5383 * target.def: ... here.
5384
5385 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
5386 text on @hook line.
5387 * doc/tm.texi: Regenerate.
5388
5389 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
5390
5391 PR c++/57869
5392 * doc/invoke.texi: Document -Wconditionally-supported.
5393
5394 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5395
5396 PR target/57844
5397 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
5398 of my_fp.
5399
5400 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5401
5402 PR target/57506
5403 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
5404 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
5405 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
5406 Remove duplicate devices.
5407 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
5408 * config/avr/t-multilib: Regenerate.
5409 * config/avr/avr-tables.opt: Regenerate.
5410 * doc/avr-mmcu.texi: Regenerate.
5411
5412 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5413
5414 PR target/56987
5415 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
5416
5417 2013-07-10 Graham Stott <graham.stott@btinternet.com>
5418
5419 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
5420 the cost of MULT when optimizing for size.
5421
5422 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5423
5424 * config/cr16/cr16-protos.h: Don't include target.h.
5425
5426 2013-07-09 Joseph Myers <joseph@codesourcery.com>
5427
5428 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
5429 adjust register size for TDmode and TFmode for VSX registers.
5430
5431 2013-07-08 Kai Tietz <ktietz@redhat.com>
5432
5433 PR target/56892
5434 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
5435 hook_bool_const_tree_true.
5436
5437 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5438
5439 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
5440 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
5441 * config/s390/s390.md: Define FPR*_REGNUM constants.
5442 Fix FPR2_REGNUM constant (18 -> 17).
5443 ("*trunc<BFP:mode><DFP_ALL:mode>2")
5444 ("*trunc<DFP_ALL:mode><BFP:mode>2")
5445 ("trunc<BFP:mode><DFP_ALL:mode>2")
5446 ("trunc<DFP_ALL:mode><BFP:mode>2")
5447 ("*extend<BFP:mode><DFP_ALL:mode>2")
5448 ("*extend<DFP_ALL:mode><BFP:mode>2")
5449 ("extend<BFP:mode><DFP_ALL:mode>2")
5450 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
5451 FPR4_REGNUM.
5452
5453 2013-07-08 Graham Stott <graham.stott@btinternet.com>
5454
5455 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
5456
5457 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5458
5459 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
5460 and cfun_fpr_bit_p to cfun_fpr_save_p.
5461 (s390_frame_area, s390_register_info, s390_frame_info)
5462 (s390_emit_prologue, s390_emit_epilogue)
5463 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
5464 register numbers.
5465 * config/s390/s390.h: Define *_REGNUM macros for floating point
5466 register numbers.
5467
5468 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
5469
5470 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
5471
5472 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
5473
5474 PR rtl-optimization/57786
5475 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
5476 and break out of the loop when it is set to false.
5477
5478 2013-07-08 Jakub Jelinek <jakub@redhat.com>
5479
5480 PR target/57819
5481 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
5482 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
5483 (const_int 63)) 0)).
5484 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
5485 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
5486 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
5487
5488 PR rtl-optimization/57829
5489 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
5490 mask bits outside of mode are just sign-extension from mode to HWI.
5491
5492 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5493
5494 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
5495 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
5496 adjust_address instead of change_address to keep info about alignment.
5497 (emit_strmov): Remove.
5498 (emit_memmov): New function.
5499 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
5500 (expand_movmem_epilogue): Likewise and return updated rtx for
5501 destination.
5502 (expand_constant_movmem_prologue): Likewise and return updated rtx for
5503 destination and source.
5504 (decide_alignment): Refactor, handle vector_loop.
5505 (ix86_expand_movmem): Likewise.
5506 (ix86_expand_setmem): Likewise.
5507 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
5508
5509 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
5510
5511 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
5512 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
5513
5514 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5515
5516 * config/i386/sse.md (sse_movlhps): Change alternative 3
5517 of operand 2 to "m".
5518
5519 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5520
5521 PR target/57807
5522 * config/i386/sse.md (iptr): New mode attribute.
5523 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
5524 (<sse>_vm<plusminus_insn><mode>3): Ditto.
5525 (<sse>_vmmul<mode>3): Ditto.
5526 (<sse>_vmdiv<mode>3): Ditto.
5527 (sse_vmrcpv4sf2): Ditto.
5528 (<sse>_vmsqrt<mode>2): Ditto.
5529 (sse_vmrsqrtv4sf2): Ditto.
5530 (<sse>_vm<code><mode>3): Ditto.
5531 (avx_vmcmp<mode>3): Ditto.
5532 (<sse>_vmmaskcmp<mode>3): Ditto.
5533 (<sse>_comi): Ditto.
5534 (<sse>_ucomi): Ditto.
5535 (*xop_vmfrcz_<mode>): Ditto.
5536 (*fmai_fmadd_<mode>): Ditto.
5537 (*fmai_fmsub_<mode>): Ditto.
5538 (*fmai_fnmadd_<mode>): Ditto.
5539 (*fmai_fnmsub_<mode>): Ditto.
5540 (*fma4i_vmfmadd_<mode>): Ditto.
5541 (*fma4i_vmfmsub_<mode>): Ditto.
5542 (*fma4i_vmfnmadd_<mode>): Ditto.
5543 (*fma4i_vmfnmsub_<mode>): Ditto.
5544 (*xop_vmfrcz_<mode>): Ditto.
5545 (sse_cvtps2pi): Ditto.
5546 (sse_cvttps2pi): Ditto.
5547 (sse_cvtss2si): Ditto.
5548 (sse_cvtss2si_2): Ditto.
5549 (sse_cvtss2siq_2): Ditto.
5550 (sse_cvttss2si): Ditto.
5551 (sse_cvttss2siq): Ditto.
5552 (sse_cvtsd2si): Ditto.
5553 (sse_cvtsd2si_2): Ditto.
5554 (sse_cvtsd2siq_2): Ditto.
5555 (sse_cvttsd2si): Ditto.
5556 (sse_cvttsd2siq): Ditto.
5557 (sse_cvtsd2ss): Ditto.
5558 (sse_cvtss2sd): Ditto.
5559 (avx2_pbroadcast<mode>): Ditto.
5560 (avx2_pbroadcast<mode>_1): Ditto.
5561 (*avx_vperm_broadcast_v4sf): Ditto.
5562
5563 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
5564 (sse_movlhps): Ditto.
5565 (sse_storehps): Ditto.
5566 (sse_loadhps): Ditto.
5567 (sse_storelps): Ditto.
5568 (sse_loadlps): Ditto.
5569 (*vec_concatv4sf): Ditto.
5570 (*vec_interleave_highv2df): Ditto.
5571 (*vec_interleave_lowv2df): Ditto.
5572 (*vec_extractv2df_1_sse): Ditto.
5573 (*vec_extractv2df_0_sse): Ditto.
5574 (sse2_storelpd): Ditto.
5575 (sse2_loadlpd): Ditto.
5576 (sse2_movsd): Ditto.
5577 (*vec_concatv4si): Ditto.
5578 (vec_concatv2di): Ditto.
5579
5580 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
5581 for Intel asm dialect.
5582 (mmx_punpcklwd): Ditto.
5583 (mmx_punpckldq): Ditto.
5584
5585 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
5586 for intel assembler dialect.
5587
5588 2013-07-06 Jakub Jelinek <jakub@redhat.com>
5589
5590 PR target/29776
5591 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
5592 for BUILT_IN_C{LZ,LRSB}*.
5593 * tree.h (CASE_INT_FN): Add FN##IMAX case.
5594 * tree-vrp.c (extract_range_basic): Handle
5595 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
5596 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
5597 fall thru to code calling set_value*.
5598 * builtins.c (expand_builtin): Remove *IMAX cases.
5599 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
5600 if width is bigger than 2*HWI.
5601
5602 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
5603
5604 PR rtl-optimization/55342
5605 * lra-int.h (lra_subreg_reload_pseudos): New.
5606 * lra.c: Add undoing optional reloads to the block diagram.
5607 (lra_subreg_reload_pseudos): New.
5608 (lra_optional_reload_pseudos): Change comments.
5609 (lra): Init and clear lra_subreg_reload_pseudos. Clear
5610 lra_optional_reload_pseudos after undo transformations.
5611 * lra-assigns.c (pseudo_prefix_title): New.
5612 (lra_setup_reg_renumber): Use it.
5613 (spill_for): Ditto. Check subreg reload pseudos too.
5614 (assign_by_spills): Consider subreg reload pseudos too.
5615 * lra-constraints.c (simplify_operand_subreg): Use
5616 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
5617 (curr_insn_transform): Recognize and do optional reloads.
5618 (undo_optional_reloads): New.
5619 (lra_undo_inheritance): Call undo_optional_reloads.
5620
5621 2013-07-05 Thomas Quinot <quinot@adacore.com>
5622
5623 * tree-complex.c (expand_complex_operations_1): Fix typo.
5624
5625 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
5626
5627 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
5628 (tune_params): New member 'const vec_costs'.
5629 * config/aarch64/aarch64.c (generic_vector_cost): New.
5630 (generic_tunings): New member 'generic_vector_cost'.
5631 (aarch64_builtin_vectorization_cost): New.
5632 (aarch64_add_stmt_cost): New.
5633 (TARGET_VECTORIZE_ADD_STMT_COST): New.
5634 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
5635
5636 2013-07-03 Jakub Jelinek <jakub@redhat.com>
5637
5638 PR target/57777
5639 * config/i386/predicates.md (vsib_address_operand): Disallow
5640 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
5641
5642 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
5643
5644 PR middle-end/55030
5645 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
5646 expand_builtin_setjmp_receiver.
5647 (expand_label): Adjust, call expand_builtin_setjmp_receiver
5648 with NULL for the label parameter.
5649 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
5650 the frame-pointer. Adjust comments.
5651 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
5652 only if LABEL is non-NULL.
5653
5654 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
5655
5656 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
5657 (ARM_ABI_AAPCS64): Ditto.
5658 (arm_abi): Ditto.
5659 (ARM_DEFAULT_ABI): Ditto.
5660
5661 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
5662
5663 * config/aarch64/aarch64-builtins.c
5664 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
5665 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
5666 (st1): Likewise.
5667 * config/aarch64/aarch64-simd.md
5668 (aarch64_ld1<VALL:mode>): New.
5669 (aarch64_st1<VALL:mode>): Likewise.
5670 * config/aarch64/arm_neon.h
5671 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
5672
5673 2013-07-02 Sriraman Tallam <tmsriram@google.com>
5674
5675 * config/i386/i386.c (gate_insert_vzeroupper): Check if
5676 target ISA is AVX.
5677 (ix86_option_override_internal):Turn on all -mavx target flags by
5678 default as they are dependent on AVX anyway.
5679
5680 2013-07-02 Cary Coutant <ccoutant@google.com>
5681
5682 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
5683 deterministic hash.
5684 (loc_checksum_ordered): Likewise.
5685 (hash_loc_operands): Remove inline keyword.
5686
5687 2013-07-02 Jakub Jelinek <jakub@redhat.com>
5688
5689 PR tree-optimization/57741
5690 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
5691 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
5692 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
5693 Allow REAL_CST step_exprs if flag_associative_math.
5694 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
5695
5696 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5697
5698 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
5699
5700 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5701
5702 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
5703
5704 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5705
5706 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
5707 encoding.
5708 (iorsi3_insn): Likewise.
5709 (arm_xorsi3): Likewise.
5710
5711 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
5712
5713 * arm.md (attribute "wtype"): Delete. Move attribute values from here
5714 to ...
5715 (attribute "type"): ... here, and prefix with "wmmx_".
5716 (attribute "core_cycles"): Update for attribute changes.
5717 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
5718 (tbcstv4hi): Likewise.
5719 (tbcstv2si): Likewise.
5720 (iwmmxt_iordi3): Likewise.
5721 (iwmmxt_xordi3): Likewise.
5722 (iwmmxt_anddi3): Likewise.
5723 (iwmmxt_nanddi3): Likewise.
5724 (iwmmxt_arm_movdi): Likewise.
5725 (iwmmxt_movsi_insn): Likewise.
5726 (mov<mode>_internal): Likewise.
5727 (and<mode>3_iwmmxt): Likewise.
5728 (ior<mode>3_iwmmxt): Likewise.
5729 (xor<mode>3_iwmmxt): Likewise.
5730 (add<mode>3_iwmmxt): Likewise.
5731 (ssaddv8qi3): Likewise.
5732 (ssaddv4hi3): Likewise.
5733 (ssaddv2si3): Likewise.
5734 (usaddv8qi3): Likewise.
5735 (usaddv4hi3): Likewise.
5736 (usaddv2si3): Likewise.
5737 (sub<mode>3_iwmmxt): Likewise.
5738 (sssubv8qi3): Likewise.
5739 (sssubv4hi3): Likewise.
5740 (sssubv2si3): Likewise.
5741 (ussubv8qi3): Likewise.
5742 (ussubv4hi3): Likewise.
5743 (ussubv2si3): Likewise.
5744 (mulv4hi3_iwmmxt): Likewise.
5745 (smulv4hi3_highpart): Likewise.
5746 (umulv4hi3_highpart): Likewise.
5747 (iwmmxt_wmacs): Likewise.
5748 (iwmmxt_wmacsz): Likewise.
5749 (iwmmxt_wmacu): Likewise.
5750 (iwmmxt_wmacuz): Likewise.
5751 (iwmmxt_clrdi): Likewise.
5752 (iwmmxt_clrv8qi): Likewise.
5753 (iwmmxt_clr4hi): Likewise.
5754 (iwmmxt_clr2si): Likewise.
5755 (iwmmxt_uavgrndv8qi3): Likewise.
5756 (iwmmxt_uavgrndv4hi3): Likewise.
5757 (iwmmxt_uavgv8qi3): Likewise.
5758 (iwmmxt_uavgv4hi3): Likewise.
5759 (iwmmxt_tinsrb): Likewise.
5760 (iwmmxt_tinsrh): Likewise.
5761 (iwmmxt_tinsrw): Likewise.
5762 (iwmmxt_textrmub): Likewise.
5763 (iwmmxt_textrmsb): Likewise.
5764 (iwmmxt_textrmuh): Likewise.
5765 (iwmmxt_textrmsh): Likewise.
5766 (iwmmxt_textrmw): Likewise.
5767 (iwmxxt_wshufh): Likewise.
5768 (eqv8qi3): Likewise.
5769 (eqv4hi3): Likewise.
5770 (eqv2si3): Likewise.
5771 (gtuv8qi3): Likewise.
5772 (gtuv4hi3): Likewise.
5773 (gtuv2si3): Likewise.
5774 (gtv8qi3): Likewise.
5775 (gtv4hi3): Likewise.
5776 (gtv2si3): Likewise.
5777 (smax<mode>3_iwmmxt): Likewise.
5778 (umax<mode>3_iwmmxt): Likewise.
5779 (smin<mode>3_iwmmxt): Likewise.
5780 (umin<mode>3_iwmmxt): Likewise.
5781 (iwmmxt_wpackhss): Likewise.
5782 (iwmmxt_wpackwss): Likewise.
5783 (iwmmxt_wpackdss): Likewise.
5784 (iwmmxt_wpackhus): Likewise.
5785 (iwmmxt_wpackwus): Likewise.
5786 (iwmmxt_wpackdus): Likewise.
5787 (iwmmxt_wunpckihb): Likewise.
5788 (iwmmxt_wunpckihh): Likewise.
5789 (iwmmxt_wunpckihw): Likewise.
5790 (iwmmxt_wunpckilb): Likewise.
5791 (iwmmxt_wunpckilh): Likewise.
5792 (iwmmxt_wunpckilw): Likewise.
5793 (iwmmxt_wunpckehub): Likewise.
5794 (iwmmxt_wunpckehuh): Likewise.
5795 (iwmmxt_wunpckehuw): Likewise.
5796 (iwmmxt_wunpckehsb): Likewise.
5797 (iwmmxt_wunpckehsh): Likewise.
5798 (iwmmxt_wunpckehsw): Likewise.
5799 (iwmmxt_wunpckelub): Likewise.
5800 (iwmmxt_wunpckeluh): Likewise.
5801 (iwmmxt_wunpckeluw): Likewise.
5802 (iwmmxt_wunpckelsb): Likewise.
5803 (iwmmxt_wunpckelsh): Likewise.
5804 (iwmmxt_wunpckelsw): Likewise.
5805 (ror<mode>3): Likewise.
5806 (ashr<mode>3_iwmmxt): Likewise.
5807 (lshr<mode>3_iwmmxt): Likewise.
5808 (ashl<mode>3_iwmmxt): Likewise.
5809 (ror<mode>3_di): Likewise.
5810 (ashr<mode>3_di): Likewise.
5811 (lshr<mode>3_di): Likewise.
5812 (ashl<mode>3_di): Likewise.
5813 (iwmmxt_wmadds): Likewise.
5814 (iwmmxt_wmaddu): Likewise.
5815 (iwmmxt_tmia): Likewise.
5816 (iwmmxt_tmiaph): Likewise.
5817 (iwmmxt_tmiabb): Likewise.
5818 (iwmmxt_tmiatb): Likewise.
5819 (iwmmxt_tmiabt): Likewise.
5820 (iwmmxt_tmiatt): Likewise.
5821 (iwmmxt_tmovmskb): Likewise.
5822 (iwmmxt_tmovmskh): Likewise.
5823 (iwmmxt_tmovmskw): Likewise.
5824 (iwmmxt_waccb): Likewise.
5825 (iwmmxt_wacch): Likewise.
5826 (iwmmxt_waccw): Likewise.
5827 (iwmmxt_waligni): Likewise.
5828 (iwmmxt_walignr): Likewise.
5829 (iwmmxt_walignr0): Likewise.
5830 (iwmmxt_walignr1): Likewise.
5831 (iwmmxt_walignr2): Likewise.
5832 (iwmmxt_walignr3): Likewise.
5833 (iwmmxt_wsadb): Likewise.
5834 (iwmmxt_wsadh): Likewise.
5835 (iwmmxt_wsadbz): Likewise.
5836 (iwmmxt_wsadhz): Likewise.
5837 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5838 (iwmmxt_wabsdiffb): Likewise.
5839 (iwmmxt_wabsdiffh): Likewise.
5840 (iwmmxt_wabsdiffw): Likewise.
5841 (iwmmxt_waddsubhx): Likewise
5842 (iwmmxt_wsubaddhx): Likewise.
5843 (addc<mode>3): Likewise.
5844 (iwmmxt_avg4): Likewise.
5845 (iwmmxt_avg4r): Likewise.
5846 (iwmmxt_wmaddsx): Likewise.
5847 (iwmmxt_wmaddux): Likewise.
5848 (iwmmxt_wmaddsn): Likewise.
5849 (iwmmxt_wmaddun): Likewise.
5850 (iwmmxt_wmulwsm): Likewise.
5851 (iwmmxt_wmulwum): Likewise.
5852 (iwmmxt_wmulsmr): Likewise.
5853 (iwmmxt_wmulumr): Likewise.
5854 (iwmmxt_wmulwsmr): Likewise.
5855 (iwmmxt_wmulwumr): Likewise.
5856 (iwmmxt_wmulwl): Likewise.
5857 (iwmmxt_wqmulm): Likewise.
5858 (iwmmxt_wqmulwm): Likewise.
5859 (iwmmxt_wqmulmr): Likewise.
5860 (iwmmxt_wqmulwmr): Likewise.
5861 (iwmmxt_waddbhusm): Likewise.
5862 (iwmmxt_waddbhusl): Likewise.
5863 (iwmmxt_wqmiabb): Likewise.
5864 (iwmmxt_wqmiabt): Likewise.
5865 (iwmmxt_wqmiatb): Likewise.
5866 (iwmmxt_wqmiatt): Likewise.
5867 (iwmmxt_wqmiabbn): Likewise.
5868 (iwmmxt_wqmiabtn): Likewise.
5869 (iwmmxt_wqmiatbn): Likewise.
5870 (iwmmxt_wqmiattn): Likewise.
5871 (iwmmxt_wmiabb): Likewise.
5872 (iwmmxt_wmiabt): Likewise.
5873 (iwmmxt_wmiatb): Likewise.
5874 (iwmmxt_wmiatt): Likewise.
5875 (iwmmxt_wmiabbn): Likewise.
5876 (iwmmxt_wmiabtn): Likewise.
5877 (iwmmxt_wmiatbn): Likewise.
5878 (iwmmxt_wmiattn): Likewise.
5879 (iwmmxt_wmiawbb): Likewise.
5880 (iwmmxt_wmiawbt): Likewise.
5881 (iwmmxt_wmiawtb): Likewise.
5882 (iwmmxt_wmiawtt): Likewise.
5883 (iwmmxt_wmiawbbn): Likewise.
5884 (iwmmxt_wmiawbtn): Likewise.
5885 (iwmmxt_wmiawtbn): Likewise.
5886 (iwmmxt_wmiawttn): Likewise.
5887 (iwmmxt_wmerge): Likewise.
5888 (iwmmxt_tandc<mode>3): Likewise.
5889 (iwmmxt_torc<mode>3): Likewise.
5890 (iwmmxt_torvsc<mode>3): Likewise.
5891 (iwmmxt_textrc<mode>3): Likewise.
5892 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5893 (wmmxt_pack): Likewise.
5894 (wmmxt_mult_c1): Likewise.
5895 (wmmxt_mult_c2): Likewise.
5896 (wmmxt_alu_c1): Likewise.
5897 (wmmxt_alu_c2): Likewise.
5898 (wmmxt_alu_c3): Likewise.
5899 (wmmxt_transfer_c1): Likewise.
5900 (wmmxt_transfer_c2): Likewise.
5901 (wmmxt_transfer_c3): Likewise.
5902 (marvell_f_iwmmxt_wstr): Likewise.
5903 (marvell_f_iwmmxt_wldr): Likewise.
5904
5905 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5906
5907 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5908
5909 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5910
5911 Revert:
5912 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5913 * lra-constraints.c (need_for_split_p): Check call used hard regs
5914 living through calls.
5915
5916 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5917 call used regs for call insn.
5918
5919 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5920
5921 PR target/57736
5922 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5923 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5924
5925 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5926
5927 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5928
5929 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5930
5931 * lra-constraints.c (need_for_split_p): Check call used hard regs
5932 living through calls.
5933
5934 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5935
5936 PR target/57744
5937 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5938 to tie with any other modes. Eliminate Altivec vector mode tests,
5939 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5940 code, to return 0 if testing MODE2 for a condition, if we've
5941 already tested MODE1 for the same condition.
5942
5943 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5944
5945 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5946 layout.
5947
5948 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5949
5950 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5951 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5952
5953 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5954
5955 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5956 Define.
5957 (aarch64_symbolic_constant_p): Remove.
5958 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5959 static. Fix line length and white space.
5960 (aarch64_symbolic_constant_p): Remove.
5961 * config/aarch64/predicates.md (aarch64_valid_symref):
5962 Use aarch64_classify_symbol_expression.
5963
5964 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5965
5966 * config/arm/constraints.md (Ts): New constraint.
5967 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5968 16-bit encodings.
5969 (compare_scc): Use "Ts" constraint for operand 0.
5970 (ior_scc_scc): Likewise.
5971 (and_scc_scc): Likewise.
5972 (and_scc_scc_nodom): Likewise.
5973 (ior_scc_scc_cmp): Likewise for operand 7.
5974 (and_scc_scc_cmp): Likewise.
5975 * config/arm/thumb2.md (thumb2_movsi_insn):
5976 Add alternatives for 16-bit encodings.
5977 (thumb2_movhi_insn): Likewise.
5978 (thumb2_movsicc_insn): Likewise.
5979 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5980 (thumb2_negscc): Use "Ts" constraint.
5981 Move mvn instruction outside cond_exec block.
5982 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5983 for 16-bit encodings.
5984
5985 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5986
5987 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5988 encoding.
5989 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5990 (mulsi3subsi): Likewise.
5991 (mulsidi3adddi): Likewise.
5992 (mulsidi3_v6): Likewise.
5993 (umulsidi3_v6): Likewise.
5994 (umulsidi3adddi_v6): Likewise.
5995 (smulsi3_highpart_v6): Likewise.
5996 (umulsi3_highpart_v6): Likewise.
5997 (mulhisi3tb): Likewise.
5998 (mulhisi3bt): Likewise.
5999 (mulhisi3tt): Likewise.
6000 (maddhisi4): Likewise.
6001 (maddhisi4tb): Likewise.
6002 (maddhisi4tt): Likewise.
6003 (maddhidi4): Likewise.
6004 (maddhidi4tb): Likewise.
6005 (maddhidi4tt): Likewise.
6006 (zeroextractsi_compare0_scratch): Likewise.
6007 (insv_zero): Likewise.
6008 (insv_t2): Likewise.
6009 (anddi_notzesidi_di): Likewise.
6010 (anddi_notsesidi_di): Likewise.
6011 (andsi_notsi_si): Likewise.
6012 (iordi_zesidi_di): Likewise.
6013 (xordi_zesidi_di): Likewise.
6014 (andsi_iorsi3_notsi): Likewise.
6015 (smax_0): Likewise.
6016 (smax_m1): Likewise.
6017 (smin_0): Likewise.
6018 (not_shiftsi): Likewise.
6019 (unaligned_loadsi): Likewise.
6020 (unaligned_loadhis): Likewise.
6021 (unaligned_loadhiu): Likewise.
6022 (unaligned_storesi): Likewise.
6023 (unaligned_storehi): Likewise.
6024 (extv_reg): Likewise.
6025 (extzv_t2): Likewise.
6026 (divsi3): Likewise.
6027 (udivsi3): Likewise.
6028 (arm_zero_extendhisi2addsi): Likewise.
6029 (arm_zero_extendqisi2addsi): Likewise.
6030 (compareqi_eq0): Likewise.
6031 (arm_extendhisi2_v6): Likewise.
6032 (arm_extendqisi2addsi): Likewise.
6033 (arm_movt): Likewise.
6034 (thumb2_ldrd): Likewise.
6035 (thumb2_ldrd_base): Likewise.
6036 (thumb2_ldrd_base_neg): Likewise.
6037 (thumb2_strd): Likewise.
6038 (thumb2_strd_base): Likewise.
6039 (thumb2_strd_base_neg): Likewise.
6040 (arm_negsi2): Add alternative for 16-bit encoding.
6041 (arm_one_cmplsi2): Likewise.
6042
6043 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6044
6045 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6046 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6047 (movdfcc): Likewise.
6048 * config/arm/vfp.md (*thumb2_movsf_vfp):
6049 Disable predication for arm_restrict_it.
6050 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6051 (*thumb2_movdfcc_vfp): Likewise.
6052 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6053 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6054 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6055 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6056 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6057 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6058 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6059 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6060 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6061 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6062 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6063 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6064 Disable predication for arm_restrict_it.
6065
6066 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6067
6068 * config/i386/bmiintrin.h (_bextr_u32): New.
6069 (_bextr_u64): Ditto.
6070
6071 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6072
6073 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6074 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6075 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6076 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6077 * config/mips/n32-elf.h: ...this new file.
6078
6079 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6080
6081 PR target/57224
6082 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6083 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6084
6085 2013-06-27 Catherine Moore <clm@codesourcery.com>
6086
6087 * config/mips/mips-tables.opt: Regenerate.
6088 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6089 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6090 * doc/invoke.texi: Add -m14kc.
6091
6092 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6093
6094 PR target/57623
6095 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6096 constraints of operand 1 and 2.
6097
6098 PR target/57623
6099 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6100 to match RTL canonicalization. Swap predicates and
6101 constraints of operand 1 and 2.
6102
6103 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6104
6105 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6106 Process OP_INOUT regs for splitting too.
6107
6108 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6109
6110 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6111 decl before the loop, initialize to NULL.
6112 (vectorizable_load): Initialize ptr_incr to NULL.
6113
6114 2013-06-27 Martin Jambor <mjambor@suse.cz>
6115
6116 PR lto/57208
6117 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6118 * ipa-ref.c (ipa_maybe_record_reference): New function.
6119 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6120 * ipa-cp.c (create_specialized_node): Record potential references from
6121 aggvals.
6122 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6123
6124 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6125
6126 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6127 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6128 'mode' to force_reg.
6129 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6130 (aarch64_expand_mov_immediate): Likewise.
6131
6132 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6133
6134 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6135 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6136
6137 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6138
6139 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6140 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6141 (struct machine_function): Add tbegin_p.
6142 (s390_canonicalize_comparison): Fold CC mode compares to
6143 conditional jump if possible.
6144 (s390_emit_jump): Return the emitted jump.
6145 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6146 Handle CCRAWmode compares.
6147 (s390_option_override): Default to -mhtm if available.
6148 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6149 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6150 FPRs instead of df_regs_ever_live_p.
6151 (s390_optimize_nonescaping_tx): New function.
6152 (s390_init_frame_layout): Extend clobbered_regs array to cover
6153 FPRs as well.
6154 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6155 (s390_expand_tbegin): New function.
6156 (enum s390_builtin): New enum definition.
6157 (code_for_builtin): New array definition.
6158 (s390_init_builtins): New function.
6159 (s390_expand_builtin): New function.
6160 (TARGET_INIT_BUILTINS): Define.
6161 (TARGET_EXPAND_BUILTIN): Define.
6162 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6163 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6164 (s390_alc_comparison): Likewise.
6165 * config/s390/s390-modes.def: Add CCRAWmode.
6166 * config/s390/s390.h (processor_flags): Add PF_TX.
6167 (TARGET_CPU_HTM): Define macro.
6168 (TARGET_HTM): Define macro.
6169 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6170 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6171 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6172 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6173 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6174 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6175 ("*cc_to_int"): Move up.
6176 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6177 constants other than 0.
6178 ("*ccraw_to_int"): New insn and splitter definition.
6179 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6180 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6181 ("tx_assist"): New expander.
6182 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6183 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6184 * config/s390/s390.opt: Add -mhtm option.
6185 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6186 * config/s390/htmxlintrin.h: New file.
6187 * config/s390/htmintrin.h: New file.
6188 * config/s390/s390intrin.h: New file.
6189 * doc/extend.texi: Document htm builtins.
6190 * config.gcc: Add the new header files to extra_headers.
6191
6192 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
6193
6194 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
6195 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
6196
6197 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6198 Pat Haugen <pthaugen@us.ibm.com>
6199 Peter Bergner <bergner@vnet.ibm.com>
6200
6201 * config/rs6000/power8.md: New.
6202 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
6203 setting for power8 entry.
6204 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
6205 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
6206 test for Power4/Power5 only.
6207 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
6208 support.
6209 (force_new_group): Adjust comment.
6210 * config/rs6000/rs6000.md: Include power8.md.
6211
6212 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
6213
6214 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
6215 * config/arm/arm-protos.h (arm_max_conditional_execute): New
6216 declaration.
6217 (tune_params): Update comment.
6218 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
6219 (arm_max_conditional_execute): New function.
6220 (thumb2_final_prescan_insn): Use max_insn_skipped and
6221 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
6222
6223 2013-06-25 Jakub Jelinek <jakub@redhat.com>
6224
6225 PR tree-optimization/57705
6226 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
6227 SSA_NAME step, provided that it is not defined inside the loop.
6228 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
6229 (get_initial_def_for_induction): Handle SSA_NAME IV step.
6230
6231 2013-06-25 Martin Jambor <mjambor@suse.cz>
6232
6233 PR middle-end/57670
6234 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
6235 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
6236 calls in the dump.
6237 (ipa_note_param_call): Initialize member_ptr flag.
6238 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
6239 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
6240 (ipa_write_indirect_edge_info): Stream member_ptr flag.
6241 (ipa_read_indirect_edge_info): Likewise.
6242
6243 2013-06-25 Richard Biener <rguenther@suse.de>
6244
6245 PR middle-end/56977
6246 * passes.c (init_optimization_passes): Move pass_fold_builtins
6247 and pass_dce earlier with -Og.
6248
6249 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
6250
6251 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
6252 <BIT_FIELD_REF>: Remove trailing TAB.
6253 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
6254 remove blank line.
6255
6256 2013-06-24 Martin Jambor <mjambor@suse.cz>
6257
6258 PR tree-optimization/57358
6259 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
6260 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
6261 (ipa_analyze_params_uses): Generate pessimistic info when true.
6262
6263 2013-06-24 Martin Jambor <mjambor@suse.cz>
6264
6265 PR tree-optimization/57539
6266 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
6267 global.inlined_to of the new node to it. All callers changed.
6268 * ipa-inline-transform.c (clone_inlined_nodes): New variable
6269 inlining_into, pass it to cgraph_clone_node.
6270 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
6271 ipa_free_edge_args_substructures.
6272 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
6273 rdesc linked list. Do not assert rdesc edges have inlined caller.
6274 Assert we have found an rdesc in the rdesc list.
6275
6276 2013-06-24 Richard Biener <rguenther@suse.de>
6277
6278 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
6279 (pointer_set_lookup): Declare.
6280 (class pointer_map): New template class implementing a
6281 generic pointer to T map.
6282 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
6283 pointer_map<T>::contains, pointer_map<T>::insert,
6284 pointer_map<T>::traverse): New functions.
6285 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
6286 (pointer_set_lookup): New function.
6287 (pointer_set_contains): Use pointer_set_lookup.
6288 (pointer_set_insert): Likewise.
6289 (insert_aux): Remove.
6290 (struct pointer_map_t): Embed a pointer_set_t.
6291 (pointer_map_create): Adjust.
6292 (pointer_map_destroy): Likewise.
6293 (pointer_map_contains): Likewise.
6294 (pointer_map_insert): Likewise.
6295 (pointer_map_traverse): Likewise.
6296 * tree-streamer.h (struct streamer_tree_cache_d): Use a
6297 pointer_map<unsigned> instead of a pointer_map_t.
6298 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
6299 (streamer_tree_cache_lookup): Likewise.
6300 (streamer_tree_cache_create): Likewise.
6301 (streamer_tree_cache_delete): Likewise.
6302 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
6303 pointer_map<unsigned> instead of a pointer_map_t.
6304 (lto_init_tree_ref_encoder): Adjust.
6305 (lto_destroy_tree_ref_encoder): Likewise.
6306 * lto-section-out.c (lto_output_decl_index): Likewise.
6307 (lto_record_function_out_decl_state): Likewise.
6308 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
6309
6310 2013-06-24 Richard Biener <rguenther@suse.de>
6311
6312 PR tree-optimization/57488
6313 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
6314
6315 2013-06-24 Alan Modra <amodra@gmail.com>
6316
6317 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
6318 (gen_easy_altivec_constant): Likewise.
6319 * config/rs6000/predicates.md (easy_vector_constant_add_self,
6320 easy_vector_constant_msb): Likewise.
6321
6322 2013-06-23 Jakub Jelinek <jakub@redhat.com>
6323
6324 PR target/57688
6325 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
6326 add missing return true.
6327
6328 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
6329
6330 PR target/52483
6331 * config/sh/predicates.md (general_extend_operand): Invoke
6332 general_movsrc_operand for memory operands.
6333 (general_movsrc_operand): Allow reg+reg addressing, do not use
6334 general_operand for memory operands.
6335
6336 2013-06-23 Sriraman Tallam <tmsriram@google.com>
6337
6338 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
6339 when current target options does not apply.
6340 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
6341 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
6342 * config/i386/bmiintrin.h: Pass appropriate target
6343 attributes to header.
6344 * config/i386/mmintrin.h: Ditto.
6345 * config/i386/nmmintrin.h: Ditto.
6346 * config/i386/avx2intrin.h: Ditto.
6347 * config/i386/fxsrintrin.h: Ditto.
6348 * config/i386/tbmintrin.h: Ditto.
6349 * config/i386/xsaveintrin.h: Ditto.
6350 * config/i386/f16cintrin.h: Ditto.
6351 * config/i386/xtestintrin.h: Ditto.
6352 * config/i386/xsaveoptintrin.h: Ditto.
6353 * config/i386/bmi2intrin.h: Ditto.
6354 * config/i386/lzcntintrin.h: Ditto.
6355 * config/i386/smmintrin.h: Ditto.
6356 * config/i386/wmmintrin.h: Ditto.
6357 * config/i386/x86intrin.h: Remove all header include guards.
6358 * config/i386/prfchwintrin.h: Ditto.
6359 * config/i386/pmmintrin.h: Ditto.
6360 * config/i386/tmmintrin.h: Ditto.
6361 * config/i386/xmmintrin.h: Ditto.
6362 * config/i386/popcntintrin.h: Ditto.
6363 * config/i386/rdseedintrin.h: Ditto.
6364 * config/i386/ammintrin.h: Ditto.
6365 * config/i386/emmintrin.h: Ditto.
6366 * config/i386/immintrin.h: Remove all header include guards.
6367 * config/i386/fma4intrin.h: Ditto.
6368 * config/i386/lwpintrin.h: Ditto.
6369 * config/i386/xopintrin.h: Ditto.
6370 * config/i386/ia32intrin.h: Ditto.
6371 * config/i386/avxintrin.h: Ditto.
6372 * config/i386/rtmintrin.h: Ditto.
6373 * config/i386/fmaintrin.h: Ditto.
6374 * config/i386/mm3dnow.h: Ditto.
6375
6376 2013-06-22 Sriraman Tallam <tmsriram@google.com>
6377
6378 * common/config/i386/i386-common.c: Handle LZCNT.
6379
6380 2013-06-22 Andi Kleen <ak@linux.intel.com>
6381
6382 * doc/extend.texi: Use __atomic_store_n instead of
6383 __atomic_store in HLE example.
6384
6385 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
6386
6387 * config/sh/sh.c: Remove <cstdlib> workaround.
6388
6389 2013-06-21 Andi Kleen <ak@linux.intel.com>
6390
6391 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
6392
6393 2013-06-21 Andi Kleen <ak@linux.intel.com>
6394
6395 * doc/extend.texi: Document that __atomic_clear and
6396 __atomic_test_and_set should only be used with bool.
6397
6398 2013-06-20 Jan Hubicka <jh@suse.cz>
6399
6400 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
6401 types_same_for_odr.
6402 * tree.c (decls_same_for_odr): New function.
6403 (same_for_edr): New function.
6404 (types_same_for_odr): New function.
6405 (get_binfo_at_offset): Use it.
6406 * tree.h (types_same_for_odr): Declare.
6407
6408 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
6409 Jason Merrill <jason@redhat.com>
6410
6411 * system.h: Include <cstdlib> as well as <stdlib.h>.
6412
6413 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
6414
6415 PR target/57655
6416 * config/i386/i386.c (construct_container): Report error if
6417 long double is used with disabled x87 float returns.
6418
6419 2013-06-20 Jan Hubicka <jh@suse.cz>
6420
6421 * lto-cgraph.c (input_symtab): Do not set cgraph state.
6422
6423 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
6424
6425 PR rtl-optimization/57425
6426 PR rtl-optimization/57569
6427 * alias.c (write_dependence_p): Remove parameters mem_mode and
6428 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
6429 Changed all callers.
6430 (canon_anti_dependence): Get comments and semantics in sync.
6431 Add parameter mem_canonicalized. Changed all callers.
6432 * rtl.h (canon_anti_dependence): Update prototype.
6433
6434 2013-06-20 Richard Biener <rguenther@suse.de>
6435
6436 * data-streamer-in.c (streamer_read_uhwi): Optimize single
6437 byte case, inline streamer_read_uchar and defer section
6438 overrun check.
6439
6440 2013-06-20 Richard Biener <rguenther@suse.de>
6441
6442 PR tree-optimization/57584
6443 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
6444 SSA names into the expanded expression that take part in
6445 abnormal coalescing.
6446
6447 2013-06-19 Sharad Singhai <singhai@google.com>
6448
6449 * gcov.c (print_usage): Handle new option.
6450 (process_args): Ditto.
6451 (get_gcov_intermediate_filename): New function.
6452 (output_intermediate_file): New function.
6453 (output_gcov_file): New function
6454 (generate_results): Handle new option.
6455 (release_function): Relase demangled name.
6456 (read_graph_file): Handle demangled name.
6457 (output_lines): Ditto.
6458 * doc/gcov.texi: Document gcov intermediate format.
6459
6460 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
6461
6462 PR bootstrap/57604
6463 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
6464 (lra_emit_add): Use the functions. Add comment about Y as an
6465 address segment.
6466
6467 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
6468
6469 PR driver/57652
6470 * collect2.c (collect_atexit): New.
6471 (collect_exit): Delete.
6472 (main): Register collect_atexit with atexit.
6473 (collect_wait): Change collect_exit to exit.
6474 (do_wait): Same.
6475 * collect2.h (collect_exit): Delete.
6476 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
6477
6478 2013-06-19 Wei Mi <wmi@google.com>
6479
6480 PR rtl-optimization/57518
6481 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
6482 if regno is used in paradoxical subreg.
6483 (update_equiv_regs): Check pdx_subregs[regno] before
6484 set a reg to be equivalent with a mem.
6485
6486 2013-06-19 Matthias Klose <doko@ubuntu.com>
6487
6488 PR driver/57651
6489 * file-find.h (find_a_file): Add a mode parameter.
6490 * file-find.c (find_a_file): Likewise.
6491 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
6492 with X_OK for the executables.
6493 * collect2.c (main): Call find_a_file with X_OK.
6494
6495 2013-06-19 Steve Ellcey <sellcey@mips.com>
6496
6497 PR target/56942
6498 * config/mips/mips.md (casesi_internal_mips16_<mode>):
6499 Use NEXT_INSN instead of next_real_insn.
6500
6501 2013-06-19 Jan Hubicka <jh@suse.cz>
6502
6503 * cgraph.h (const_value_known_p): Replace by ...
6504 (ctor_for_folding): .. this one.
6505 * cgraphunit.c (process_function_and_variable_attributes): Use it.
6506 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
6507 * expr.c (expand_expr_real_1): Likewise.
6508 (string_constant): Likewise.
6509 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
6510 * ipa.c (process_references): Likewise.
6511 (symtab_remove_unreachable_nodes): Likewise.
6512 * ipa-inline-analysis.c (param_change_prob): Likewise.
6513 * gimple-fold.c (canonicalize_constructor_val): Likewise.
6514 (get_base_constructor): Likwise.
6515 * varpool.c (varpool_remove_node): Likewise.
6516 (varpool_remove_initializer): LIkewise.
6517 (dump_varpool_node): LIkwise.
6518 (const_value_known_p): Rewrite to ...
6519 (ctor_for_folding): ... this one.
6520
6521 2013-06-19 Jakub Jelinek <jakub@redhat.com>
6522
6523 PR driver/57651
6524 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
6525 PERSONALITY in $PATH derived prefixes.
6526
6527 2013-06-19 Jeff Law <law@redhat.com>
6528
6529 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
6530 in comment.
6531
6532 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
6533 (simplify_bitwise_binary): Use it to simpify certain binary ops on
6534 booleans.
6535
6536 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
6537
6538 * config/arm/vfp.md: Move VFP instruction classification documentation
6539 to ...
6540 * config/arm/arm.md: ... here. Update instruction classification
6541 documentation.
6542
6543 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
6544
6545 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
6546 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
6547 pattern. Use more efficient sequences on ARMv5 and Thumb2.
6548
6549 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
6550
6551 PR target/57609
6552 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6553 with NEXT_INSN. Use tablejump_p to check for jump table data
6554 insns.
6555
6556 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
6557
6558 PR c++/56544
6559 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
6560 that now in C++ the value is correct per the C++ standards.
6561
6562 2013-06-19 Richard Biener <rguenther@suse.de>
6563
6564 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
6565 for global context.
6566
6567 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6568
6569 Revert:
6570 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6571
6572 PR target/57609
6573 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6574 with next_active_insn.
6575
6576 2013-06-18 Sriraman Tallam <tmsriram@google.com>
6577
6578 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
6579 functions are inlined during failures to flag an error.
6580 * tree-inline.c (expand_call_inline): Allow the error to be flagged
6581 in early inline pass.
6582
6583 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
6584
6585 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
6586 in comments.
6587
6588 2013-06-18 Julian Brown <julian@codesourcery.com>
6589
6590 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
6591 Permit virtual register pre-reload if !strict.
6592 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
6593 change.
6594 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
6595 prototype.
6596 * config/arm/neon.md (movmisalign<mode>): Use
6597 neon_perm_struct_or_reg_operand instead of
6598 neon_struct_or_register_operand.
6599 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
6600 neon_permissive_struct_operand instead of neon_struct_operand.
6601 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
6602 neon_vector_mem_operand.
6603 * config/arm/predicates.md (neon_struct_operand): Adjust call to
6604 neon_vector_mem_operand.
6605 (neon_permissive_struct_operand): New.
6606 (neon_struct_or_register_operand): Rename to...
6607 (neon_perm_struct_or_reg_operand): This. Adjust call to
6608 neon_vector_mem_operand.
6609
6610 2013-06-18 Richard Biener <rguenther@suse.de>
6611
6612 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
6613 * lto-streamer.h: Include pointer-set.h.
6614 (struct lto_decl_slot): Remove.
6615 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
6616 Remove next_index entry.
6617 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6618 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6619 (lto_init_tree_ref_encoder): Adjust.
6620 (lto_destroy_tree_ref_encoder): Likewise.
6621 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6622 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6623 (lto_output_decl_index): Adjust.
6624 (lto_new_out_decl_state): Likewise.
6625 (lto_record_function_out_decl_state): Likewise.
6626 * lto-streamer-out.c (copy_function): Likewise.
6627
6628 2013-06-18 Richard Biener <rguenther@suse.de>
6629
6630 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
6631 * cgraphunit.c: Include cfgloop.h.
6632 (init_lowered_empty_function): Initialize the loop tree.
6633 (assemble_thunk): Insert new BBs into loops.
6634
6635 2013-06-18 Richard Biener <rguenther@suse.de>
6636
6637 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
6638 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
6639 the map from cache entry to cache index optional.
6640 (streamer_tree_cache_replace_tree): Adjust accordingly.
6641 (streamer_tree_cache_append): Likewise.
6642 (streamer_tree_cache_delete): Likewise.
6643 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
6644 streamer cache map from cache entry to cache index.
6645 * lto-streamer-out.c (create_output_block): Adjust.
6646
6647 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
6648
6649 * config/arm/arm.md (attribute "insn"): Move multiplication and
6650 division attributes to...
6651 (attribute "type"): ... here. Remove mult.
6652 (attribute "mul32"): New attribute.
6653 (attribute "mul64"): Add umaal.
6654 (*arm_mulsi3): Update attributes.
6655 (*arm_mulsi3_v6): Likewise.
6656 (*thumb_mulsi3): Likewise.
6657 (*thumb_mulsi3_v6): Likewise.
6658 (*mulsi3_compare0): Likewise.
6659 (*mulsi3_compare0_v6): Likewise.
6660 (*mulsi_compare0_scratch): Likewise.
6661 (*mulsi_compare0_scratch_v6): Likewise.
6662 (*mulsi3addsi): Likewise.
6663 (*mulsi3addsi_v6): Likewise.
6664 (*mulsi3addsi_compare0): Likewise.
6665 (*mulsi3addsi_compare0_v6): Likewise.
6666 (*mulsi3addsi_compare0_scratch): Likewise.
6667 (*mulsi3addsi_compare0_scratch_v6): Likewise.
6668 (*mulsi3subsi): Likewise.
6669 (*mulsidi3adddi): Likewise.
6670 (*mulsi3addsi_v6): Likewise.
6671 (*mulsidi3adddi_v6): Likewise.
6672 (*mulsidi3_nov6): Likewise.
6673 (*mulsidi3_v6): Likewise.
6674 (*umulsidi3_nov6): Likewise.
6675 (*umulsidi3_v6): Likewise.
6676 (*umulsidi3adddi): Likewise.
6677 (*umulsidi3adddi_v6): Likewise.
6678 (*smulsi3_highpart_nov6): Likewise.
6679 (*smulsi3_highpart_v6): Likewise.
6680 (*umulsi3_highpart_nov6): Likewise.
6681 (*umulsi3_highpart_v6): Likewise.
6682 (mulhisi3): Likewise.
6683 (*mulhisi3tb): Likewise.
6684 (*mulhisi3bt): Likewise.
6685 (*mulhisi3tt): Likewise.
6686 (maddhisi4): Likewise.
6687 (*maddhisi4tb): Likewise.
6688 (*maddhisi4tt): Likewise.
6689 (maddhidi4): Likewise.
6690 (*maddhidi4tb): Likewise.
6691 (*maddhidi4tt): Likewise.
6692 (divsi3): Likewise.
6693 (udivsi3): Likewise.
6694 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
6695 (thumb2_mulsi_short_compare0): Likewise.
6696 (thumb2_mulsi_short_compare0_scratch): Likewise.
6697 * config/arm/arm1020e.md (1020mult1): Update attribute change.
6698 (1020mult2): Likewise.
6699 (1020mult3): Likewise.
6700 (1020mult4): Likewise.
6701 (1020mult5): Likewise.
6702 (1020mult6): Likewise.
6703 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
6704 change.
6705 (cortex_a15_mult64): Likewise.
6706 (cortex_a15_sdiv): Likewise.
6707 (cortex_a15_udiv): Likewise.
6708 * config/arm/arm1026ejs.md (mult1): Update attribute change.
6709 (mult2): Likewise.
6710 (mult3): Likewise.
6711 (mult4): Likewise.
6712 (mult5): Likewise.
6713 (mult6): Likewise.
6714 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
6715 (pj4_ir_div): Likewise.
6716 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
6717 (11_mult2): Likewise.
6718 (11_mult3): Likewise.
6719 (11_mult4): Likewise.
6720 (11_mult5): Likewise.
6721 (11_mult6): Likewise.
6722 (11_mult7): Likewise.
6723 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
6724 (cortex_a8_mla): Likewise.
6725 (cortex_a8_mull): Likewise.
6726 (cortex_a8_smulwy): Likewise.
6727 (cortex_a8_smlald): Likewise.
6728 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
6729 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
6730 (cortex_r4_mul_3): Likewise.
6731 (cortex_r4_mla_4): Likewise.
6732 (cortex_r4_mla_3): Likewise.
6733 (cortex_r4_smlald): Likewise.
6734 (cortex_r4_mull): Likewise.
6735 (cortex_r4_sdiv): Likewise.
6736 (cortex_r4_udiv): Likewise.
6737 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
6738 (cortex_a7_idiv): Likewise.
6739 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
6740 (9_mult2): Likewise.
6741 (9_mult3): Likewise.
6742 (9_mult4): Likewise.
6743 (9_mult5): Likewise.
6744 (9_mult6): Likewise.
6745 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
6746 (cortex_a53_sdiv): Likewise.
6747 (cortex_a53_udiv): Likewise.
6748 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
6749 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
6750 (mp626_mult2): Likewise.
6751 (mp626_mult3): Likewise.
6752 (mp626_mult4): Likewise.
6753 * config/arm/fa526.md (526_mult1): Update attribute change.
6754 (526_mult2): Likewise.
6755 * config/arm/arm-generic.md (mult): Update attribute change.
6756 (mult_ldsched_strongarm): Likewise.
6757 (mult_ldsched): Likewise.
6758 (multi_cycle): Likewise.
6759 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
6760 * config/arm/fa606te.md (606te_mult1): Update attribute change.
6761 (606te_mult2): Likewise.
6762 (606te_mult3): Likewise.
6763 (606te_mult4): Likewise.
6764 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
6765 (cortex_a9_mac16): Likewise.
6766 (cortex_a9_multiply): Likewise.
6767 (cortex_a9_mac): Likewise.
6768 (cortex_a9_multiply_long): Likewise.
6769 * config/arm/fa626te.md (626te_mult1): Update attribute change.
6770 (626te_mult2): Likewise.
6771 (626te_mult3): Likewise.
6772 (626te_mult4): Likewise.
6773
6774 2013-06-18 Richard Biener <rguenther@suse.de>
6775
6776 PR lto/57334
6777 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
6778
6779 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6780
6781 PR target/57609
6782 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6783 with next_active_insn.
6784
6785 2013-06-18 Alan Modra <amodra@gmail.com>
6786
6787 * config/rs6000/rs6000.h (enum data_align): New.
6788 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
6789 (DATA_ABI_ALIGNMENT): Define.
6790 (CONSTANT_ALIGNMENT): Correct comment.
6791 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
6792 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
6793
6794 2013-06-17 David Malcolm <dmalcolm@redhat.com>
6795
6796 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
6797 ATTRIBUTE_UNUSED marking.
6798
6799 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6800
6801 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6802 alternative and update.
6803 (aarch64_dup_lanedi): Delete.
6804 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6805 * config/aarch64/aarch64-simd-builtins.def: Update.
6806
6807 2013-06-17 Richard Biener <rguenther@suse.de>
6808
6809 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6810 (lto_input_scc): Declare.
6811 (lto_input_tree_1): Likewise.
6812 (struct lto_stats_d): Add num_tree_bodies_output and
6813 num_pickle_refs_output.
6814 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6815 (lto_read_tree_1): Split out from ...
6816 (lto_read_tree): ... this.
6817 (lto_input_scc): New function.
6818 (lto_input_tree_1): Split out from ...
6819 (lto_input_tree): ... this. Handle LTO_tree_scc.
6820 (lto_data_in_create): Create the streamer cache without hashes.
6821 * lto-streamer-out.c (create_output_block): Create the streamer
6822 cache with hashes when not doing WPA.
6823 (lto_write_tree_1): Split out from ...
6824 (lto_write_tree): ... this.
6825 (get_symbol_initial_value): New function.
6826 (lto_output_tree_1): Split out from ...
6827 (lto_output_tree): ... this. Write trees as series of SCCs
6828 using a DFS walk via DFS_write_tree.
6829 (struct sccs, struct scc_entry): New types.
6830 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6831 (DFS_write_tree_body): New function.
6832 (DFS_write_tree): Likewise.
6833 (hash_tree): Likewise.
6834 (scc_entry_compare): Likewise.
6835 (hash_scc): Likewise.
6836 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6837 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6838 TREE_CHAIN as regular reference.
6839 (streamer_read_integer_cst): Remove.
6840 (streamer_get_pickled_tree): Adjust.
6841 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6842 of DECL_EXTERNALs in BLOCK_VARS for now.
6843 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6844 reference.
6845 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6846 Add hash value argument and record that if hashes are recorded
6847 in the cache.
6848 (streamer_tree_cache_insert_1): Adjust.
6849 (streamer_tree_cache_insert): Likewise.
6850 (streamer_tree_cache_insert_at): Rename to ...
6851 (streamer_tree_cache_replace_tree): ... this and adjust.
6852 (streamer_tree_cache_append): Adjust.
6853 (record_common_node): Likewise.
6854 (streamer_tree_cache_create): Add argument whether to
6855 record hash values together with trees.
6856 (streamer_tree_cache_delete): Adjust.
6857 * tree-streamer.h (struct streamer_tree_cache_d): Add
6858 vector of hashes.
6859 (streamer_read_integer_cst): Remove.
6860 (streamer_tree_cache_insert): Adjust.
6861 (streamer_tree_cache_append): Likewise.
6862 (streamer_tree_cache_insert_at): Rename to ...
6863 (streamer_tree_cache_replace_tree): ... this and adjust.
6864 (streamer_tree_cache_create): Add argument whether to record hashes.
6865 (streamer_tree_cache_get): Rename to ...
6866 (streamer_tree_cache_get_tree): ... this.
6867 (streamer_tree_cache_get_hash): New function.
6868 * tree.c (cache_integer_cst): New function.
6869 * tree.h (cache_integer_cst): Declare.
6870 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6871 * lto-symtab.c (lto_varpool_replace_node): Only release
6872 DECL_INITIAL of non-prevailing decls.
6873 * varpool.c (varpool_remove_initializer): Do not release
6874 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6875
6876 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6877
6878 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6879 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6880 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6881 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6882 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6883 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6884 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6885 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6886 instead of TARGET_64BIT.
6887 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6888 Require ISA_HAS_<D>DIV.
6889
6890 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6891
6892 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6893 (mips*-*-linux*): Move default with_llsc setting to where other
6894 defaults are set.
6895 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6896 with_arch block.
6897 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6898 Likewise. Remove default with_tune setting. Move default float
6899 setting to its own block. Handle with_llsc in the same block as above.
6900
6901 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6902
6903 PR rtl-optimization/57425
6904 PR rtl-optimization/57569
6905 * alias.c (write_dependence_p): Add new parameters mem_mode,
6906 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6907 Changed all callers.
6908 (canon_anti_dependence): New function.
6909 * cse.c (check_dependence): Use canon_anti_dependence.
6910 * cselib.c (cselib_invalidate_mem): Likewise.
6911 * rtl.h (canon_anti_dependence): Declare.
6912
6913 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6914
6915 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6916 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6917 ".set mips3" for 64-bit targets.
6918
6919 2013-06-15 Dehao Chen <dehao@google.com>
6920
6921 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6922 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6923 (gimple_check_call_args): Likewise.
6924 * value-prof.c (check_ic_target): Likewise.
6925 * ipa-inline.c (early_inliner): Likewise.
6926 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6927 * cgraph.c (cgraph_create_edge_1): Likewise.
6928 (cgraph_make_edge_direct): Likewise.
6929
6930 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6931
6932 PR target/57615
6933 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6934 rs6000_output_move_128bit to handle emitting quad memory
6935 operations. Set attribute length to 8 bytes.
6936
6937 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6938
6939 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6940 New pattern.
6941 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6942 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6943 (aarch64_<su>mlsl<mode>): Likewise.
6944
6945 2013-06-14 Mike Stump <mikestump@comcast.net>
6946
6947 * Makefile.in (TARGET_H): Add insn-codes.h.
6948
6949 2013-06-14 Alan Modra <amodra@gmail.com>
6950
6951 PR middle-end/57134
6952 PR middle-end/57586
6953 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6954 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6955 bitfield expansion when EXPAND_MEMORY.
6956 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6957
6958 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6959
6960 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6961 test for clearing quad memory on 32-bit later.
6962
6963 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6964
6965 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6966 (fold_negate_expr): Likewise.
6967 (fold_real_zero_addition_p): Handle vectors.
6968 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6969
6970 2013-06-14 Alan Modra <amodra@gmail.com>
6971
6972 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6973
6974 2013-06-13 Jan Hubicka <jh@suse.cz>
6975
6976 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6977 Local comdats are not externally visible.
6978 * symtab.c (dump_symtab_base): Dump externally visible.
6979 (verify_symtab_base): Verify back links in the symtab hash.
6980
6981 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6982
6983 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6984 CONVERT_EXPR as equal nodes.
6985
6986 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6987
6988 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6989
6990 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6991
6992 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6993 Generalize to complex and vector.
6994 * tree.c (build_all_ones_cst): New function.
6995 * tree.h (build_all_ones_cst): Declare it.
6996
6997 2013-06-13 Alan Modra <amodra@gmail.com>
6998
6999 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
7000 * config/rs6000/rs6000.md (signbittf2): New insn.
7001 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
7002 (abstf2_internal, cmptf_internal2): Likewise.
7003 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
7004
7005 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7006 Pat Haugen <pthaugen@us.ibm.com>
7007 Peter Bergner <bergner@vnet.ibm.com>
7008
7009 * config/rs6000/rs6000.c (emit_load_locked): Add support for
7010 power8 byte, half-word, and quad-word atomic instructions.
7011 (emit_store_conditional): Likewise.
7012 (rs6000_expand_atomic_compare_and_swap): Likewise.
7013 (rs6000_expand_atomic_op): Likewise.
7014
7015 * config/rs6000/sync.md (larx): Add new modes for power8.
7016 (stcx): Likewise.
7017 (AINT): New mode iterator to include TImode as well as normal
7018 integer modes on power8.
7019 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
7020 that VSX registers are not considered. Use AINT mode iterator
7021 instead of INT1 to allow inclusion of quad word atomic operations
7022 on power8.
7023 (load_locked<mode>): Likewise.
7024 (store_conditional<mode>): Likewise.
7025 (atomic_compare_and_swap<mode>): Likewise.
7026 (atomic_exchange<mode>): Likewise.
7027 (atomic_nand<mode>): Likewise.
7028 (atomic_fetch_<fetchop_name><mode>): Likewise.
7029 (atomic_nand_fetch<mode>): Likewise.
7030 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7031 each type.
7032 (ATOMIC): On power8, add QImode, HImode modes.
7033 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7034 modes that promote to SImode.
7035 (load_lockedti): Convert TImode arguments to PTImode, so that we
7036 get a guaranteed even/odd register pair.
7037 (load_lockedpti): Likewise.
7038 (store_conditionalti): Likewise.
7039 (store_conditionalpti): Likewise.
7040
7041 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7042 atomic load/store instructions.
7043 (HSI): Likewise.
7044
7045 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7046
7047 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7048 loads.
7049 (insn_count): New attribute, with most cases extracted from...
7050 (length): ...here. Redefine most cases in terms of insn_count.
7051 (single_insn): Delete.
7052 (can_delay): Use insn_count to check for single instructions.
7053 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7054 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7055 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7056 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7057 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7058 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7059 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7060 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7061 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7062 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7063 rather than "length".
7064 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7065 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7066 Use "insn_count" rather than "length".
7067 * config/mips/mips-dsp.md
7068 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7069 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7070 length attributes.
7071
7072 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7073
7074 PR tree-optimization/57361
7075 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7076
7077 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7078
7079 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7080 to split.
7081 (aarch64_simd_combine<mode>): New instruction expansion.
7082 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7083 function prototype.
7084 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7085 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7086
7087 2013-06-12 Jan Hubicka <jh@suse.cz>
7088
7089 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7090 decl has when in streaming stage.
7091 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7092 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7093
7094 2013-06-12 Roland Stigge <stigge@antcom.de>
7095
7096 PR target/57578
7097 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7098
7099 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7100
7101 PR tree-optimization/57537
7102 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7103 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7104
7105 2013-06-12 Richard Biener <rguenther@suse.de>
7106
7107 * data-streamer.h (streamer_write_char_stream): CSE
7108 obs->current_pointer.
7109 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7110 streamer_write_char_stream manually and optimize the resulting loop.
7111 (streamer_write_hwi_stream): Likewise.
7112
7113 2013-06-12 Jan Hubicka <jh@suse.cz>
7114
7115 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7116 * cgraph.h (varpool_create_empty_node): Declare.
7117 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7118 duplicated nodes.
7119 * symtab.c (symtab_unregister_node): Be lax about missin entries
7120 in node hash.
7121 (symtab_get_node): Update comment.
7122 * varpool.c (varpool_create_empty_node): Break out from ...
7123 (varpool_node_for_decl): ... here.
7124 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7125
7126 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7127
7128 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7129 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7130 part. Use straight-line flow at the end.
7131 <COMPONENT_REF>: Remove superfluous else.
7132 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7133
7134 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7135
7136 PR target/56564
7137 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7138 target hook even for !TREE_PUBLIC decls. If no resolution info
7139 is available, return false for common and external decls.
7140
7141 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7142
7143 * config/rl78/constraints.md (U): New constraint.
7144 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7145 valloc attribute.
7146
7147 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7148
7149 PR target/57589
7150 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7151 to allow returning address to AT_PLATFORM name.
7152
7153 2013-06-11 Jan Hubicka <jh@suse.cz>
7154
7155 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7156 * cgraph.h (symtab_node_base): Add weakref flag.
7157 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7158 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7159 (output_weakrefs): Use weakref flag.
7160 * fold-const.c (simple_operand_p): Handle WEAK.
7161 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7162 * ipa.c (varpool_externally_visible_p): Drop weakref.
7163 (function_and_variable_visibility): Update comment; fix weakref
7164 sanity checks; do not clear DECL_WEAK on them.
7165 * lto-cgraph.c (lto_output_node): update.
7166 (lto_output_varpool_node): Update.
7167 (input_overwrite_node): Update.
7168 (input_node): Update.
7169 (input_varpool_node): Update.
7170 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7171 (lto_symtab_merge_symbols): Add sanity check.
7172 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7173 * passes.c (rest_of_decl_compilation): Set static flag, too.
7174 * symtab.c (dump_symtab_base): Dump weakref.
7175 (verify_symtab_base): Sanity check weakrefs.
7176 (symtab_make_decl_local): Remove duplicated code.
7177 (symtab_alias_ultimate_target): Simplify.
7178 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7179
7180 2013-06-11 DJ Delorie <dj@redhat.com>
7181
7182 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7183 (rl78_unwind_word_mode): New.
7184
7185 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7186
7187 * final.c (debug_prefix_maps): Make static.
7188
7189 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7190
7191 * function.c (initial_trampoline): Remove stray copy.
7192
7193 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
7194
7195 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
7196
7197 2013-06-11 Martin Jambor <mjambor@suse.cz>
7198
7199 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
7200 within bounds at the beginning of the function.
7201
7202 2013-06-11 Alan Modra <amodra@gmail.com>
7203
7204 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
7205 reporting.
7206 (get_named_section): Don't NULL !DECL_P decl.
7207
7208 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
7209
7210 * doc/invoke.texi (core-avx2): Document.
7211 (slm): Likewise.
7212 (atom): Updated with MOVBE.
7213
7214 2013-06-11 Richard Biener <rguenther@suse.de>
7215
7216 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
7217
7218 2013-06-11 Anton Blanchard <anton@samba.org>
7219
7220 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
7221 correct shift value in little-endian mode.
7222
7223 2013-06-11 Jakub Jelinek <jakub@redhat.com>
7224
7225 PR target/56564
7226 * varasm.c (get_variable_align): Move #endif to the right place.
7227
7228 2013-06-10 Cary Coutant <ccoutant@google.com>
7229
7230 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
7231 for hash so that hash table traversal order is deterministic.
7232
7233 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
7234 Pat Haugen <pthaugen@us.ibm.com>
7235 Peter Bergner <bergner@vnet.ibm.com>
7236
7237 * config/rs6000/vector.md (GPR move splitter): Do not split moves
7238 of vectors in GPRS if they are direct moves or quad word load or
7239 store moves.
7240
7241 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
7242 declaration.
7243 (direct_move_p): Likewise.
7244 (quad_load_store_p): Likewise.
7245
7246 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
7247 classes into bins based on the physical register type.
7248 (reg_class_to_reg_type): Likewise.
7249 (IS_STD_REG_TYPE): Likewise.
7250 (IS_FP_VECT_REG_TYPE): Likewise.
7251 (reload_fpr_gpr): Arrays to determine what insn to use if we can
7252 use direct move instructions.
7253 (reload_gpr_vsx): Likewise.
7254 (reload_vsx_gpr): Likewise.
7255 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
7256 information that is a simplification of register classes. Also
7257 precalculate direct move reload helpers.
7258 (direct_move_p): New function to return true if the operation can
7259 be done as a direct move instruciton.
7260 (quad_load_store_p): New function to return true if the operation
7261 is a quad memory operation.
7262 (rs6000_legitimize_address): If quad memory, only allow register
7263 indirect for TImode addresses.
7264 (rs6000_legitimate_address_p): Likewise.
7265 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
7266 (rs6000_reload_register_type): Likewise.
7267 (register_to_reg_type): Return register type.
7268 (rs6000_secondary_reload_simple_move): New helper function for
7269 secondary reload and secondary memory needed to identify anything
7270 that is a simple move, and does not need reloading.
7271 (rs6000_secondary_reload_direct_move): New helper function for
7272 secondary reload to identify cases that can be done with several
7273 instructions via the direct move instructions.
7274 (rs6000_secondary_reload_move): New helper function for secondary
7275 reload to identify moves between register types that can be done.
7276 (rs6000_secondary_reload): Add support for quad memory operations
7277 and for direct move.
7278 (rs6000_secondary_memory_needed): Likewise.
7279 (rs6000_debug_secondary_memory_needed): Change argument names.
7280 (rs6000_output_move_128bit): New function to return the move to
7281 use for 128-bit moves, including knowing about the various
7282 limitations of quad memory operations.
7283
7284 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
7285 memory operations. call rs6000_output_move_128bit for the actual
7286 instruciton(s) to generate.
7287 (vsx_movti_64bit): Likewise.
7288
7289 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
7290 (UNSPEC_P8V_MTVSRWZ): Likewise.
7291 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
7292 (UNSPEC_P8V_MTVSRD): Likewise.
7293 (UNSPEC_P8V_XXPERMDI): Likewise.
7294 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
7295 (UNSPEC_FUSION_GPR): Likewise.
7296 (FMOVE128_GPR): New iterator for direct move.
7297 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
7298 (f32_sv): Likewise.
7299 (f32_dm): Likewise.
7300 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
7301 loads and direct move instructions.
7302 (zero_extendsidi2_lfiwzx): Likewise.
7303 (extendsidi2_lfiwax): Likewise.
7304 (extendsidi2_nocell): Likewise.
7305 (floatsi<mode>2_lfiwax): Likewise.
7306 (lfiwax): Likewise.
7307 (floatunssi<mode>2_lfiwzx): Likewise.
7308 (lfiwzx): Likewise.
7309 (fix_trunc<mode>_stfiwx): Likewise.
7310 (fixuns_trunc<mode>_stfiwx): Likewise.
7311 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
7312 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
7313 (parity<mode>2_cmpb): Set length/type attr.
7314 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
7315 for 'mr.' to fast_compare.
7316 (bpermd_<mode>): Change type attr to popcnt.
7317 (p8_fmrgow_<mode>): New insns for power8 direct move support.
7318 (p8_mtvsrwz_1): Likewise.
7319 (p8_mtvsrwz_2): Likewise.
7320 (reload_fpr_from_gpr<mode>): Likewise.
7321 (p8_mtvsrd_1): Likewise.
7322 (p8_mtvsrd_2): Likewise.
7323 (p8_xxpermdi_<mode>): Likewise.
7324 (reload_vsx_from_gpr<mode>): Likewise.
7325 (reload_vsx_from_gprsf): Likewise.
7326 (p8_mfvsrd_3_<mode>): LIkewise.
7327 (reload_gpr_from_vsx<mode>): Likewise.
7328 (reload_gpr_from_vsxsf): Likewise.
7329 (p8_mfvsrd_4_disf): Likewise.
7330 (multi-word GPR splits): Do not split direct moves or quad memory
7331 operations.
7332
7333 2013-06-10 David Malcolm <dmalcolm@redhat.com>
7334
7335 * tree-into-ssa.c (interesting_blocks): Make static.
7336
7337 2013-06-10 Jakub Jelinek <jakub@redhat.com>
7338
7339 PR target/56564
7340 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
7341 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
7342 Use DATA_ABI_ALIGNMENT for that case instead if defined.
7343 (get_variable_align): New function.
7344 (get_variable_section, emit_bss, emit_common,
7345 assemble_variable_contents, place_block_symbol): Use
7346 get_variable_align instead of DECL_ALIGN.
7347 (assemble_noswitch_variable): Add align argument, use it
7348 instead of DECL_ALIGN.
7349 (assemble_variable): Adjust caller. Use get_variable_align
7350 instead of DECL_ALIGN.
7351 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
7352 caller.
7353 (DATA_ABI_ALIGNMENT): Define.
7354 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
7355 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
7356 opt is false, only return the psABI mandated alignment increase.
7357 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
7358 (DATA_ABI_ALIGNMENT): ... this.
7359 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
7360 (DATA_ABI_ALIGNMENT): ... this.
7361 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
7362 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
7363 (DATA_ABI_ALIGNMENT): ... this.
7364 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
7365 * doc/tm.texi: Regenerated.
7366
7367 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
7368
7369 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
7370 cmp_code to construct REG_EQUAL note.
7371
7372 2013-06-09 Jakub Jelinek <jakub@redhat.com>
7373
7374 PR target/57568
7375 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
7376 that operands[2] doesn't overlap with operands[0].
7377
7378 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
7379 Jan Hubicka <jh@suse.cz>
7380
7381 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
7382 hack to mark symbols as used.
7383
7384 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
7385
7386 PR rtl-optimization/57559
7387 * lra-constraints.c (process_alt_operands): Don't discourage
7388 memory with known offset for offsetable memory constraint.
7389 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
7390
7391 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
7392
7393 * varasm.c (struct oc_local_state): Reorder fields.
7394 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
7395 and adjust accordingly.
7396 (output_constructor): Reorder initialization code and adjust call to
7397 output_constructor_bitfield.
7398
7399 2013-06-07 Jan Hubicka <jh@suse.cz>
7400
7401 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
7402
7403 2013-06-07 David Malcolm <dmalcolm@redhat.com>
7404
7405 * tree-object-size.c (unknown): Make const.
7406
7407 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7408
7409 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
7410 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
7411 for last alternative in the cpu_facility attribute.
7412
7413 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7414
7415 PR target/56315
7416 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
7417 (xordi3): Change operand 2 constraint to arm_xordi_operand.
7418 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
7419 * config/arm/constraints.md (Dg): New constraint.
7420 * config/arm/neon.md (xordi3_neon): Remove.
7421 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
7422 * config/arm/predicates.md (arm_xordi_operand): New predicate.
7423
7424 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7425
7426 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
7427 Clean up alternatives.
7428
7429 2013-06-07 Alan Modra <amodra@gmail.com>
7430
7431 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
7432 va_list_gpr_size.
7433
7434 2013-06-07 Alan Modra <amodra@gmail.com>
7435
7436 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
7437
7438 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7439
7440 * config/arm/constraints.md (Df): New constraint.
7441 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
7442 Correct length attribute for last two alternatives.
7443
7444 2013-06-07 Alan Modra <amodra@gmail.com>
7445
7446 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7447 override user -mfp-in-toc.
7448 (offsettable_ok_by_alignment): Consider just the current access
7449 rather than the whole object, unless BLKmode. Handle
7450 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
7451 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
7452 for -mcmodel=medium.
7453 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
7454 override user -mfp-in-toc or -msum-in-toc. Default to
7455 -mno-fp-in-toc for -mcmodel=medium.
7456
7457 2013-06-06 DJ Delorie <dj@redhat.com>
7458
7459 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
7460 TARGET_VALID_POINTER_MODE.
7461
7462 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
7463 Pat Haugen <pthaugen@us.ibm.com>
7464 Peter Bergner <bergner@vnet.ibm.com>
7465
7466 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7467 Document new power8 builtins.
7468
7469 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
7470 condition code register, to allow 128-bit logical operations to be
7471 done in the VSX or GPR registers.
7472 (nor<mode>3): Use the canonical form for nor.
7473 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
7474 vclz*, and vpopcnt* vector instructions.
7475 (nand<mode>3): Likewise.
7476 (orc<mode>3): Likewise.
7477 (clz<mode>2): LIkewise.
7478 (popcount<mode>2): Likewise.
7479
7480 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
7481 that only the GPRs are recognized.
7482
7483 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7484 support for new power8 builtins.
7485
7486 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
7487 builtin functions.
7488 (xscvdpspn): Likewise.
7489 (vclz): Likewise.
7490 (vclzb): Likewise.
7491 (vclzh): Likewise.
7492 (vclzw): Likewise.
7493 (vclzd): Likewise.
7494 (vpopcnt): Likewise.
7495 (vpopcntb): Likewise.
7496 (vpopcnth): Likewise.
7497 (vpopcntw): Likewise.
7498 (vpopcntd): Likewise.
7499 (vgbbd): Likewise.
7500 (vmrgew): Likewise.
7501 (vmrgow): Likewise.
7502 (eqv): Likewise.
7503 (eqv_v16qi3): Likewise.
7504 (eqv_v8hi3): Likewise.
7505 (eqv_v4si3): Likewise.
7506 (eqv_v2di3): Likewise.
7507 (eqv_v4sf3): Likewise.
7508 (eqv_v2df3): Likewise.
7509 (nand): Likewise.
7510 (nand_v16qi3): Likewise.
7511 (nand_v8hi3): Likewise.
7512 (nand_v4si3): Likewise.
7513 (nand_v2di3): Likewise.
7514 (nand_v4sf3): Likewise.
7515 (nand_v2df3): Likewise.
7516 (orc): Likewise.
7517 (orc_v16qi3): Likewise.
7518 (orc_v8hi3): Likewise.
7519 (orc_v4si3): Likewise.
7520 (orc_v2di3): Likewise.
7521 (orc_v4sf3): Likewise.
7522 (orc_v2df3): Likewise.
7523
7524 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
7525 allow power8 quad mode in 64-bit.
7526 (rs6000_builtin_vectorized_function): Add support to vectorize
7527 ISA 2.07 count leading zeros, population count builtins.
7528 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
7529 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
7530 (builtin_function_type): Add vgbbd builtin function which takes an
7531 unsigned argument.
7532 (altivec_expand_vec_perm_const): Add support for new power8 merge
7533 instructions.
7534
7535 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
7536 that does not include TImdoe for use with 32-bit.
7537 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
7538 instructions.
7539 (UNSPEC_VSX_CVDPSPN): Likewise.
7540 (vsx_xscvdpspn): Likewise.
7541 (vsx_xscvspdpn): Likewise.
7542 (vsx_xscvdpspn_scalar): Likewise.
7543 (vsx_xscvspdpn_directmove): Likewise.
7544 (vsx_and<mode>3): Split logical operations into 32-bit and
7545 64-bit. Add support to do logical operations on TImode as well as
7546 VSX vector types. Allow logical operations to be done in either
7547 VSX registers or in general purpose registers in 64-bit mode. Add
7548 splitters if GPRs were used. For AND, add clobber of CCmode to
7549 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
7550 encoding.
7551 (vsx_and<mode>3_32bit): Likewise.
7552 (vsx_and<mode>3_64bit): Likewise.
7553 (vsx_ior<mode>3): Likewise.
7554 (vsx_ior<mode>3_32bit): Likewise.
7555 (vsx_ior<mode>3_64bit): Likewise.
7556 (vsx_xor<mode>3): Likewise.
7557 (vsx_xor<mode>3_32bit): Likewise.
7558 (vsx_xor<mode>3_64bit): Likewise.
7559 (vsx_one_cmpl<mode>2): Likewise.
7560 (vsx_one_cmpl<mode>2_32bit): Likewise.
7561 (vsx_one_cmpl<mode>2_64bit): Likewise.
7562 (vsx_nor<mode>3): Likewise.
7563 (vsx_nor<mode>3_32bit): Likewise.
7564 (vsx_nor<mode>3_64bit): Likewise.
7565 (vsx_andc<mode>3): Likewise.
7566 (vsx_andc<mode>3_32bit): Likewise.
7567 (vsx_andc<mode>3_64bit): Likewise.
7568 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
7569 and xxlorc instructions.
7570 (vsx_eqv<mode>3_64bit): Likewise.
7571 (vsx_nand<mode>3_32bit): Likewise.
7572 (vsx_nand<mode>3_64bit): Likewise.
7573 (vsx_orc<mode>3_32bit): Likewise.
7574 (vsx_orc<mode>3_64bit): Likewise.
7575
7576 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
7577
7578 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
7579 instruction.
7580 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
7581 (p8_vmrgow): Likewise.
7582 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
7583 GPRs to be split under VSX.
7584 (p8v_clz<mode>2): Add power8 count leading zero support.
7585 (p8v_popcount<mode>2): Add power8 population count support.
7586 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
7587 support.
7588
7589 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
7590 instruction.
7591
7592 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
7593 builtin functions.
7594 (vec_nand): Likewise.
7595 (vec_vclz): Likewise.
7596 (vec_vclzb): Likewise.
7597 (vec_vclzd): Likewise.
7598 (vec_vclzh): Likewise.
7599 (vec_vclzw): Likewise.
7600 (vec_vgbbd): Likewise.
7601 (vec_vmrgew): Likewise.
7602 (vec_vmrgow): Likewise.
7603 (vec_vpopcnt): Likewise.
7604 (vec_vpopcntb): Likewise.
7605 (vec_vpopcntd): Likewise.
7606 (vec_vpopcnth): Likewise.
7607 (vec_vpopcntw): Likewise.
7608
7609 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7610
7611 PR rtl-optimization/57468
7612 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
7613 spilled pseudos.
7614
7615 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7616
7617 PR rtl-optimization/57459
7618 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
7619 type when setting live regs.
7620
7621 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7622
7623 * config/s390/s390.opt (mlra): New option.
7624 * config/s390/s390.c (s390_decompose_address): Check displacement
7625 for all registers for LRA.
7626 (s390_secondary_reload): Don't used secondary reloads for LRA.
7627 (s390_lra_p): New function.
7628 (TARGET_LRA_P): Define.
7629 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
7630 of attribute cpu_facility to zarch for the last alternative.
7631 (*cmpmem_short): Ditto.
7632
7633 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
7634
7635 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
7636 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
7637 (arm_expand_prologue): Likewise.
7638
7639 2013-06-06 Teresa Johnson <tejohnson@google.com>
7640
7641 PR c++/53743
7642 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
7643 as this is now done by redirect_edge_and_branch_force.
7644 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
7645 barriers, and fix interaction with splitting.
7646 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
7647 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
7648 reflect changes made in the routine.
7649 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
7650 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
7651 since this is called in cfglayout mode, and replace partition fixup
7652 with assert as that is now done by force_nonfallthru_and_redirect.
7653 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
7654 already be marked with region crossing note.
7655 (insert_section_boundary_note): Make non-static, gate on flag
7656 has_bb_partition, rewrite to also check for multiple partitions.
7657 (rest_of_handle_reorder_blocks): Remove call to
7658 insert_section_boundary_note, now done later during free_cfg.
7659 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
7660 * bb-reorder.h (insert_section_boundary_note): Declare.
7661 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
7662 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
7663 invoke insert_section_boundary_note.
7664 (try_redirect_by_replacing_jump): Remove unnecessary
7665 check for region crossing note.
7666 (fixup_partition_crossing): New function.
7667 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
7668 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
7669 in non-cfglayout mode.
7670 (force_nonfallthru_and_redirect): Fixup partition boundaries,
7671 remove old code that tried to do this. Emit barrier correctly
7672 when we are in cfglayout mode.
7673 (last_bb_in_partition): New function.
7674 (rtl_split_edge): Correctly fixup partition boundaries.
7675 (commit_one_edge_insertion): Remove old code that tried to
7676 fixup region crossing edge since this is now handled in
7677 split_block, and set up insertion point correctly since
7678 block may now end in a jump.
7679 (verify_hot_cold_block_grouping): Guard against checking when not in
7680 linearized RTL mode.
7681 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
7682 notes.
7683 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
7684 rtl_verify_flow_info, so not called in cfglayout mode.
7685 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
7686 (fixup_reorder_chain): Remove old code that attempted to fixup region
7687 crossing note as this is now handled in force_nonfallthru_and_redirect.
7688 (duplicate_insn_chain): Don't duplicate switch section notes.
7689 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
7690 note.
7691 * basic-block.h (emit_barrier_after_bb): Declare.
7692
7693 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7694
7695 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
7696 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
7697 arm_usatsihi): Adjust alternatives for arm_restrict_it.
7698
7699 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7700
7701 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
7702 where appropriate.
7703 * config/arm/ldmstm.md: Regenerate.
7704
7705 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7706
7707 * config/arm/sync.md (atomic_loaddi_1):
7708 Disable predication for arm_restrict_it.
7709 (arm_load_exclusive<mode>): Likewise.
7710 (arm_load_exclusivesi): Likewise.
7711 (arm_load_exclusivedi): Likewise.
7712 (arm_load_acquire_exclusive<mode>): Likewise.
7713 (arm_load_acquire_exclusivesi): Likewise.
7714 (arm_load_acquire_exclusivedi): Likewise.
7715 (arm_store_exclusive<mode>): Likewise.
7716 (arm_store_exclusive<mode>): Likewise.
7717 (arm_store_release_exclusivedi): Likewise.
7718 (arm_store_release_exclusive<mode>): Likewise.
7719
7720 2013-06-06 Richard Biener <rguenther@suse.de>
7721
7722 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
7723 after LTO_null.
7724 (lto_tag_is_tree_code_p): Adjust.
7725 (lto_tag_is_gimple_code_p): Likewise.
7726 (lto_gimple_code_to_tag): Likewise.
7727 (lto_tag_to_gimple_code): Likewise.
7728 (lto_tree_code_to_tag): Likewise.
7729 (lto_tag_to_tree_code): Likewise.
7730 * data-streamer.h (streamer_write_hwi_in_range): Use
7731 uhwi streaming to stream the normalized range.
7732 (streamer_read_hwi_in_range): Likewise.
7733
7734 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7735
7736 * config/arm/arm.md (enabled_for_depr_it): New attribute.
7737 (predicable_short_it): Likewise.
7738 (predicated): Likewise.
7739 (enabled): Handle above.
7740 (define_cond_exec): Set predicated attribute to yes.
7741
7742 2013-06-05 Mike Stump <mikestump@comcast.net>
7743
7744 * gdbinit.in (__FUNCTION__): Add.
7745
7746 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
7747
7748 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
7749 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
7750
7751 2013-06-05 Jan Hubicka <jh@suse.cz>
7752
7753 * varasm.c (mark_decl_referenced): Revert the removal until targets
7754 are fixed.
7755
7756 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
7757
7758 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
7759 instead of mark_decl_referenced.
7760
7761 2013-06-05 Jan Hubicka <jh@suse.cz>
7762
7763 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
7764 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
7765 and symtab_used_from_object_file_p.
7766 (cgraph_make_node_local_1): Clear forced_by_abi.
7767 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
7768 * cgraph.h (symtab_node_base): Add forced_by_abi.
7769 (decide_is_variable_needed): Remove.
7770 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
7771 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
7772 (decide_is_symbol_needed): ... this one; handle symbols in general;
7773 always analyze virtuals; honnor forced_by_abi.
7774 (cgraph_finalize_function): Update.
7775 (varpool_finalize_decl): Update.
7776 (symbol_defined_and_needed): Remove.
7777 (analyze_functions): Update.
7778 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
7779 output_refs, input_overwrite_node): Handle forced_by_abi.
7780 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
7781 (address_taken_from_non_vtable_p): ... this one.
7782 (comdat_can_be_unshared_p_1): New function.
7783 (cgraph_comdat_can_be_unshared_p): Rename to ...
7784 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
7785 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
7786 (function_and_variable_visibility): Clear forced_by_abi as needed.
7787 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
7788 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
7789 * symtab.c (dump_symtab_base): Dump forced_by_abi.
7790 * varpool.c (decide_is_variable_needed): Remove.
7791
7792 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7793
7794 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
7795 (arm_option_override): Override arm_restrict_it where appropriate.
7796 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
7797 * config/arm/arm.opt (mrestrict-it): New command-line option.
7798 * doc/invoke.texi: Document -mrestrict-it.
7799
7800 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7801
7802 * tsan.c (tsan_atomic_table): Make const.
7803
7804 2013-06-05 Richard Biener <rguenther@suse.de>
7805
7806 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7807 index associated with the tree we are supposed to replace.
7808 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7809 TREE_ASM_WRITTEN as zero for everything but SSA names.
7810
7811 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7812
7813 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7814
7815 2013-06-04 Jan Hubicka <jh@suse.cz>
7816
7817 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7818 (input_node, input_varpool_node): Handle correctly external same
7819 body aliases.
7820 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7821 nodes at ltrans stage.
7822
7823 2013-06-04 Jan Hubicka <jh@suse.cz>
7824
7825 * ipa-inline.c (update_caller_keys): Fix availability test.
7826 (update_callee_keys): Likewise.
7827 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7828 to follow ELF standard.
7829
7830 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7831
7832 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7833 (mips64r5900el-*-elf*): New configurations.
7834 * config/mips/mips-cpus.def (r5900): New processor.
7835 * config/mips/mips-tables.opt: Regenerate.
7836 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7837 (mips_issue_rate): Handle PROCESSOR_R5900.
7838 (mips_reorg_process_insns): Force reorder mode for the R5900.
7839 * config/mips/mips.h (TARGET_MIPS5900): Define.
7840 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7841 TARGET_MIPS5900.
7842 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7843 TARGET_MIPS5900.
7844 * config/mips/mips.md (processor): Add r5900.
7845 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7846
7847 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7848
7849 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7850 into function to generate MOVI instruction.
7851 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7852 (aarch64_preferred_simd_mode): Turn into wrapper.
7853 (aarch64_output_scalar_simd_mov_immediate): New function.
7854 * config/aarch64/aarch64-protos.h: Add prototype for above.
7855
7856 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7857
7858 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7859 element_char member.
7860 (sizetochar): Return signed char.
7861 (aarch64_simd_valid_immediate): Remove elchar and other
7862 unnecessary variables.
7863 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7864 Calculate element_char as required.
7865 * config/aarch64/aarch64-protos.h: Update and move prototype
7866 for aarch64_output_simd_mov_immediate.
7867 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7868 Update arguments.
7869
7870 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7871
7872 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7873 information completed by aarch64_simd_valid_immediate.
7874 (aarch64_legitimate_constant_p): Update arguments.
7875 (aarch64_simd_valid_immediate): Work with struct rather than many
7876 pointers.
7877 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7878 (aarch64_simd_make_constant): Update arguments.
7879 (aarch64_output_simd_mov_immediate): Work with struct rather than
7880 many pointers. Output immediate directly rather than as operand.
7881 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7882 Update prototype.
7883 * config/aarch64/constraints.md (Dn): Update arguments.
7884
7885 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7886
7887 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7888 longer static.
7889 (aarch64_simd_immediate_valid_for_move): Remove.
7890 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7891 (aarch64_simd_make_constant): Update call.
7892 (aarch64_output_simd_mov_immediate): Update call.
7893 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7894 Add prototype.
7895 * config/aarch64/constraints.md (Dn): Update call.
7896
7897 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7898
7899 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7900 return type to bool for prototype.
7901 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7902 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7903 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7904
7905 2013-06-04 Catherine Moore <clm@codesourcery.com>
7906
7907 * config/mips/mips.opt (meva): New.
7908 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7909 (ASM_SPEC): Handle -meva.
7910 * doc/invoke.texi (meva): Document.
7911
7912 2013-06-04 Alan Modra <amodra@gmail.com>
7913
7914 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7915 constant output.
7916
7917 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7918
7919 * rtl.def: Add extra fourth optional field to define_cond_exec.
7920 * gensupport.c (process_one_cond_exec): Process attributes from
7921 define_cond_exec.
7922 * doc/md.texi: Document fourth field in define_cond_exec.
7923
7924 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7925
7926 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7927 out the processing order as in store_bit_field_1.
7928
7929 2013-06-04 Jan Hubicka <jh@suse.cz>
7930
7931 PR middle-end/57500
7932 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7933 non-VAR_DECL node if it does not exist yet.
7934
7935 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7936
7937 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7938 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7939 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7940 target_cpu_default setting.
7941
7942 2013-06-03 Teresa Johnson <tejohnson@google.com>
7943
7944 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7945 default to -fopt-info=optimized instead of all.
7946 * doc/invoke.texi: Ditto.
7947 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7948 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7949 (execute_vect_slp): Emit BB vectorization success under
7950 MSG_OPTIMIZED_LOCATIONS.
7951 * tree-vect-slp.c (vect_slp_transform_bb): Change
7952 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7953 * tree-vect-loop.c (vect_transform_loop): Ditto.
7954
7955 2013-06-03 Jason Merrill <jason@redhat.com>
7956
7957 PR c++/57415
7958 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7959 Use TARGET_EXPR for C++.
7960
7961 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7962
7963 PR rtl-optimization/57268
7964 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7965 if DEBUG_INSN_P (insn).
7966
7967 Reapply
7968 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7969
7970 PR rtl-optimization/57268
7971 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7972 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7973
7974 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7975
7976 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7977 (ix86_avoid_lea_for_addr): Likewise.
7978 (exact_dependency_1): Likewise.
7979 (ix86_adjust_cost): Likewise.
7980 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7981 removed.
7982 (do_reorder_for_imul): Fix typo, formatting and
7983 !reload_completed check removed.
7984 (ix86_sched_reorder): Fix typo and formatting.
7985 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7986 list.
7987
7988 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7989
7990 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7991
7992 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7993
7994 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7995 <STRING_CST>: Likewise.
7996 <VECTOR_CST>: Likewise.
7997
7998 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7999 Mikael Morin <mikael@gcc.gnu.org>
8000
8001 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
8002 * config.in: Regenerated.
8003 * configure: Regenerated.
8004
8005 2013-06-01 Jan Hubicka <jh@suse.cz>
8006
8007 PR middle-end/57366
8008 * cgraphunit.c (compile): When weakref is not supported,
8009 set up transparent aliases before final output pass.
8010 * varasm.c (assemble_alias): Do not try to do it here.
8011
8012 2013-06-01 Jan Hubicka <jh@suse.cz>
8013
8014 PR middle-end/57467
8015 * passes.c (for_per_function): Skip unanalyzed functions.
8016
8017 2013-06-01 Jan Hubicka <jh@suse.cz>
8018
8019 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
8020 (lto_symtab_merge_symbols_1): ... this one.
8021 (lto_symtab_merge_cgraph_nodes): Rename to ...
8022 (lto_symtab_merge_symbols): ... this one; simplify.
8023 * cgraph.c (same_body_aliases_done): Rename to ...
8024 (cpp_implicit_aliases_done): ... this one.
8025 (cgraph_create_function_alias): Update.
8026 (cgraph_same_body_alias): Update.
8027 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8028 (verify_edge_corresponds_to_fndecl): Simplify.
8029 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8030 (cgraph_node): Remove same_body_alias.
8031 (varpool_node): Remove alias_of and extra_name_alias.
8032 (same_body_aliases_done): Rename to ..
8033 (cpp_implicit_aliases_done): ... this one.
8034 (symtab_alias_ultimate_target): Add default parameter.
8035 (symtab_resolve_alias): New function.
8036 (fixup_same_cpp_alias_visibility): Declare.
8037 (cgraph_function_node): Add default parameter.
8038 (cgraph_node_asm_name): Likewise.
8039 (cgraph_function_or_thunk_node): Add default parameter; do
8040 not ICE when it is NULL.
8041 (varpool_variable_node): Likewise.
8042 * tree-emutls.c (create_emultls_var): Update.
8043 (ipa_lower_emutls): Update.
8044 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8045 (cgraph_reset_node): Reset alias info.
8046 (cgraph_finalize_function): Update.
8047 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8048 (analyze_function): Simplify.
8049 (cgraph_process_same_body_aliases): Simplify.
8050 (analyze_functions): Fixup same body aliases.
8051 (handle_alias_pairs): Simplify.
8052 (assemble_thunk): Update.
8053 (assemble_thunks_and_aliases): Update.
8054 (output_weakrefs): Rewrite.
8055 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8056 (lto_output_varpool_node): Likewise.
8057 (compute_ltrans_boundary): Remve assert.
8058 (get_alias_symbol): New functoin.
8059 (input_node): Rewrite alias handling.
8060 (input_varpool_node): Likewise.
8061 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8062 * ipa.c (process_references): Handle weakrefs correctly.
8063 (symtab_remove_unreachable_nodes): Likewise.
8064 * trans-mem.c (get_cg_data): Update.
8065 (ipa_tm_create_version_alias): Update.
8066 (ipa_tm_execute): Update.
8067 * symtab.c (dump_symtab_base): Dump aliases.
8068 (verify_symtab_base): Verify aliases.
8069 (symtab_node_availability): New function.
8070 (symtab_alias_ultimate_target): Simplify.
8071 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8072 handle all the fixup cases.
8073 (symtab_resolve_alias): New function.
8074 * passes.c (ipa_write_summaries): Handle weakrefs.
8075 * varpool.c (varpool_analyze_node): Simplify.
8076 (assemble_aliases): Update.
8077 (varpool_create_variable_alias): Simplify.
8078 (varpool_extra_name_alias): Simplify.
8079 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8080 (lto_symtab_merge_symbols): ... this one.
8081
8082 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8083
8084 Revert
8085 PR rtl-optimization/57268
8086 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8087 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8088
8089 2013-06-01 Tobias Burnus <burnus@net-b.de>
8090
8091 Partially reverted:
8092 2013-05-31 Tobias Burnus <burnus@net-b.de>
8093
8094 PR middle-end/57073
8095 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8096 further up.
8097
8098 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8099
8100 PR rtl-optimization/57268
8101 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8102 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8103
8104 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8105
8106 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8107 unordered comparison operators when -fno-trapping-math is in effect
8108 on the e500.
8109 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8110 and implement unordered comparison operators properly on the e500.
8111
8112 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8113
8114 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8115 for constant scalar integers.
8116 (simplify_relational_operation_1): Likewise.
8117
8118 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8119
8120 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8121 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8122 Fix comment.
8123
8124 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8125 Igor Zamyatin <igor.zamyatin@intel.com>
8126
8127 Silvermont (SLM) architecture performance tuning.
8128 * config/i386/i386.h (enum ix86_tune_indices): Add
8129 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8130 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8131
8132 * config/i386/i386.c (initial_ix86_tune_features)
8133 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8134 (ix86_lea_outperforms): Handle Silvermont tuning.
8135 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8136 call.
8137 (ix86_use_lea_for_mov): Likewise.
8138 (ix86_avoid_lea_for_addr): Likewise.
8139 (ix86_lea_for_add_ok): Likewise.
8140 (exact_dependency_1): New function.
8141 (exact_store_load_dependency): Likewise.
8142 (ix86_adjust_cost): Handle Silvermont tuning.
8143 (do_reoder_for_imul): Likewise.
8144 (swap_top_of_ready_list): New function.
8145 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8146
8147 * config/i386/i386.md (peepholes that split memory operand in fp
8148 converts): New.
8149
8150 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8151
8152 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8153 Remove un-necessary braces.
8154
8155 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8156
8157 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8158 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8159
8160 2013-05-31 Tobias Burnus <burnus@net-b.de>
8161
8162 PR middle-end/57073
8163 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8164
8165 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8166
8167 PR target/56315
8168 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8169 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8170 * config/arm/neon.md (iordi3_neon): Remove.
8171 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8172 * config/arm/predicates.md (imm_for_neon_logic_operand):
8173 Move to earlier in the file.
8174 (neon_logic_op2): Likewise.
8175 (arm_iordi_operand_neon): New predicate.
8176
8177 2013-05-31 Richard Biener <rguenther@suse.de>
8178
8179 PR tree-optimization/57478
8180 PR tree-optimization/57453
8181 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8182 are life as well.
8183
8184 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8185
8186 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8187 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
8188
8189 2013-05-30 Tobias Burnus <burnus@net-b.de>
8190 Thomas Koenig <tkoenig@gcc.gnu.org>
8191
8192 PR middle-end/57073
8193 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
8194 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
8195
8196 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
8197
8198 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
8199
8200 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
8201
8202 * target.def (register_usage_leveling_p): New hook.
8203 * targhooks.c (default_register_usage_leveling_p): New.
8204 * targhooks.h (default_register_usage_leveling_p): New prototype.
8205 * lra-assigns.c (register_usage_leveling_p): Use the hook.
8206 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
8207 * doc/tm.texi: Update.
8208 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
8209
8210 2013-05-30 Ian Bolton <ian.bolton@arm.com>
8211
8212 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
8213 (*insv_reg<mode>): New define_insn.
8214
8215 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
8216
8217 PR rtl-optimization/57439
8218 * postreload.c (move2add_valid_value_p): Check that we have
8219 a zero subreg_regno_offset when accessing the register in
8220 the requested mode.
8221
8222 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8223 Igor Zamyatin <igor.zamyatin@intel.com>
8224
8225 Silvermont (SLM) architecture pipeline model, tuning and
8226 insn selection.
8227 * config.gcc: Add slm config options and target.
8228
8229 * config/i386/slm.md: New.
8230
8231 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
8232
8233 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
8234 PROCESSOR_SLM.
8235 (ix86_target_macros_internal): Likewise.
8236
8237 * gcc/config/i386/i386.c (slm_cost): New cost.
8238 (m_SLM): New macro flag.
8239 (initial_ix86_tune_features): Set m_SLM.
8240 (x86_accumulate_outgoing_args): Likewise.
8241 (x86_arch_always_fancy_math_387): Likewise.
8242 (processor_target_table): Add slm cost.
8243 (cpu_names): Add slm cpu name.
8244 (x86_option_override_internal): Set SLM ISA.
8245 (ix86_issue_rate): New case PROCESSOR_SLM.
8246 (ia32_multipass_dfa_lookahead): Likewise.
8247 (fold_builtin_cpu): Add slm.
8248
8249 * config/i386/i386.h (TARGET_SLM): New target macro.
8250 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
8251 (processor_type): Add PROCESSOR_SLM.
8252
8253 * config/i386/i386.md (cpu): Add new value "slm".
8254 (slm.md): Include slm.md.
8255
8256 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
8257 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8258
8259 * config/arm/arm-protos.h: Add and update function protos.
8260 * config/arm/arm.c (use_simple_return_p): New added.
8261 (thumb2_expand_return): Check simple_return flag.
8262 * config/arm/arm.md: Add simple_return and conditional simple_return.
8263 * config/arm/iterators.md: Add iterator for return and simple_return.
8264
8265 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8266
8267 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
8268 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
8269 (arm_emit_vfp_multi_reg_pop): Likewise.
8270 (thumb2_emit_ldrd_pop): Likewise.
8271 (arm_expand_epilogue): Add misc REG_CFA notes.
8272 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
8273
8274 2013-05-29 Lawrence Crowl <crowl@google.com>
8275
8276 * config/arm/t-arm: Update for below.
8277
8278 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
8279 Change type to hash_table. Update dependent calls and types.
8280
8281 * config/i386/t-cygming: Update for below.
8282
8283 * config/i386/t-interix: Update for below.
8284
8285 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
8286 Change type to hash_table. Update dependent calls and types.
8287 (i386_find_on_wrapper_list::wrappers): Likewise.
8288
8289 * config/ia64/t-ia64: Update for below.
8290
8291 * config/ia64/ia64.c (bundle_state_table):
8292 Change type to hash_table. Update dependent calls and types.
8293
8294 * config/mips/mips.c (mips_reorg_process_insns::htab):
8295 Change type to hash_table. Update dependent calls and types.
8296
8297 * config/sol2.c (solaris_comdat_htab):
8298 Change type to hash_table. Update dependent calls and types.
8299
8300 * config/t-sol2: Update for above.
8301
8302 2013-05-29 Teresa Johnson <tejohnson@google.com>
8303
8304 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
8305 functions are not yet marked as defined.
8306
8307 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
8308 Pat Haugen <pthaugen@us.ibm.com>
8309 Peter Bergner <bergner@vnet.ibm.com>
8310
8311 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
8312 instructions.
8313 (VEC_A): Likewise.
8314 (VEC_C): Likewise.
8315 (vrotl<mode>3): Likewise.
8316 (vashl<mode>3): Likewise.
8317 (vlshr<mode>3): Likewise.
8318 (vashr<mode>3): Likewise.
8319
8320 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8321 support for power8 V2DI builtins.
8322
8323 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
8324 power8 V2DI builtins.
8325 (vupkhsw): Likewise.
8326 (vupklsw): Likewise.
8327 (vaddudm): Likewise.
8328 (vminsd): Likewise.
8329 (vmaxsd): Likewise.
8330 (vminud): Likewise.
8331 (vmaxud): Likewise.
8332 (vpkudum): Likewise.
8333 (vpksdss): Likewise.
8334 (vpkudus): Likewise.
8335 (vpksdus): Likewise.
8336 (vrld): Likewise.
8337 (vsld): Likewise.
8338 (vsrd): Likewise.
8339 (vsrad): Likewise.
8340 (vsubudm): Likewise.
8341 (vcmpequd): Likewise.
8342 (vcmpgtsd): Likewise.
8343 (vcmpgtud): Likewise.
8344 (vcmpequd_p): Likewise.
8345 (vcmpgtsd_p): Likewise.
8346 (vcmpgtud_p): Likewise.
8347 (vupkhsw): Likewise.
8348 (vupklsw): Likewise.
8349 (vaddudm): Likewise.
8350 (vmaxsd): Likewise.
8351 (vmaxud): Likewise.
8352 (vminsd): Likewise.
8353 (vminud): Likewise.
8354 (vpksdss): Likewise.
8355 (vpksdus): Likewise.
8356 (vpkudum): Likewise.
8357 (vpkudus): Likewise.
8358 (vrld): Likewise.
8359 (vsld): Likewise.
8360 (vsrad): Likewise.
8361 (vsrd): Likewise.
8362 (vsubudm): Likewise.
8363
8364 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
8365 support for power8 V2DI instructions.
8366
8367 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
8368 power8 V2DI instructions. Combine pack and unpack insns to use an
8369 iterator for each mode. Check whether a particular mode supports
8370 Altivec instructions instead of just checking TARGET_ALTIVEC.
8371 (UNSPEC_VPKUWUM): Likewise.
8372 (UNSPEC_VPKSHSS): Likewise.
8373 (UNSPEC_VPKSWSS): Likewise.
8374 (UNSPEC_VPKUHUS): Likewise.
8375 (UNSPEC_VPKSHUS): Likewise.
8376 (UNSPEC_VPKUWUS): Likewise.
8377 (UNSPEC_VPKSWUS): Likewise.
8378 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
8379 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
8380 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
8381 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
8382 (UNSPEC_VUPKHSB): Likewise.
8383 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
8384 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
8385 (UNSPEC_VUPKHSH): Likewise.
8386 (UNSPEC_VUPKLSB): Likewise.
8387 (UNSPEC_VUPKLSH): Likewise.
8388 (VI2): Likewise.
8389 (VI_char): Likewise.
8390 (VI_scalar): Likewise.
8391 (VI_unit): Likewise.
8392 (VP): Likewise.
8393 (VP_small): Likewise.
8394 (VP_small_lc): Likewise.
8395 (VU_char): Likewise.
8396 (add<mode>3): Likewise.
8397 (altivec_vaddcuw): Likewise.
8398 (altivec_vaddu<VI_char>s): Likewise.
8399 (altivec_vadds<VI_char>s): Likewise.
8400 (sub<mode>3): Likewise.
8401 (altivec_vsubcuw): Likewise.
8402 (altivec_vsubu<VI_char>s): Likewise.
8403 (altivec_vsubs<VI_char>s): Likewise.
8404 (altivec_vavgs<VI_char>): Likewise.
8405 (altivec_vcmpbfp): Likewise.
8406 (altivec_eq<mode>): Likewise.
8407 (altivec_gt<mode>): Likewise.
8408 (altivec_gtu<mode>): Likewise.
8409 (umax<mode>3): Likewise.
8410 (smax<mode>3): Likewise.
8411 (umin<mode>3): Likewise.
8412 (smin<mode>3): Likewise.
8413 (altivec_vpkuhum): Likewise.
8414 (altivec_vpkuwum): Likewise.
8415 (altivec_vpkshss): Likewise.
8416 (altivec_vpkswss): Likewise.
8417 (altivec_vpkuhus): Likewise.
8418 (altivec_vpkshus): Likewise.
8419 (altivec_vpkuwus): Likewise.
8420 (altivec_vpkswus): Likewise.
8421 (altivec_vpks<VI_char>ss): Likewise.
8422 (altivec_vpks<VI_char>us): Likewise.
8423 (altivec_vpku<VI_char>us): Likewise.
8424 (altivec_vpku<VI_char>um): Likewise.
8425 (altivec_vrl<VI_char>): Likewise.
8426 (altivec_vsl<VI_char>): Likewise.
8427 (altivec_vsr<VI_char>): Likewise.
8428 (altivec_vsra<VI_char>): Likewise.
8429 (altivec_vsldoi_<mode>): Likewise.
8430 (altivec_vupkhsb): Likewise.
8431 (altivec_vupkhs<VU_char>): Likewise.
8432 (altivec_vupkls<VU_char>): Likewise.
8433 (altivec_vupkhsh): Likewise.
8434 (altivec_vupklsb): Likewise.
8435 (altivec_vupklsh): Likewise.
8436 (altivec_vcmpequ<VI_char>_p): Likewise.
8437 (altivec_vcmpgts<VI_char>_p): Likewise.
8438 (altivec_vcmpgtu<VI_char>_p): Likewise.
8439 (abs<mode>2): Likewise.
8440 (vec_unpacks_hi_v16qi): Likewise.
8441 (vec_unpacks_hi_v8hi): Likewise.
8442 (vec_unpacks_lo_v16qi): Likewise.
8443 (vec_unpacks_hi_<VP_small_lc>): Likewise.
8444 (vec_unpacks_lo_v8hi): Likewise.
8445 (vec_unpacks_lo_<VP_small_lc>): Likewise.
8446 (vec_pack_trunc_v8h): Likewise.
8447 (vec_pack_trunc_v4si): Likewise.
8448 (vec_pack_trunc_<mode>): Likewise.
8449
8450 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
8451 V2DI builtins.
8452 (vec_vmaxsd): Likewise.
8453 (vec_vmaxud): Likewise.
8454 (vec_vminsd): Likewise.
8455 (vec_vminud): Likewise.
8456 (vec_vpksdss): Likewise.
8457 (vec_vpksdus): Likewise.
8458 (vec_vpkudum): Likewise.
8459 (vec_vpkudus): Likewise.
8460 (vec_vrld): Likewise.
8461 (vec_vsld): Likewise.
8462 (vec_vsrad): Likewise.
8463 (vec_vsrd): Likewise.
8464 (vec_vsubudm): Likewise.
8465 (vec_vupkhsw): Likewise.
8466 (vec_vupklsw): Likewise.
8467
8468 2013-05-29 Jan Hubicka <jh@suse.cz>
8469
8470 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
8471 flags; reorder rest of fields in more consistent way.
8472 (varpool_node): Remove analyzed, finalized and alias.
8473 (cgraph_ndoe): Likewise.
8474 (symtab_alias_ultimate_target): New function.
8475 (cgraph_function_node): Move offline.
8476 (cgraph_reset_node): Declare.
8477 (cgraph_comdat_can_be_unshared_p): Remove.
8478 (varpool_remove_initializer): Declare.
8479 (varpool_first_defined_variable, varpool_next_defined_variable
8480 cgraph_first_defined_function, cgraph_next_defined_function): Update.
8481 (cgraph_function_with_gimple_body_p): Update.
8482 (varpool_all_refs_explicit_p): Update.
8483 (symtab_alias_target): New function.
8484 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
8485 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
8486 (cgraph_function_or_thunk_node): Simplify using
8487 symtab_alias_ultimate_target.
8488 (varpool_variable_node): Likewise.
8489 * cgraph.c (cgraph_create_function_alias): Update.
8490 (cgraph_add_thunk): Update.
8491 (cgraph_remove_node): Update.
8492 (dump_cgraph_node): Do not dump removed flags.
8493 (cgraph_function_body_availability): Update.
8494 (cgraph_propagate_frequency): Update.
8495 (verify_cgraph_node): Check sanity of local flag.
8496 (cgraph_function_node): Move here from cgraph.h; revamp for
8497 cgraph_function_or_thunk_node.
8498 * lto-symtab.c (lto_varpool_replace_node): Update.
8499 (lto_symtab_resolve_can_prevail_p): Update.
8500 (lto_symtab_merge_cgraph_nodes): Update.
8501 * ipa-cp.c (determine_versionability, initialize_node_lattices,
8502 propagate_constants_accross_call, devirtualization_time_bonus,
8503 ipcp_propagate_stage): Update.
8504 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
8505 * ipa-inline-transform.c (clone_inlined_nodes,
8506 preserve_function_body_p): Update.
8507 * ipa-reference.c (propagate): Update.
8508 (write_node_summary_p): Update.
8509 * toplev.c (wrapup_global_declaration_2): Update.
8510 * cgraphunit.c (cgraph_analyze_function): Rename to ...
8511 (analyze_function) ... this one.
8512 (cgraph_process_new_functions): Update.
8513 (cgraph_reset_node): Export.
8514 (cgraph_finalize_function): Update.
8515 (cgraph_add_new_function): Update.
8516 (process_function_and_variable_attributes): Update.
8517 (varpool_finalize_decl): Update.
8518 (symbol_finalized): Remove.
8519 (symbol_finalized_and_needed): Rename to ...
8520 (symbol_defined_and_needed): ... update.
8521 (cgraph_analyze_functions): Update.
8522 (handle_alias_pairs): Update.
8523 (mark_functions_to_output): Update.
8524 (assemble_thunk): Update.
8525 (output_in_order): Update.
8526 (output_weakrefs): Update.
8527 (finalize_compilation_unit): Update.
8528 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
8529 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
8530 input_node, input_varpool_node): Update.
8531 * dbxout.c (dbxout_expand_expr): Update.
8532 * cgraphclones.c (cgraph_clone_node): Update.
8533 (cgraph_copy_node_for_versioning): Update.
8534 (cgraph_materialize_clone): Update.
8535 (cgraph_materialize_all_clones): Update.
8536 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
8537 propagate_pure_const, propagate_nothrow): Update.
8538 * lto-streamer-out.c (lto_output, write_symbol): Update.
8539 * ipa-utils.c (ipa_reverse_postorder): Update.
8540 * ipa-inline.c (can_inline_edge_p): Update.
8541 (update_caller_keys, ipa_inline): Update.
8542 * dwarf2out.c (reference_to_unused,
8543 premark_types_used_by_global_vars_helper): Update.
8544 * tree-eh.c (tree_could_trap_p): Update.
8545 * ipa-split.c (consider_split, execute_split_functions): Update.
8546 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
8547 has_addr_references_p): Update; move ahead in file for better
8548 readability.
8549 (process_references): Simplify.
8550 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
8551 bodies are removed.
8552 (cgraph_comdat_can_be_unshared_p): Make static.
8553 (cgraph_externally_visible_p): Update.
8554 (varpool_externally_visible_p): Update.
8555 (function_and_variable_visibility): Update.
8556 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
8557 ipa_tm_mark_force_output_node): Update.
8558 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
8559 estimate_edge_devirt_benefit, inline_generate_summary,
8560 inline_write_summary): Update.
8561 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
8562 * ipa-prop.c (ipa_compute_jump_functions): Update.
8563 (ipa_print_node_params, ipa_prop_read_section,
8564 ipa_update_after_lto_read, read_replacements_section): Update.
8565 * varasm.c (mark_decl_referenced): Update.
8566 (assemble_alias, dump_tm_clone_pairs): Update.
8567 * tree-inline.c (copy_bb): Update.
8568 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
8569 Update.
8570 * symtab.c (dump_symtab_base): Print new flags.
8571 (verify_symtab_base): Verify new flags.
8572 (symtab_alias_ultimate_target): New function.
8573 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
8574 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
8575 Update.
8576 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
8577 Update.
8578 * i386.c (ix86_get_function_versions_dispatcher,
8579 ix86_generate_version_dispatcher_body): Update.
8580 (fold_builtin_cpu): Use varpool_add_new_variable.
8581 * varpool.c (varpool_remove_initializer): Break out from ...
8582 (varpool_remove_node): ... this one.
8583 (dump_varpool_node, varpool_node_for_asm,
8584 cgraph_variable_initializer_availability, varpool_analyze_node,
8585 varpool_assemble_decl, varpool_remove_unreferenced_decls,
8586 varpool_finalize_named_section_flags, varpool_create_variable_alias):
8587 Update.
8588
8589 2013-05-29 Jan Hubicka <jh@suse.cz>
8590
8591 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
8592
8593 2013-05-29 Easwaran Raman <eraman@google.com>
8594
8595 PR tree-optimization/57442
8596 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
8597 when control exits the main loop.
8598
8599 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
8600
8601 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
8602 and RX600.
8603 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
8604 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
8605 * rx/t-rx: Add rx100 under multi library matches option for nofpu
8606 option.
8607
8608 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8609
8610 PR tree-optimization/57441
8611 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
8612 Don't limit size of incr_vec to number of candidates.
8613
8614 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
8615
8616 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
8617 and mips16 directories.
8618 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
8619 (MULTILIB_DIRNAMES): Ditto.
8620 (MULTILIB_EXCEPTIONS): Add new exceptions.
8621 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
8622 (MULTILIB_DIRNAMES): Ditto.
8623 (MULTILIB_EXCEPTIONS): Add new exceptions.
8624
8625 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8626 Marcus Shawcroft <marcus.shawcroft@arm.com>
8627
8628 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
8629 SYMBOL_TINY_ABSOLUTE.
8630 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
8631 SYMBOL_TINY_ABSOLUTE.
8632 (aarch64_expand_mov_immediate): Likewise.
8633 (aarch64_classify_symbol): Likewise.
8634 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
8635 Permit SYMBOL_TINY_ABSOLUTE.
8636 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
8637
8638 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8639 Marcus Shawcroft <marcus.shawcroft@arm.com>
8640
8641 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
8642 Refactor if/switch. Replace gcc_assert with if.
8643
8644 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8645
8646 * config/i386/i386.c (initial_ix86_tune_features): Enable
8647 FP Reassociation for AMD bdver1 and bdver2.
8648
8649 2013-05-29 Martin Jambor <mjambor@suse.cz>
8650
8651 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
8652 and IMAGPART_EXPR do not occur within other handled_components.
8653
8654 2013-05-29 Richard Biener <rguenther@suse.de>
8655
8656 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
8657 access on whether the use is in the BB we currently try to
8658 vectorize.
8659 (vect_bb_vectorization_profitable_p): Pass the BB we currently
8660 vectorize to vect_bb_slp_scalar_cost.
8661
8662 2013-05-29 Richard Biener <rguenther@suse.de>
8663
8664 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
8665 computing scalar cost offsetted by stmts that are kept live
8666 by scalar uses.
8667 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
8668 for computation of scalar cost.
8669
8670 2013-05-28 Steve Ellcey <sellcey@mips.com>
8671
8672 * config/mips/mips-cpus.def (mips32r2): Change processor type.
8673
8674 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8675
8676 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
8677 array notation built-in reduction functions.
8678 * doc/passes.texi (Passes): Added documentation about changes done
8679 for Cilk Plus.
8680 * doc/invoke.texi (C Dialect Options): Added documentation about
8681 the -fcilkplus flag.
8682 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
8683 (BUILTINS_DEF): Depend on cilkplus.def.
8684 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
8685 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
8686 * cilkplus.def: New file.
8687
8688 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
8689
8690 PR rtl-optimization/57439
8691 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
8692
8693 2013-05-28 Easwaran Raman <eraman@google.com>
8694
8695 PR tree-optimization/57337
8696 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
8697 (find_insert_point): Correctly identify the insertion point
8698 when two statements with the same UID is compared.
8699
8700 2013-05-28 Richard Biener <rguenther@suse.de>
8701
8702 PR tree-optimization/56787
8703 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
8704 from the list of data references.
8705 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
8706 clobbers.
8707 (vect_analyze_loop_operations): Likewise.
8708 (vect_transform_loop): Remove clobbers.
8709
8710 2013-05-28 Martin Jambor <mjambor@suse.cz>
8711
8712 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
8713 and REALPART_EXPRs have scalar type.
8714
8715 2013-05-28 Richard Biener <rguenther@suse.de>
8716
8717 PR tree-optimization/57411
8718 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
8719 virtual operands.
8720 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
8721 virtual operand propagation.
8722
8723 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8724
8725 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
8726 destination register for bmasksi_vis.
8727 (vector_init_bshuffle): Likewise.
8728 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
8729
8730 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8731
8732 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
8733 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
8734 mode if the instruction isn't available in the original mode.
8735 * config/sparc/sparc.opt (mfix-ut699): New option.
8736 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
8737 (divdf3): Turn into expander.
8738 (divdf3_nofix): New insn.
8739 (divdf3_fix): Likewise.
8740 (divsf3): Disable if -mfix-ut699.
8741 (sqrtdf2): Turn into expander.
8742 (sqrtdf2_nofix): New insn.
8743 (sqrtdf2_fix): Likewise.
8744 (sqrtsf2): Disable if -mfix-ut699.
8745
8746 2013-05-27 Richard Biener <rguenther@suse.de>
8747
8748 PR middle-end/57412
8749 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
8750 block for the new loop.
8751
8752 2013-05-27 Richard Biener <rguenther@suse.de>
8753
8754 PR tree-optimization/57343
8755 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
8756 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
8757 (number_of_iterations_cond): Do not build the folded tree.
8758
8759 2013-05-27 Richard Biener <rguenther@suse.de>
8760
8761 Revert
8762 PR middle-end/57381
8763 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8764 OEP_CONSTANT_ADDRESS_OF retained.
8765
8766 PR tree-optimization/57417
8767 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
8768 for unchanged base.
8769 (set_ssa_val_to): Compare addresses using
8770 get_addr_base_and_unit_offset.
8771
8772 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
8773
8774 PR rtl-optimization/56833
8775 * postreload.c (move2add_record_mode): New function.
8776 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
8777 (move2add_use_add2_insn): Use move2add_record_sym_value.
8778 (move2add_use_add3_insn): Likewise.
8779 (reload_cse_move2add): Use move2add_valid_value_p and
8780 move2add_record_mode. Invalidate call-clobbered and REG_INC
8781 affected regs by setting reg_mode to VOIDmode.
8782 (move2add_note_store): Don't pretend the inside of a SUBREG is
8783 the actual destination. Invalidate single/leading registers by
8784 setting reg_mode to VOIDmode.
8785 Use move2add_record_sym_value, move2add_valid_value_p and
8786 move2add_record_mode.
8787
8788 2013-05-27 Richard Biener <rguenther@suse.de>
8789
8790 PR tree-optimization/57396
8791 * tree-affine.c (double_int_constant_multiple_p): Properly
8792 return false for val == 0 and div != 0.
8793
8794 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8795
8796 * config/mips/mips.h: Use #elif in preprocessor conditions.
8797
8798 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8799
8800 PR target/53916
8801 * config/mips/constraints.md (kl): New constraint.
8802 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8803 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8804 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8805 from LO for MIPS16.
8806 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8807
8808 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8809
8810 PR target/55777
8811 * config/mips/mips.c (mips_can_inline_p): New function.
8812 (TARGET_CAN_INLINE_P): Define.
8813
8814 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8815
8816 * sched-int.h (ds_t, dw_t): Make unsigned int.
8817 Fix documentation that describes how all the ds_t bits are used.
8818 Reserve the last bit for delayed-branch scheduling.
8819 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8820 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8821 (gen_dep_weak_1): Remove prototype.
8822 * sched-deps.c (get_dep_weak_1): Make static.
8823 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8824 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8825 * doc/tm.texi: Regenerate.
8826 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8827
8828 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8829
8830 PR debug/56950
8831 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8832
8833 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8834 Sandra Loosemore <sandra@codesourcery.com>
8835
8836 * config.gcc (powerpc-*): Allow native for with-cpu.
8837
8838 2013-05-24 Jeff Law <law@redhat.com>
8839
8840 PR tree-optimization/57124
8841 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8842 conversion feeding a condition if the range has an overflow
8843 if -fstrict-overflow. Add warnings for when we do make the
8844 transformation.
8845
8846 2013-05-24 Dehao Chen <dehao@google.com>
8847
8848 * tree-cfg.c (locus_discrim_map): Fix the typo.
8849 (locus_discrim_hasher): Likewise.
8850 (locus_discrim_hasher::hash): Likewise.
8851 (locus_discrim_hasher::equal): Likewise.
8852
8853 2013-05-24 Martin Jambor <mjambor@suse.cz>
8854
8855 PR tree-optimization/57294
8856 * cgraph.h (ipa_record_stmt_references): Declare.
8857 * cgraphbuild.c (ipa_record_stmt_references): New function.
8858 (build_cgraph_edges): Use ipa_record_stmt_references.
8859 (rebuild_cgraph_edges): Likewise.
8860 (cgraph_rebuild_references): Likewise.
8861 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8862 associated with the old statement and build references from the
8863 newly built statements.
8864 * ipa-ref.c (ipa_remove_stmt_references): New function.
8865 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8866
8867 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8868
8869 * lra-constraints.c (emit_spill_move): Use smaller mode for
8870 mem-mem moves.
8871 (check_and_process_move): Consider mem-reg moves for secondary
8872 too.
8873 (curr_insn_transform): Don't lose insns emitted before for
8874 secondary memory moves.
8875 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8876 reg set up in the current insn.
8877
8878 2013-05-24 Dehao Chen <dehao@google.com>
8879
8880 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8881 hash function.
8882 (locus_descrim_hasher::equal): Likewise.
8883 (build_gimple_cfg): New discriminator assignment algorithm.
8884 (make_edges): Likewise.
8885 (next_discriminator_for_locus): Likewise.
8886 (same_line_p): Likewise.
8887 (assign_discriminators): Likewise.
8888 (make_cond_expr_edges): Likewise.
8889 (make_gimple_switch_edges): Likewise.
8890 (make_goto_expr_edges): Likewise.
8891 (make_gimple_asm_edges): Likewise.
8892
8893 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8894
8895 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8896 X format specifier to only display bottom 16 bits.
8897 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8898 immediate to match for operand 2, since it will be masked.
8899
8900 2013-05-24 Richard Biener <rguenther@suse.de>
8901
8902 PR tree-optimization/57287
8903 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8904 all SSA names that occur in abnormal PHIs.
8905
8906 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8907
8908 PR tree-ssa/57385
8909 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8910 that index is not negative.
8911
8912 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8913
8914 PR rtl-optimization/55177
8915 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8916 (simplify_byte_swapping_operation): New.
8917 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8918 (simplify_relational_operation_1): Deal with BSWAP.
8919
8920 2013-05-23 Richard Henderson <rth@redhat.com>
8921
8922 PR target/56742
8923 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8924 (ix86_reorg): Call it.
8925
8926 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8927
8928 PR target/57379
8929 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8930 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8931 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8932
8933 2013-05-23 Christian Bruel <christian.bruel@st.com>
8934
8935 PR debug/57351
8936 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8937
8938 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8939 Marcus Shawcroft <marcus.shawcroft@arm.com>
8940
8941 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8942 * config/aarch64/constraints.md (Usa): Remove.
8943 * doc/md.texi (AArch64 Usa): Remove.
8944
8945 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8946 Marcus Shawcroft <marcus.shawcroft@arm.com>
8947
8948 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8949 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8950 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8951 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8952
8953 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8954
8955 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8956 instruction (AdvSIMD).
8957 * config/aarch64/aarch64-builtins.c
8958 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8959 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8960
8961 2013-05-23 Martin Jambor <mjambor@suse.cz>
8962
8963 PR middle-end/57347
8964 * tree.h (contains_bitfld_component_ref_p): Declare.
8965 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8966 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8967 caller.
8968 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8969 not access a bit-field. Assert all final offsets are byte-aligned.
8970
8971 2013-05-23 Richard Biener <rguenther@suse.de>
8972
8973 PR tree-optimization/57380
8974 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8975 least one invariant or re-used load.
8976 * passes.c (init_optimization_passes): Move pass_phiprop before
8977 pass_forwprop.
8978
8979 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8980
8981 * config/aarch64/aarch64-simd.md
8982 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8983
8984 2013-05-23 Richard Biener <rguenther@suse.de>
8985
8986 PR middle-end/57381
8987 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8988 OEP_CONSTANT_ADDRESS_OF retained.
8989
8990 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8991
8992 PR middle-end/57344
8993 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8994 don't lower unit. Handle unit not being always BITS_PER_WORD.
8995
8996 2013-05-23 Richard Biener <rguenther@suse.de>
8997
8998 PR rtl-optimization/57341
8999 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
9000 instead of true_dependence.
9001
9002 2013-05-22 David Malcolm <dmalcolm@redhat.com>
9003
9004 * bb-reorder.c (branch_threshold): Make const.
9005 (exec_threshold): Ditto.
9006
9007 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9008 Pat Haugen <pthaugen@us.ibm.com>
9009 Peter Bergner <bergner@vnet.ibm.com>
9010
9011 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
9012 documentation for the power8 crypto builtins.
9013
9014 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
9015
9016 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
9017 macros for defining power8 builtin functions.
9018 (BU_P8V_AV_2): Likewise.
9019 (BU_P8V_AV_P): Likewise.
9020 (BU_P8V_VSX_1): Likewise.
9021 (BU_P8V_OVERLOAD_1): Likewise.
9022 (BU_P8V_OVERLOAD_2): Likewise.
9023 (BU_CRYPTO_1): Likewise.
9024 (BU_CRYPTO_2): Likewise.
9025 (BU_CRYPTO_3): Likewise.
9026 (BU_CRYPTO_OVERLOAD_1): Likewise.
9027 (BU_CRYPTO_OVERLOAD_2): Likewise.
9028 (XSCVSPDP): Fix typo, point to the correct instruction.
9029 (VCIPHER): Add power8 crypto builtins.
9030 (VCIPHERLAST): Likewise.
9031 (VNCIPHER): Likewise.
9032 (VNCIPHERLAST): Likewise.
9033 (VPMSUMB): Likewise.
9034 (VPMSUMH): Likewise.
9035 (VPMSUMW): Likewise.
9036 (VPERMXOR_V2DI): Likewise.
9037 (VPERMXOR_V4SI: Likewise.
9038 (VPERMXOR_V8HI: Likewise.
9039 (VPERMXOR_V16QI: Likewise.
9040 (VSHASIGMAW): Likewise.
9041 (VSHASIGMAD): Likewise.
9042 (VPMSUM): Likewise.
9043 (VPERMXOR): Likewise.
9044 (VSHASIGMA): Likewise.
9045
9046 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9047 __CRYPTO__ if the crypto instructions are available.
9048 (altivec_overloaded_builtins): Add support for overloaded power8
9049 builtins.
9050
9051 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9052 support for power8 crypto builtins.
9053 (builtin_function_type): Likewise.
9054 (altivec_init_builtins): Add support for builtins that take vector
9055 long long (V2DI) arguments.
9056
9057 * config/rs6000/crypto.md: New file, define power8 crypto
9058 instructions.
9059
9060 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9061 Pat Haugen <pthaugen@us.ibm.com>
9062 Peter Bergner <bergner@vnet.ibm.com>
9063
9064 * doc/invoke.texi (Option Summary): Add power8 options.
9065 (RS/6000 and PowerPC Options): Likewise.
9066
9067 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9068 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9069 wm, wn, wr documentation.
9070
9071 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9072 registers if direct move instructions are enabled.
9073 (wn): New constraint for no registers.
9074 (wq): New constraint for quad word even GPR registers.
9075 (wr): New constraint if 64-bit instructions are enabled.
9076 (wv): New constraint if power8 vector instructions are enabled.
9077 (wQ): New constraint for quad word memory locations.
9078
9079 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9080 constraint for 0..15 for crypto instructions.
9081 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9082 as GPR and floating point registers.
9083 (int_reg_operand): New predicate to match only GPR registers.
9084 (base_reg_operand): New predicate to match base registers.
9085 (quad_int_reg_operand): New predicate to match even GPR registers
9086 for quad memory operations.
9087 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9088 operations in both GPR and VSX registers.
9089 (quad_memory_operand): New predicate for quad memory operations.
9090 (reg_or_indexed_operand): New predicate for direct move support.
9091
9092 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9093 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9094 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9095 (POWERPC_MASKS): Add power8 options.
9096 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9097 various options.
9098
9099 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9100 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9101
9102 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9103 (-mpower8-fusion): New power8 options.
9104 (-mpower8-fusion-sign): Likewise.
9105 (-mpower8-vector): Likewise.
9106 (-mcrypto): Likewise.
9107 (-mdirect-move): Likewise.
9108 (-mquad-memory): Likewise.
9109
9110 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9111 power8.
9112 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9113 registers.
9114 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9115 (rs6000_debug_vector_unit): Add p8_vector.
9116 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9117 definitions. Also print fusion state.
9118 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9119 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9120 (rs6000_option_override_internal): Add support for power8.
9121 (rs6000_common_init_builtins): Add debugging for skipped builtins
9122 if -mdebug=builtin.
9123 (rs6000_adjust_cost): Add power8 support.
9124 (rs6000_issue_rate): Likewise.
9125 (insn_must_be_first_in_group): Likewise.
9126 (insn_must_be_last_in_group): Likewise.
9127 (force_new_group): Likewise.
9128 (rs6000_register_move_cost): Likewise.
9129 (rs6000_opt_masks): Likewise.
9130
9131 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9132 power8 capable assembler, default to power7 options.
9133 (TARGET_DIRECT_MOVE): Likewise.
9134 (TARGET_CRYPTO): Likewise.
9135 (TARGET_P8_VECTOR): Likewise.
9136 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9137 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9138 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9139 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9140 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9141 (TARGET_XSCVDPSPN): Likewise.
9142 (TARGET_XSCVSPDPN): Likewsie.
9143 (TARGET_SYNC_HI_QI): Likewise.
9144 (TARGET_SYNC_TI): Likewise.
9145 (MASK_CRYPTO): Likewise.
9146 (MASK_DIRECT_MOVE): Likewise.
9147 (MASK_P8_FUSION): Likewise.
9148 (MASK_P8_VECTOR): Likewise.
9149 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9150 temporary used by some of the direct move instructions to get two FP
9151 temporary registers does not force creation of a stack frame.
9152 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9153 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9154 that any VSX registers are tieable, even if they are also an
9155 Altivec vector mode.
9156 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9157 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9158 (RS6000_BTM_CRYPTO): Likewise.
9159 (RS6000_BTM_COMMON): Likewise.
9160
9161 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9162 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9163 (enum rs6000_vector): Add power8 vector support.
9164
9165 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9166
9167 PR target/19599
9168 PR target/57340
9169 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9170 (any_sibcall_could_use_r3): this and handle indirect calls.
9171 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9172
9173 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9174
9175 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9176
9177 2013-05-22 Richard Biener <rguenther@suse.de>
9178
9179 PR middle-end/57349
9180 * profile.c (branch_prob): Do not split blocks that are
9181 abnormally receiving from ECF_RETURNS_TWICE functions.
9182
9183 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9184
9185 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9186 address mode. Move pointer mode initialization to the same place.
9187
9188 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9189
9190 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
9191 while it has any effect.
9192
9193 2013-05-21 Easwaran Raman <eraman@google.com>
9194
9195 PR tree-optimization/57322
9196 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
9197 UID of the statement added to the BB to be 1.
9198
9199 2013-05-21 Jakub Jelinek <jakub@redhat.com>
9200
9201 PR tree-optimization/57331
9202 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
9203 of conversion from pointer type to integral type with integer.
9204
9205 2013-05-21 Martin Jambor <mjambor@suse.cz>
9206
9207 PR lto/57289
9208 * ipa-prop.c (ipa_read_node_info): Process param_used and
9209 controlled_uses in the same order as when writing.
9210
9211 2013-05-21 Magnus Granberg <baldrick@free.fr>
9212
9213 PR plugins/56754
9214 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
9215
9216 2013-05-21 Richard Biener <rguenther@suse.de>
9217
9218 PR tree-optimization/57318
9219 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
9220 estimate stmts with side-effects as likely eliminated.
9221
9222 2013-05-21 Richard Biener <rguenther@suse.de>
9223
9224 PR tree-optimization/57330
9225 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
9226 preserve the call stmts fntype.
9227
9228 2013-05-21 Richard Biener <rguenther@suse.de>
9229
9230 PR tree-optimization/57303
9231 * tree-ssa-sink.c (statement_sink_location): Improve killing
9232 stmt detection and properly handle self-assignments.
9233
9234 2013-05-21 Christian Bruel <christian.bruel@st.com>
9235
9236 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
9237 spanning registers. LEAF_REG_REMAP is supported only for contiguous
9238 registers. Set register size out of the PARALLEL loop.
9239
9240 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
9241
9242 PR target/56547
9243 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
9244 (*fmasf4, *fmasf4_media): New insns.
9245
9246 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9247
9248 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
9249 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
9250 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
9251 (mips_idiv_insns): Update the comments to say that the returned
9252 instruction counts are in units of BASE_INSN_LENGTH.
9253 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
9254 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
9255 using 2 rather than 4 as the length of indirect MIPS16 and
9256 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
9257 length of a NOP. Don't divide MIPS16 lengths by 2.
9258 (mips16_split_long_branches): Assume a branch is long if the
9259 length is greater than 4 rather than 8.
9260 * config/mips/mips.md (length): Give MIPS16 lengths directly,
9261 rather than multiplying them by 2. Multiply instruction counts
9262 by BASE_INSN_LENGTH rather than 4.
9263 (*jump_mips16, tls_get_tp_mips16_<mode>)
9264 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
9265
9266 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9267
9268 * config/mips/mips.md (extended_mips16): Remove branch case.
9269 (length): Remove duplicated extended_mips16 test.
9270
9271 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9272
9273 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
9274
9275 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
9276
9277 * recog.h (Recog_data): Rename to...
9278 (recog_data_d): ...this.
9279 (recog_data): Update accordingly.
9280 * recog.c (recog_data): Likewise.
9281 * reload.c (save_recog_data): Likewise.
9282 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
9283 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
9284
9285 2013-05-17 Julian Brown <julian@codesourcery.com>
9286
9287 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
9288 found in a REG_EQUAL note, invalidate it.
9289
9290 2013-05-17 Easwaran Raman <eraman@google.com>
9291
9292 * tree-ssa-reassoc.c (find_insert_point): New function.
9293 (insert_stmt_after): Likewise.
9294 (get_def_stmt): Likewise.
9295 (ensure_ops_are_available): Likewise.
9296 (not_dominated_by): Likewise.
9297 (rewrite_expr_tree): Do not move statements beyond what is
9298 necessary. Remove call to swap_ops_for_binary_stmt...
9299 (reassociate_bb): ... and move it here.
9300 (build_and_add_sum): Assign UIDs for new statements.
9301 (linearize_expr): Likewise.
9302 (do_reassoc): Renumber gimple statement UIDs.
9303
9304 2013-05-17 Jan Hubicka <jh@suse.cz>
9305
9306 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
9307 weakrefs.
9308 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
9309 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
9310 weakrefs.
9311 (output_weakrefs): Update.
9312
9313 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
9314 Martin Jambor <mjambor@suse.cz>
9315
9316 PR middle-end/57276
9317 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
9318 value that corresponds to the given aggval is found in values vector.
9319
9320 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
9321
9322 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
9323 sse, sse2, sse3, ssse3 and sse4a flags to options.
9324
9325 2013-05-17 David Malcolm <dmalcolm@redhat.com>
9326
9327 * gengtype-state.c: (s_expr_writer): New class, to handle
9328 prettifying of output layout of s-expressions.
9329 (state_writer): New class, to write out gtype.state.
9330 (state_written_type_count): Move this variable into member data of
9331 state_writer.
9332 (s_expr_writer::s_expr_writer): New code: constructor for new class
9333 (state_writer::state_writer(): ditto
9334 (s_expr_writer::write_new_line): New function
9335 (s_expr_writer::write_any_indent): ditto
9336 (s_expr_writer::begin_s_expr): ditto
9337 (s_expr_writer::end_s_expr): ditto
9338 (write_state_fileloc): convert to method of state_writer...
9339 (state_writer:: write_state_fileloc): ...and use methods of
9340 s_expr_writer to write indentation into the gtype.state output file
9341 to visually represent the hierarchical structure of the list
9342 structures
9343 (write_state_fields): ditto, renaming to...
9344 (state_writer::write_state_fields)
9345 (write_state_a_string): ditto, renaming to...
9346 (state_writer::write_state_a_string)
9347 (write_state_string_option): ditto, renaming to...
9348 (state_writer::write_state_string_option)
9349 (write_state_type_option): ditto, renaming to...
9350 (state_writer::write_state_type_option)
9351 (write_state_nested_option): ditto, renaming to...
9352 (state_writer::write_state_nested_option)
9353 (write_state_option): ditto, renaming to...
9354 (state_writer::write_state_option)
9355 (write_state_options): ditto, renaming to...
9356 (state_writer::write_state_options)
9357 (write_state_lang_bitmap): ditto, renaming to...
9358 (state_writer::write_state_lang_bitmap)
9359 (write_state_version): ditto, renaming to...
9360 (state_writer::write_state_version)
9361 (write_state_scalar_type): ditto, renaming to...
9362 (state_writer::write_state_scalar_type)
9363 (write_state_string_type): ditto, renaming to...
9364 (state_writer::write_state_string_type)
9365 (write_state_undefined_type): ditto, renaming to...
9366 (state_writer::write_state_undefined_type)
9367 (write_state_struct_union_type): ditto, renaming to...
9368 (state_writer::write_state_struct_union_type)
9369 (write_state_struct_type): ditto, renaming to...
9370 (state_writer::write_state_struct_type)
9371 (write_state_user_struct_type): ditto, renaming to...
9372 (state_writer::write_state_user_struct_type)
9373 (write_state_lang_struct_type): ditto, renaming to...
9374 (state_writer::write_state_lang_struct_type)
9375 (write_state_param_struct_type): ditto, renaming to...
9376 (state_writer::write_state_param_struct_type)
9377 (write_state_pointer_type): ditto, renaming to...
9378 (state_writer::write_state_pointer_type)
9379 (write_state_array_type): ditto, renaming to...
9380 (state_writer::write_state_array_type)
9381 (write_state_gc_used): ditto, renaming to...
9382 (state_writer::write_state_gc_used)
9383 (write_state_common_type_content): ditto, renaming to...
9384 (state_writer::write_state_common_type_content)
9385 (write_state_type): ditto, renaming to...
9386 (state_writer::write_state_type)
9387 (write_state_pair_list): ditto, renaming to...
9388 (state_writer::write_state_pair_list)
9389 (write_state_pair): ditto, renaming to...
9390 (state_writer::write_state_pair)
9391 (write_state_typedefs): ditto, renaming to...
9392 (state_writer::write_state_typedefs)
9393 (write_state_structures): ditto, renaming to...
9394 (state_writer::write_state_structures)
9395 (write_state_param_structs): ditto, renaming to...
9396 (state_writer::write_state_param_structs)
9397 (write_state_variables): ditto, renaming to...
9398 (state_writer::write_state_variables)
9399 (write_state_srcdir): ditto, renaming to...
9400 (state_writer::write_state_srcdir)
9401 (write_state_files_list): ditto, renaming to...
9402 (state_writer::write_state_files_list)
9403 (write_state_languages): ditto, renaming to...
9404 (state_writer::write_state_languages)
9405 (write_state): create a state_writer instance and use it when
9406 writing out the state file
9407
9408 2013-05-17 Mike Stump <mikestump@comcast.net>
9409
9410 PR rtl-optimization/57304
9411 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
9412 accessing DF_REF_REAL_LOC.
9413
9414 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9415
9416 PR rtl-optimization/57281
9417 PR rtl-optimization/57300
9418 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
9419 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
9420 what the other splitter did if the registers are dead.
9421
9422 2013-05-17 Richard Biener <rguenther@suse.de>
9423
9424 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
9425 MEM_REF offsets.
9426
9427 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9428
9429 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
9430 linking.
9431
9432 2013-05-17 Marek Polacek <polacek@redhat.com>
9433
9434 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
9435 length when doing non-zero store of storing '\0' to '\0'.
9436
9437 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9438
9439 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
9440 vect_external_def oprnd1 with loop_vinfo, try to emit
9441 optional cast, negation and and stmts on the loop preheader
9442 edge instead of into the pattern def seq.
9443
9444 PR tree-optimization/57051
9445 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
9446 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
9447
9448 2013-05-16 Nick Clifton <nickc@redhat.com>
9449
9450 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
9451 (rl78_is_naked_func): New function.
9452 (rl78_expand_prologue): Skip prologue generation for naked functions.
9453 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
9454 * doc/extend.texi (naked): Add RL78 to the list of processors
9455 that supports this attribute.
9456
9457 2013-05-16 Jeff Law <law@redhat.com>
9458
9459 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
9460
9461 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9462
9463 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
9464 cache parameters using detect_caches_amd also for CYRIX,
9465 NSC and TM2 signatures.
9466
9467 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9468 Dzianis Kahanovich <mahatma@eu.by>
9469
9470 PR target/45359
9471 PR target/46396
9472 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9473 VIA/Centaur processors and determine their cache parameters
9474 using detect_caches_amd.
9475
9476 2013-05-16 Teresa Johnson <tejohnson@google.com>
9477
9478 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
9479 (rtl_verify_edges): New function.
9480 (rtl_verify_bb_insns): Ditto.
9481 (rtl_verify_bb_pointers): Ditto.
9482 (rtl_verify_bb_insn_chain): Ditto.
9483 (rtl_verify_fallthru): Ditto.
9484 (rtl_verify_bb_layout): Ditto.
9485 (rtl_verify_flow_info_1): Outline checks into new functions.
9486 (rtl_verify_flow_info): Ditto.
9487
9488 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
9489
9490 * cfghooks.c (copy_bbs): Add update_dominance argument.
9491 * cfghooks.h (copy_bbs): Update prototype.
9492 * tree-cfg.c (gimple_duplicate_sese_region):
9493 Add update_dominance argument.
9494 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
9495 * tree-ssa-loop-ch.c (copy_loop_headers): Update
9496 gimple_duplicate_sese_region call.
9497 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
9498 Update copy_bbs call.
9499 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
9500 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
9501
9502 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9503
9504 * tree-vectorizer.h (NUM_PATTERNS): Increment.
9505 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
9506 vect_recog_rotate_pattern.
9507 (vect_recog_rotate_pattern): New function.
9508
9509 2013-05-16 Jason Merrill <jason@redhat.com>
9510
9511 * Makefile.in (LLINKER): New variable.
9512 (mostlyclean): Remove link mutex.
9513 * configure.ac: Handle --enable-link-mutex.
9514 * lock-and-run.sh: New script.
9515
9516 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9517
9518 PR target/19599
9519 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
9520 for NULL decl.
9521
9522 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9523
9524 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
9525
9526 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
9527
9528 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
9529 * config/arm/arm.c (next_consecutive_mem): New function.
9530 (gen_movmem_ldrd_strd): Likewise.
9531 * config/arm/arm.md (movmemqi): Update condition and code.
9532 (unaligned_loaddi, unaligned_storedi): New patterns.
9533
9534 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9535
9536 * config.gcc: Obsolete *-*-solaris2.9*.
9537 * doc/install.texi (Specific, *-*-solaris2*): Document it.
9538
9539 2013-05-16 Richard Biener <rguenther@suse.de>
9540
9541 * passes.c (init_optimization_passes): Move pass_parallelize_loops
9542 earlier, after GRAPHITE transforms and IV canonicalization.
9543
9544 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9545
9546 * omp-low.c (extract_omp_for_data): For collapsed loops,
9547 if at least one of the loops is known at compile time to
9548 iterate zero times, set count to 0.
9549 (expand_omp_regimplify_p): New function.
9550 (expand_omp_for_generic): For collapsed loops, if at least
9551 one of the loops isn't known to iterate at least once,
9552 add runtime check with setting count to 0.
9553 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
9554 For unsigned types if it isn't known at compile time that
9555 the loop will iterate at least once, add runtime check to bypass
9556 the whole loop if initial condition isn't true.
9557
9558 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
9559
9560 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
9561
9562 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
9563
9564 PR middle-end/57286
9565 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
9566 transformations to avoid an infinite loop.
9567
9568 2013-05-16 Marek Polacek <polacek@redhat.com>
9569
9570 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
9571
9572 2013-05-15 Leif Ekblad <leif@rdos.net>
9573
9574 * config/i386/i386.c (ix86_decompose_address): Use
9575 DEFAULT_TLS_SEG_REG to access TLS segment register.
9576 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
9577 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
9578 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
9579
9580 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
9581
9582 PR target/57260
9583 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
9584 sibling calls to functions that would normally be lazily bound,
9585 unless $gp is call-clobbered.
9586
9587 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
9588
9589 * config/i386/i386.c (ix86_option_override_internal): Update
9590 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
9591 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
9592 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
9593 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
9594 of TARGET_3DNOW.
9595 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
9596
9597 2013-05-15 Andreas Schwab <schwab@suse.de>
9598
9599 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
9600 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
9601 third operand.
9602
9603 2013-05-15 Teresa Johnson <tejohnson@google.com>
9604
9605 * loop-unroll.c (report_unroll_peel): Check decision before
9606 emitting unroll/peel message.
9607
9608 2013-05-15 Teresa Johnson <tejohnson@google.com>
9609
9610 * function.h (has_bb_partition): New rtl_data flag.
9611 (bb_reorder_complete): Ditto.
9612 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
9613 instead of flag_reorder_blocks_and_partition.
9614 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
9615 with some enhancements.
9616 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
9617 * bb-reorder.c (connect_traces): Check for has_bb_partition
9618 instead of flag_reorder_blocks_and_partition.
9619 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
9620 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
9621 verify_hot_cold_block_grouping.
9622 (partition_hot_cold_basic_blocks): Set has_bb_partition.
9623
9624 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9625
9626 PR target/19599
9627 * config/arm/predicates.md (call_insn_operand): New predicate.
9628 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
9629 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
9630 if insn is not a tail call.
9631 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
9632 registers.
9633 * config/arm/arm.h (enum reg_class): New caller save register class.
9634 (REG_CLASS_NAMES): Likewise.
9635 (REG_CLASS_CONTENTS): Likewise.
9636 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
9637 without decls.
9638
9639 2013-05-15 Richard Biener <rguenther@suse.de>
9640
9641 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
9642 of MSG_OPTIMIZED_LOCATIONS.
9643 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
9644 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
9645 message.
9646 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
9647 of MSG_OPTIMIZED_LOCATIONS.
9648 (execute_vect_slp): Likewise.
9649 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
9650 (vect_create_cond_for_alias_checks): Likewise.
9651 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
9652 (vect_recog_widen_mult_pattern): Likewise.
9653 (vect_recog_widen_sum_pattern): Likewise.
9654 (vect_recog_over_widening_pattern): Likewise.
9655 (vect_recog_widen_shift_pattern): Likewise.
9656 (vect_recog_vector_vector_shift_pattern): Likewise.
9657 (vect_recog_divmod_pattern): Likewise.
9658 (vect_recog_mixed_size_cond_pattern): Likewise.
9659 (vect_recog_bool_pattern): Likewise.
9660 (vect_pattern_recog_1): Likewise.
9661
9662 2013-05-15 Martin Jambor <mjambor@suse.cz>
9663
9664 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
9665 non-functions to builtin_unreachable.
9666 * ipa-inline-transform.c (inline_call): Do not assert estimates were
9667 correct when new direct edges were discovered.
9668
9669 2013-05-15 Martin Jambor <mjambor@suse.cz>
9670
9671 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
9672 header, print symbol order instead of node uid, print more information
9673 about indirect edge targets.
9674 (ipa_make_edge_direct_to_target): Print symbol order instead of node
9675 uids.
9676 (ipa_make_edge_direct_to_target): Likewise.
9677 (remove_described_reference): Likewise.
9678 (propagate_controlled_uses): Likewise.
9679 (ipa_print_node_params): Also print symbol order.
9680 (ipcp_transform_function): Print symbol order instead of node uids.
9681 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
9682 (cgraph_get_create_real_symbol_node): Likewise.
9683 * ipa-cp.c (print_lattice): Likewise.
9684 (print_all_lattices): Likewise.
9685 (determine_versionability): Likewise.
9686 (initialize_node_lattices): Likewise.
9687 (estimate_local_effects): Likewise.
9688 (update_profiling_info): Likewise.
9689 (create_specialized_node): Likewise.
9690 (perhaps_add_new_callers): Likewise.
9691 (decide_about_value): Likewise.
9692 (decide_whether_version_node): Likewise.
9693 (identify_dead_nodes): Likewise.
9694 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
9695 (dump_inline_summary): Likewise.
9696 (estimate_node_size_and_time): Likewise.
9697 (inline_analyze_function): Likewise.
9698 * ipa-inline.c (report_inline_failed_reason): Likewise.
9699 (want_early_inline_function_p): Likewise.
9700 (edge_badness): Likewise.
9701 (update_edge_key): Likewise.
9702 (inline_small_functions): Likewise. Add dumping of order to two other
9703 dumps.
9704 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
9705 instead of node uids.
9706 (propagate_pure_const): Likewise.
9707 (propagate_pure_const): Likewise.
9708 * ipa-utils.c (dump_cgraph_node_set): Likewise.
9709 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
9710 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
9711 of node uids.
9712 * tree-pretty-print.c (dump_function_header): Likewise.
9713 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
9714 Print symbol order instead of node uids.
9715
9716 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9717
9718 * config/s390/s390.c (s390_register_move_cost): Don't impose the
9719 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
9720
9721 2013-05-15 Richard Biener <rguenther@suse.de>
9722
9723 PR tree-optimization/57275
9724 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
9725 return value for fail to do runtime alias checks for gather loads.
9726
9727 2013-05-15 Jan Hubicka <jh@suse.cz>
9728
9729 PR lto/57038
9730 PR lto/47375
9731 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
9732 weakrefs are not external.
9733 (lto_symtab_merge_decls): Fix thinko when dealing with
9734 non-lto_symtab decls.
9735 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
9736 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
9737 * varpool.c (dump_varpool_node): Dump more flags.
9738
9739 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9740
9741 * config/i386/i386.c (processor_alias_table): Add instruction
9742 FSGSBASE for AMD bdver3 architecture.
9743
9744 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9745
9746 * tree.c (warn_deprecated_use): Print file:line using locus color.
9747 * diagnostic.c (diagnostic_report_current_module): Print file:line
9748 and file:line:column using locus color.
9749
9750 2013-05-14 Mike Stump <mikestump@comcast.net>
9751
9752 * gdbinit.in: Add __null.
9753
9754 2013-05-14 Mike Stump <mikestump@comcast.net>
9755
9756 * recog.h: Rename struct recog_data to Recog_data.
9757 * recog.c: Likewise.
9758 * reload.c (can_reload_into): Likewise.
9759 * config/picochip/picochip.c: Likewise.
9760
9761 2013-05-14 Mike Stump <mikestump@comcast.net>
9762
9763 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
9764
9765 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
9766
9767 * resource.h (struct resources): Remove unch_memory member.
9768 (CLEAR_RESOURCE): Don't clear unch_memory.
9769 * resource.c (mark_referenced_resources): Don't set it.
9770 (mark_set_resources): Likewise.
9771 (mark_target_live_regs): Don't clear it.
9772 (init_resource_info): Likewise.
9773 * reorg.c (resource_conflicts_p): Don't compare it.
9774 (redundant_insn): Don't set it.
9775
9776 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
9777 Remove prototypes.
9778 * emit-rtl.c (next_label): Remove unused function.
9779 (skip_consecutive_labels, link_cc0_insns): Move to ...
9780 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
9781 only place where these functions are used, and make them static.
9782
9783 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9784
9785 * fold-const.c (fold_negate_expr): Handle vectors.
9786 (fold_truth_not_expr): Make it static.
9787 (fold_invert_truthvalue): New static function.
9788 (invert_truthvalue_loc): Handle vectors. Do not call
9789 fold_truth_not_expr directly.
9790 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
9791 <TRUTH_NOT_EXPR>: Do not cast to boolean.
9792 (fold_comparison): Handle vector constants.
9793 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
9794 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
9795 * tree.h (fold_truth_not_expr): Remove declaration.
9796
9797 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
9798
9799 * config/aarch64/aarch64-simd.md
9800 (aarch64_vcond_internal<mode>): Rename to...
9801 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9802 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9803 float modes. Clarify all iterator modes.
9804 (vcond<mode><mode>): Use new name for vcond expanders.
9805 (vcond<v_cmp_result><mode>): Likewise.
9806 (vcondu<mode><mode>: Likewise.
9807 * config/aarch64/iterators.md (VDQF_COND): New.
9808
9809 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9810
9811 PR bootstrap/57266
9812 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9813 variable for the shift amount. Check that we shift by non-negative
9814 amounts.
9815
9816 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9817
9818 PR target/42017
9819 * config/arm/arm.h (EPILOGUE_USES): Only return true
9820 for LR_REGNUM after epilogue_completed.
9821
9822 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9823
9824 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9825 is error_mark_node.
9826
9827 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9828
9829 PR target/57261
9830 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9831 and Solaris 11+/x86 with gld.
9832 * configure: Regenerate.
9833
9834 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9835
9836 * expmed.c (expand_shift_1): Canonicalize rotates by
9837 constant bitsize / 2 to bitsize - 1.
9838 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9839 case ROTATERT>: Likewise.
9840
9841 Revert:
9842 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9843
9844 * config/i386/i386.md (rotateinv): New code attr.
9845 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9846 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9847 roll $31, %eax, etc.
9848
9849 2013-05-14 Richard Biener <rguenther@suse.de>
9850
9851 PR middle-end/57235
9852 * tree-eh.c (sink_clobbers): Give up for successors with
9853 multiple predecessors and no virtual uses.
9854
9855 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9856
9857 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9858 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9859
9860 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9861
9862 PR middle-end/57251
9863 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9864 the case when both op0 and op1 have VOIDmode.
9865
9866 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9867
9868 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9869 in multiply-accumulate mode.
9870
9871 2013-05-13 Guozhi Wei <carrot@google.com>
9872
9873 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9874
9875 2013-05-13 Kai Tietz <ktietz@redhat.com>
9876
9877 PR target/56975
9878 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9879 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9880 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9881 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9882 (ix86_expand_prologue): Likewise.
9883 (ix86_expand_split_stack_prologue): Likewise.
9884 (legitimate_pic_address_disp_p): Likewise.
9885 (legitimize_pic_address): Likewise.
9886 (legitimize_tls_address): Likewise.
9887 (legitimize_pe_coff_symbol): Likewise.
9888 (output_pic_addr_const): Likewise.
9889 (construct_plt_address): Likewise.
9890 (ix86_expand_call): Likewise.
9891 (x86_output_mi_thunk): Likewise.
9892 (x86_function_profiler): Likewise.
9893
9894 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9895
9896 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9897 similar switch cases.
9898 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9899 (aarch64_simd_mov_to_<mode>low): Delete.
9900 (aarch64_simd_mov_to_<mode>high): Delete.
9901 (move_lo_quad_<mode>): Add w<-r alternative.
9902 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9903 (aarch64_simd_mov_from_*): Update type attribute.
9904 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9905 statement.
9906
9907 2013-05-13 Jan Hubicka <jh@suse.cz>
9908
9909 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9910 * config/i386/i386.c (ix86_compute_frame_layout,
9911 ix86_expand_epilogue, emit_i387_cw_initialization,
9912 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9913 ix86_local_alignment): Fix use of size/speed predicates.
9914
9915 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9916
9917 PR tree-optimization/45216
9918 PR tree-optimization/57157
9919 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9920 the (-Y) & (B - 1) variant if OP is |.
9921 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9922 return shifted. Use (-op1) & (prec - 1) as other_amount
9923 instead of prec - op1.
9924
9925 2013-05-13 Martin Jambor <mjambor@suse.cz>
9926
9927 PR middle-end/42371
9928 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9929 (ipa_constant_data): New type.
9930 (ipa_jump_func): Use ipa_constant_data to hold information about
9931 constant jump functions.
9932 (ipa_get_jf_constant): Adjust to jump function type changes.
9933 (ipa_get_jf_constant_rdesc): New function.
9934 (ipa_param_descriptor): New field controlled_uses.
9935 (ipa_get_controlled_uses): New function.
9936 (ipa_set_controlled_uses): Likewise.
9937 * ipa-ref.h (ipa_find_reference): Declare.
9938 * ipa-prop.c (ipa_cst_ref_desc): New type.
9939 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9940 changes.
9941 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9942 New parameter cs. Adjust all callers.
9943 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9944 (remove_described_reference): New function.
9945 (jfunc_rdesc_usable): Likewise.
9946 (try_make_edge_direct_simple_call): Decrement controlled use count,
9947 attempt to remove reference if it hits zero.
9948 (combine_controlled_uses_counters): New function.
9949 (propagate_controlled_uses): Likewise.
9950 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9951 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9952 (ipa_print_node_params): Print described use counter.
9953 (ipa_write_jump_function): Adjust to jump function type changes.
9954 (ipa_read_jump_function): New parameter CS, pass it to
9955 ipa_set_jf_constant. Adjust caller.
9956 (ipa_write_node_info): Stream controlled use count
9957 (ipa_read_node_info): Likewise.
9958 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9959 asserting.
9960 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9961 count. Remove cloning-added reference if it reaches zero.
9962 * ipa-ref.c (ipa_find_reference): New function.
9963
9964 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9965
9966 * config/i386/i386.c (processor_target_table): Modified default
9967 alignment values for AMD BD and BT architectures.
9968
9969 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9970
9971 * tree-vect-generic.c (uniform_vector_p): Move ...
9972 * tree.c (uniform_vector_p): ... here.
9973 * tree.h (uniform_vector_p): Declare it.
9974 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9975 into a scalar.
9976
9977 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9978
9979 PR tree-optimization/57230
9980 * tree-ssa-strlen.c (handle_char_store): Record length for
9981 array store from STRING_CST.
9982
9983 PR tree-optimization/57230
9984 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9985 check.
9986
9987 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9988
9989 * config/epiphany/epiphany.c (epiphany_init): Check size of
9990 NUM_MODES_FOR_MODE_SWITCHING.
9991 (epiphany_expand_prologue):
9992 Remove CONFIG_REGNUM initial value handling code.
9993 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9994 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9995 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9996 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9997 Don't return 1 for FP_MODE_NONE.
9998 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9999 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
10000 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
10001 * config/epiphany/epiphany.md (save_config): New pattern.
10002
10003 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
10004
10005 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
10006
10007 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
10008
10009 * config/i386/i386.md (memory): Handle sseishft1.
10010 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
10011 (*vec_extractv2di_1): Ditto.
10012
10013 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
10014
10015 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
10016 saved registers.
10017
10018 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
10019
10020 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
10021 Add mthumb/march=armv7-a multilib.
10022 Add mthumb/march=armv7-r multilib.
10023 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10024
10025 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10026
10027 * config/v850/t-rtems: Add more multilibs.
10028
10029 2013-05-10 Richard Biener <rguenther@suse.de>
10030
10031 PR tree-optimization/57214
10032 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10033 not propagate from SSA names that occur in abnormal PHI nodes.
10034
10035 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10036
10037 * stor-layout.c (element_precision): New function.
10038 * machmode.h (element_precision): Declare it.
10039 * tree.c (build_minus_one_cst): New function.
10040 (element_precision): Likewise.
10041 * tree.h (build_minus_one_cst): Declare new function.
10042 (element_precision): Likewise.
10043 * fold-const.c (operand_equal_p): Use element_precision.
10044 (fold_binary_loc): Handle vector types.
10045 * convert.c (convert_to_integer): Use element_precision.
10046 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10047 separately.
10048
10049 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10050
10051 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10052 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10053 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10054 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10055 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10056 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10057 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10058 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10059 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10060 (Uuw8): New constraints.
10061 (Usb4): Move into alphabetical order.
10062 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10063 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10064 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10065 previously unnamed patterns.
10066 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10067 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10068 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10069 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10070 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10071 of specific lengths.
10072
10073 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10074
10075 * config/i386/i386.md (rotateinv): New code attr.
10076 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10077 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10078 roll $31, %eax, etc.
10079
10080 PR tree-optimization/45216
10081 PR tree-optimization/57157
10082 * tree-ssa-forwprop.c (simplify_rotate): New function.
10083 (ssa_forward_propagate_and_combine): Call it.
10084
10085 2013-05-10 Richard Biener <rguenther@suse.de>
10086
10087 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10088 disable peeling when we version for aliasing.
10089 (vector_alignment_reachable_p): Honor explicit user alignment.
10090 (vect_supportable_dr_alignment): Likewise.
10091 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10092 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10093 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10094 then peeling to arrange for the cost-model check to come first.
10095
10096 2013-05-10 Alan Modra <amodra@gmail.com>
10097
10098 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10099 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10100 * configure: Regenerate.
10101
10102 2013-05-10 Alan Modra <amodra@gmail.com>
10103
10104 PR target/55033
10105 * varasm.c (default_elf_select_section): Move !DECL_P check..
10106 (get_named_section): ..to here before calling get_section_name.
10107 Adjust assertion.
10108 (default_section_type_flags): Add DECL_P check.
10109 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10110 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10111
10112 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10113
10114 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10115 When using gen_stack_adjust_str with a register offset, add a
10116 REG_FRAME_RELATED_EXPR note.
10117
10118 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10119
10120 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10121 (*vec_extractv4si_zext_mem): Ditto.
10122 (*vec_extractv2di): Add 0->x and x->x alternatives.
10123 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10124 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10125
10126 2013-05-09 Jason Merrill <jason@redhat.com>
10127
10128 N3639 C++1y VLA support
10129 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10130 DECL_VALUE_EXPR.
10131
10132 * tree.c (build_constructor_va): New.
10133 * tree.h: Declare it.
10134
10135 2013-05-09 Martin Jambor <mjambor@suse.cz>
10136
10137 PR lto/57084
10138 * gimple-fold.c (canonicalize_constructor_val): Call
10139 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10140
10141 2013-05-09 Jan Hubicka <jh@suse.cz>
10142 Richard Biener <rguenther@suse.de>
10143
10144 PR lto/54095
10145 * symtab.c (symtab_make_decl_local): Do not add private names.
10146
10147 2013-05-09 Jan Hubicka <jh@suse.cz>
10148
10149 PR lto/54095
10150 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10151 (unlink_from_assembler_name_hash): Likewise.
10152 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10153 symtab_unregister_node, symtab_initialize_asm_name_hash,
10154 change_decl_assembler_name): Update.
10155
10156 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10157
10158 * config/aarch64/aarch64.md: New movtf split.
10159 (*movtf_aarch64): Update.
10160 (aarch64_movdi_tilow): Handle TF modes and rename to
10161 aarch64_movdi_<mode>low.
10162 (aarch64_movdi_tihigh): Handle TF modes and rename to
10163 aarch64_movdi_<mode>high
10164 (aarch64_movtihigh_di): Handle TF modes and rename to
10165 aarch64_mov<mode>high_di
10166 (aarch64_movtilow_di): Handle TF modes and rename to
10167 aarch64_mov<mode>low_di
10168 (aarch64_movtilow_tilow): Remove spurious whitespace.
10169 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10170 splits.
10171 (aarch64_print_operand): Update.
10172
10173 2013-05-09 Alan Modra <amodra@gmail.com>
10174
10175 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10176 powerpc64le.
10177 * configure: Regenerate.
10178
10179 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10180
10181 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10182 splitter preparation statements.
10183 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10184 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10185 adjust_address_nv.
10186
10187 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10188
10189 * gimple-ssa-strength-reduction.c (count_candidates): Change
10190 return value to int.
10191 (analyze_candidates_and_replace): Change type of length to int.
10192
10193 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10194
10195 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
10196 (*vec_extract<mode>): Use VI12_128 mode iterator.
10197 (*vec_extract<mode>_mem): Ditto.
10198 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
10199 attribute.
10200
10201 2013-05-08 Diego Novillo <dnovillo@google.com>
10202
10203 PR bootstrap/54659
10204
10205 Revert:
10206 2012-08-17 Diego Novillo <dnovillo@google.com>
10207
10208 PR bootstrap/54281
10209 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
10210 * config.in: Regenerate.
10211 * configure: Regenerate.
10212 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
10213
10214 2013-05-08 Jan Hubicka <jh@suse.cz>
10215
10216 PR lto/54095
10217 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
10218 * cgraph.h (symtab_node_base): Add unique_name.
10219 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10220 input_overwrite_node, input_varpool_node): Stream unique_name.
10221 * cgraphclones.c (cgraph_create_virtual_clone,
10222 cgraph_function_versioning): Set unique_name.
10223 * ipa.c (function_and_variable_visibility): Set unique_name.
10224
10225 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10226
10227 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
10228 (alloc_cand_and_find_basis): Restrict conditional candidate
10229 processing to CAND_MULTs.
10230
10231 2013-05-08 Jan Hubicka <jh@suse.cz>
10232
10233 PR lto/54095
10234 lto-symtab.c (lto_symtab_symbol_p): New function.
10235 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
10236 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
10237 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
10238 Skip static symbols.
10239
10240 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
10241
10242 PR tree-optimization/57200
10243 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
10244 Only call inform if the preceding warning_at returns true.
10245
10246 2013-05-07 Han Shen <shenhan@google.com>
10247
10248 * cfgexpand.c (record_or_union_type_has_array_p): New function.
10249 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
10250 * common.opt (fstack-protector-strong): New option.
10251 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
10252 * doc/invoke.texi (Optimization Options): Document
10253 "-fstack-protector-strong".
10254 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
10255
10256 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10257
10258 * config/mips/mips.c (mips_machine_reorg2): Return 0.
10259
10260 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
10261
10262 * ira.c (update_equiv_regs): Add insn having equiv memory even if
10263 it is not lhs of the insn.
10264 (setup_reg_equiv): Remove insn having equiv memory which it is not
10265 lhs of the insn.
10266 * lra-constraints.c (process_address): Try to improve generation
10267 code for address base + disp.
10268 (lra_constraints): Make correct the code for checking insn setting
10269 up backward equivalence. Remove insn only if it is in the init
10270 insn list.
10271 * lra-eliminations.c (update_reg_eliminate): Change return value.
10272 (lra_eliminate): Use the result.
10273
10274 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
10275
10276 * config/i386/sse.md (ssescalarnummask): New mode attribute.
10277 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
10278 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
10279 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
10280 register target operands.
10281 (*vec_extractv8hi_sse2): New pattern.
10282 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
10283 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
10284 (*vec_extract<mode>_mem): New insn and split pattern.
10285
10286 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
10287
10288 * config/arm/arm.c (arm_asan_shadow_offset): New function.
10289 (TARGET_ASAN_SHADOW_OFFSET): Define.
10290 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
10291 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
10292
10293 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10294
10295 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
10296 (incr_vec_index): Return -1 if increment not found.
10297 (create_add_on_incoming_edge): Assert if increment not found.
10298 (record_increment): Limit number of increments recorded.
10299 (all_phi_incrs_profitable): Return false if an increment not found.
10300 (replace_profitable_candidates): Don't process increments that were
10301 not recorded.
10302 (analyze_candidates_and_replace): Limit size of incr_vec.
10303
10304 2013-05-07 Richard Biener <rguenther@suse.de>
10305
10306 * calls.c (special_function_p): setjmp-like functions are leaf.
10307 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
10308 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
10309
10310 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
10311
10312 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
10313 (aarch64_simd_mov<mode>): New expander.
10314 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
10315 (aarch64_simd_mov_to_<mode>high): Likewise.
10316 (aarch64_simd_mov_from_<mode>low): Likewise.
10317 (aarch64_simd_mov_from_<mode>high): Likewise.
10318 (aarch64_dup_lane<mode>): Update.
10319 (aarch64_dup_lanedi): New instruction pattern.
10320 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
10321 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
10322
10323 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10324
10325 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
10326 (replace_mult_candidate): Remove unnecessary argument; remove
10327 unnecessary parameter from call to introduce_cast_before_cand.
10328 (replace_unconditional_candidate): Remove unnecessary parameter
10329 from call to replace_mult_candidate.
10330 (replace_conditional_candidate): Likewise.
10331 (insert_initializers): Use make_temp_ssa_name.
10332 (introduce_cast_before_cand): Remove unnecessary argument; use
10333 make_temp_ssa_name.
10334 (replace_one_candidate): Remove unnecessary argument; remove
10335 unnecessary parameter from calls to introduce_cast_before_cand.
10336 (replace_profitable_candidates): Remove unnecessary parameters
10337 from calls to replace_one_candidate.
10338
10339 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10340
10341 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
10342 phi def as possibly hiding a basis for a CAND_ADD whose operands
10343 have been commuted in the analysis.
10344 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
10345
10346 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10347
10348 * config/aarch64/aarch64.md
10349 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
10350 shift value between 0-4.
10351
10352 2013-05-07 Richard Biener <rguenther@suse.de>
10353
10354 * double-int.h (rshift): New overload.
10355 * double-int.c (rshift): New function.
10356 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
10357 (create_reference_ops_from_ref): Remove.
10358 (vn_reference_insert): Use shared ops for constructing the
10359 reference and copy it.
10360
10361 2013-05-07 Richard Biener <rguenther@suse.de>
10362
10363 PR middle-end/57190
10364 * tree-eh.c (sink_clobbers): Properly propagate
10365 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
10366
10367 2013-05-07 Jakub Jelinek <jakub@redhat.com>
10368
10369 PR tree-optimization/57149
10370 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
10371 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
10372 collect_phi_def_edges, execute_late_warn_uninitialized): Use
10373 uninit_undefined_value_p instead of ssa_undefined_value_p.
10374
10375 PR debug/57184
10376 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
10377 for modifier == EXPAND_INITIALIZER.
10378
10379 2013-05-07 Anton Blanchard <anton@samba.org>
10380
10381 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
10382 for powerpc64 little endian.
10383 * configure: Regenerate.
10384
10385 2013-05-06 Graham Stott <grahams@btinternet.com>
10386
10387 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
10388 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
10389 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
10390 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
10391
10392 2013-05-06 Graham Stott <grahams@btinternet.com>
10393
10394 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
10395 codes which allow non-lvalues.
10396
10397 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
10398
10399 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
10400 components are all 1s.
10401 (integer_minus_onep): New function.
10402 * tree.h (integer_minus_onep): Declare it.
10403 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
10404 integer_minus_onep instead of integer_all_onesp.
10405
10406 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10407
10408 PR target/52933
10409 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
10410 variations of these patterns.
10411
10412 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10413
10414 * config/i386/i386.md (isa): Add x64_sse4 member.
10415 (enabled): Handle x64_sse4.
10416 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
10417 instruction for 64bit SSE4_1 targets. Update insn attributes.
10418 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
10419 instruction for SSE4_1 targets. Update insn attributes.
10420 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10421 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
10422 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
10423 const_1 selector.
10424 (*vec_extractv4si): Rename from *sse4_1_pextrd.
10425 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
10426 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
10427
10428 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10429
10430 PR target/57108
10431 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
10432
10433 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
10434
10435 * final.c (do_assembler_dialects): Don't handle curly braces and
10436 vertical bar escaped by % as dialect delimiters.
10437 (output_asm_insn): Print curly braces and vertical bar if escaped
10438 by % and ASSEMBLER_DIALECT defined.
10439 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
10440 * doc/tm.texi: Regenerated.
10441
10442 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10443
10444 * config/mips/mips.c: Include tree-pass.h.
10445 (mips_reorg): Split in pre- and post-dbr_schedule parts.
10446 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
10447 (pass_mips_machine_reorg2): New machine specific pass.
10448 (insert_pass_mips_machine_reorg2): New pass plugin definition.
10449 (mips_option_override): Register the new pass.
10450 * rtl.h (cleanup_barriers): Remove prototype.
10451 (dbr_schedule): Likewise.
10452 * jump.c (cleanup_barriers): Make static.
10453 * reorg.c (dbr_schedule): Likewise.
10454
10455 2013-05-06 Richard Biener <rguenther@suse.de>
10456
10457 PR tree-optimization/57185
10458 * tree-parloops.c (add_field_for_reduction): Handle anonymous
10459 SSA names properly.
10460
10461 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10462
10463 PR target/57106
10464 * config/i386/i386.c (add_parameter_dependencies): Add dependence
10465 between "first_arg" and "insn", not "last" and "insn".
10466
10467 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10468
10469 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
10470 (find_candidates_in_block): Re-enable slsr_process_phi.
10471 (create_phi_basis): Fix double counting of candidate adjustment.
10472
10473 2013-05-06 Richard Biener <rguenther@suse.de>
10474
10475 PR middle-end/57147
10476 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
10477 the edge is also fallthru, preserve it and just clear the
10478 abnormal flag.
10479 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
10480 also complex, preserve that and just clear the fallthru flag.
10481 * tree-inline.c (update_ssa_across_abnormal_edges): Also
10482 update virtual operands.
10483
10484 2013-05-06 Alan Modra <amodra@gmail.com>
10485
10486 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
10487 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
10488 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
10489 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
10490 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10491 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
10492
10493 2013-05-06 Alan Modra <amodra@gmail.com>
10494
10495 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
10496 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
10497 (DEFAULT_ASM_ENDIAN): Define.
10498 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10499 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
10500 Update -K PIC clause from sysv4.h.
10501 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
10502 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
10503
10504 2013-05-06 Alan Modra <amodra@gmail.com>
10505
10506 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
10507 twice for little-endian.
10508 (ashrdi3_no_power, ashrdi3): Support little-endian.
10509
10510 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10511
10512 PR target/55303
10513 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
10514 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
10515 related expanders.
10516 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
10517 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
10518 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
10519 New predicates.
10520
10521 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
10522 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
10523
10524 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
10525 * config/pa/pa.opt: Make mbig-switch a no-op.
10526 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
10527 (CASE_VECTOR_MODE): Always return SImode.
10528 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
10529 for the !TARGET_BIG_SWITCH case.
10530 * config/pa/pa-linux.h: Likewise.
10531 * config/pa/pa-openbsd.h: Likewise.
10532 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
10533 * config/pa/pa.md (short_jump): Remove define_insn.
10534 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
10535 (casesi0): Remove define_insn.
10536 (type): Remove btable_branch.
10537 (pa_combine_type): Likewise.
10538 (in_nullified_branch_delay): Likewise.
10539 (in_call_delay): Likewise.
10540 (define_delay): Likewise.
10541 (define_insn_reservation "Z3"): Likewise.
10542 (define_insn_reservation "Z4"): Likewise.
10543 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
10544 (pa_adjust_insn_length): Remove adjustment for btable branches.
10545 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
10546 and mno-big-switch
10547
10548 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
10549
10550 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10551 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
10552 Add m->r,x alternatives.
10553 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
10554 splitters using SWI48x mode iterator.
10555 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
10556 TARGET_64BIT. Add m->x alternative.
10557 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
10558 Add o->x alternative. Enable for TARGET_SSE.
10559 (sse_storeq): Remove expander.
10560 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
10561 with memory input operand.
10562 (*vec_extractv2di_1 splitter): New.
10563 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
10564 * config/i386/i386.md (ssevecmodelower): New mode attribute.
10565
10566 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10567
10568 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
10569 (INT_LOWPART): Delete.
10570 (extract_MB): Adjust.
10571 (extract_ME): Adjust.
10572 (print_operand): Adjust.
10573
10574 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10575
10576 * config/rs6000/predicates.md (reg_or_add_cint_operand,
10577 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10578 (reg_or_logical_cint_operand, easy_fp_constant,
10579 logical_const_operand): Delete "CONST_DOUBLE" case.
10580 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
10581 "HOST_BITS_PER_WIDE_INT == 64" test.
10582 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
10583 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
10584 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
10585 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
10586 test.
10587 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
10588 CONST_DOUBLE DImode/VOIDmode case.
10589 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
10590 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10591 CONST_DOUBLE VOIDmode case.
10592 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10593 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
10594 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
10595 Delete CONST_DOUBLE case.
10596 (splitters for mov FMOVE64 const_double): Delete
10597 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10598 "HOST_BITS_PER_WIDE_INT >= 64" test.
10599 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
10600 case.
10601 (mov DI const_double): Delete.
10602
10603 2013-05-04 Jakub Jelinek <jakub@redhat.com>
10604
10605 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
10606 on op shows all bits zero in mode of a lowpart subreg, return zero.
10607
10608 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
10609
10610 PR target/57150
10611 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
10612 to save TFmode registers and DImode to save TImode registers for
10613 caller save operations.
10614 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
10615 mark being partially clobbered since they only use the first
10616 double word.
10617
10618 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
10619 and TDmode only use the upper 64-bits of each VSX register.
10620
10621 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10622
10623 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
10624 (find_candidates_in_block): Disable slsr_process_phi.
10625
10626 2013-05-03 Guozhi Wei <carrot@google.com>
10627
10628 * coverage.c (coverage_obj_init): Move the construction of gcov
10629 constructor to ...
10630 (build_init_ctor): ... here.
10631
10632 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10633
10634 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
10635 (slsr_cand_d): Redefine def_phi.
10636 (stride_status, phi_adjust_status, count_phis_status): New enums.
10637 (find_phi_def): New.
10638 (find_basis_for_base_expr): New.
10639 (find_basis_for_candidate): Handle hidden bases.
10640 (alloc_cand_and_find_basis): Handle phi candidates.
10641 (slsr_process_phi): New.
10642 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
10643 (create_mul_imm_cand): Likewise.
10644 (create_add_ssa_cand): Exclude phi base candidates.
10645 (create_add_imm_cand): Likewise.
10646 (slsr_process_cast): Likewise.
10647 (slsr_process_copy): Likewise.
10648 (find_candidates_in_block): Handle phi candidates.
10649 (dump_candidate): Likewise.
10650 (unconditional_cands): Delete.
10651 (unconditional_cands_with_known_stride_p): Delete.
10652 (phi_dependent_cand_p): New.
10653 (cand_increment): Handle phi-dependent candidates.
10654 (replace_dependent): Delete.
10655 (replace_mult_candidate): New.
10656 (replace_unconditional_candidate): New.
10657 (incr_vec_index): Move to avoid forward reference.
10658 (create_add_on_incoming_edge): New.
10659 (create_phi_basis): New.
10660 (replace_dependents): Delete.
10661 (replace_conditional_candidate): New.
10662 (phi_add_costs): New.
10663 (replace_uncond_cands_and_profitable_phis): New.
10664 (record_increment): Handle phi adjustments.
10665 (record_phi_increments): New.
10666 (record_increments): Handle phi adjustments.
10667 (phi_incr_cost): New.
10668 (lowest_cost_path): Handle phis.
10669 (total_savings): Likewise.
10670 (analyze_increments): Likewise.
10671 (ncd_with_phi): New.
10672 (ncd_of_cand_and_phis): New.
10673 (nearest_common_dominator_for_cands): Handle phi increments.
10674 (all_phi_incrs_profitable): New.
10675 (replace_profitable_candidates): Handle phi-dependent candidates.
10676 (analyze_candidates_and_replace): Likewise.
10677
10678 2013-05-03 Teresa Johnson <tejohnson@google.com>
10679
10680 PR bootstrap/57154
10681 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
10682 do not exceed REG_BR_PROB_BASE.
10683
10684 2013-05-03 Jeff Law <law@redhat.com>
10685
10686 PR tree-optimization/57144
10687 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
10688 operand of the condition will bit into the new type when eliminating
10689 a cast feeding a condition.
10690
10691 2013-05-03 Jakub Jelinek <jakub@redhat.com>
10692
10693 PR rtl-optimization/57130
10694 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
10695 of COMPARE as in_code to the recursive call if needed.
10696
10697 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
10698
10699 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
10700 (enabled): Handle new members.
10701 * config/i386/sse.md (*vec_concatv2si): Merge from
10702 *vec_concatv2si_sse2 and vec_concatv2si_sse.
10703 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
10704
10705 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
10706
10707 PR tree-optimization/57027
10708 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
10709 for fnms opportunity, check we got the prerequisite kind
10710 of tree / gimple before using accessor functions.
10711
10712 2013-05-03 Richard Biener <rguenther@suse.de>
10713
10714 * double-int.h (lshift): New overload without precision
10715 and arith argument.
10716 (operator *=, operator +=, operator -=): Move ...
10717 * double-int.c (operator *=, operator +=, operator -=): ... here
10718 and implement more efficiently.
10719 (mul_double_with_sign): Remove.
10720 (lshift_double): Adjust to take unsinged shift argument, push
10721 dispatching code to callers.
10722 (mul_double_wide_with_sign): Add early out for callers that
10723 are not interested in high parts or overflow.
10724 (lshift): New function.
10725 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
10726 dispatch code here.
10727 (lrotate, rrotate): Use logical shifts.
10728 * expr.c (get_inner_reference): Use lshift.
10729 * fixed-value.c (do_fixed_divide): Likewise.
10730 * tree-dfa.c (get_ref_base_and_extent): Likewise.
10731 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
10732 (indirect_refs_may_alias_p): Likewise.
10733 (stmt_kills_ref_p_1): Likewise.
10734
10735 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10736
10737 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
10738
10739 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10740
10741 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
10742 scalar form of FABD instruction.
10743
10744 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
10745
10746 * lra-constraints.c (process_alt_operands): Add checking alt
10747 number to choose the best alternative.
10748
10749 2013-05-02 Richard Biener <rguenther@suse.de>
10750
10751 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
10752 bitmap and its handling.
10753 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
10754
10755 2013-05-02 Richard Biener <rguenther@suse.de>
10756
10757 PR middle-end/57140
10758 * tree-inline.c (copy_loops): Properly handle removed loops.
10759 (copy_cfg_body): Mark destination loops for fixup if source
10760 loops needed fixup.
10761
10762 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
10763
10764 PR target/56732
10765 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
10766 generating simple_return for naked functions.
10767
10768 2013-05-02 Martin Jambor <mjambor@suse.cz>
10769
10770 PR middle-end/56988
10771 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
10772 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
10773 flags match.
10774 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
10775 ipa_agg_replacement_value structures.
10776 (known_aggs_to_agg_replacement_list): Likewise.
10777 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
10778 (read_agg_replacement_chain): Likewise.
10779 (ipcp_transform_function): Also check that by_ref flags match.
10780
10781 2013-05-02 Richard Biener <rguenther@suse.de>
10782
10783 * graphds.h (struct graph): Add obstack member.
10784 * graphds.c (new_graph): Initialize obstack and allocate
10785 vertices from it.
10786 (add_edge): Allocate edge from the obstack.
10787 (free_graph): Free the obstack instead of all edges and vertices.
10788
10789 2013-05-02 Teresa Johnson <tejohnson@google.com>
10790
10791 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
10792 divides.
10793 * cfg.c (update_bb_profile_for_threading): Ditto.
10794 * tree-inline.c (copy_bb): Ditto.
10795 (copy_edges_for_bb): Ditto.
10796 (initialize_cfun): Ditto.
10797 (copy_cfg_body): Ditto.
10798 (expand_call_inline): Ditto.
10799 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10800 (estimate_node_size_and_time): Ditto.
10801 (inline_merge_summary): Ditto.
10802 * cgraphclones.c (cgraph_clone_edge): Ditto.
10803 (cgraph_clone_node): Ditto.
10804 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10805 (compute_trg_info): Ditto.
10806
10807 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10808
10809 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10810 S reg when fp attribute set.
10811 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10812
10813 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10814
10815 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10816 New pattern.
10817 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10818 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10819 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10820
10821 2013-05-02 Richard Biener <rguenther@suse.de>
10822
10823 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10824 (struct instantiate_cache_entry): New type.
10825 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10826 (struct instantiate_cache_type): New type.
10827 (set_instantiated_value, get_instantiated_value): Remove.
10828 (get_instantiated_value_entry): New function.
10829 (instantiate_scev_name): Use the new cache and adjust.
10830 (instantiate_scev_poly): Adjust.
10831 (instantiate_scev_binary): Likewise.
10832 (instantiate_array_ref): Likewise.
10833 (instantiate_scev_convert): Likewise.
10834 (instantiate_scev_not): Likewise.
10835 (instantiate_scev_3): Likewise.
10836 (instantiate_scev_2): Likewise.
10837 (instantiate_scev_r): Likewise.
10838 (instantiate_scev): Likewise.
10839 (resolve_mixers): Likewise.
10840
10841 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10842
10843 PR target/57091
10844 * lra-constraints.c (best_small_class_operands_num): Remove.
10845 (process_alt_operands): Remove small_class_operands_num. Take
10846 small classes operands into losers and only if the operand is not
10847 matched. Modify debugging output.
10848 (curr_insn_transform): Remove best_small_class_operands_num.
10849 Print insn name.
10850
10851 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10852
10853 * config/aarch64/aarch64-builtins.c
10854 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10855 * config/aarch64/aarch64-simd-builtins.def
10856 (reduc_splus_): Add new modes.
10857 (reduc_uplus_): New.
10858 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10859 (reduc_uplus_v4sf): Likewise.
10860 (reduc_splus_v4sf): Likewise.
10861 (aarch64_addv<mode>): Likewise.
10862 (reduc_uplus_<mode>): Likewise.
10863 (reduc_splus_<mode>): Likewise.
10864 (aarch64_addvv2di): Likewise.
10865 (reduc_uplus_v2di): Likewise.
10866 (reduc_splus_v2di): Likewise.
10867 (aarch64_addvv2si): Likewise.
10868 (reduc_uplus_v2si): Likewise.
10869 (reduc_splus_v2si): Likewise.
10870 (reduc_<sur>plus_<mode>): New.
10871 (reduc_<sur>plus_v2di): Likewise.
10872 (reduc_<sur>plus_v2si): Likewise.
10873 (reduc_<sur>plus_v4sf): Likewise.
10874 (aarch64_addpv4sf): Likewise.
10875 * config/aarch64/arm_neon.h
10876 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10877 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10878 add UNSPEC_SADDV, UNSPEC_UADDV.
10879 (SUADDV): New.
10880 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10881
10882 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10883
10884 * config/aarch64/arm_neon.h
10885 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10886
10887 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10888
10889 * config/aarch64/aarch64-builtins
10890 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10891
10892 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10893
10894 * config/aarch64/aarch64-simd-builtins.def
10895 (reduc_smax_): New.
10896 (reduc_smin_): Likewise.
10897 (reduc_umax_): Likewise.
10898 (reduc_umin_): Likewise.
10899 (reduc_smax_nan_): Likewise.
10900 (reduc_smin_nan_): Likewise.
10901 (fmax): Remove.
10902 (fmin): Likewise.
10903 (smax): Update for V2SF, V4SF and V2DF modes.
10904 (smin): Likewise.
10905 (smax_nan): New.
10906 (smin_nan): Likewise.
10907 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10908 (<su><maxmin><mode>3): ...This, refactor.
10909 (s<maxmin><mode>3): New.
10910 (<maxmin_uns><mode>3): Likewise.
10911 (reduc_<maxmin_uns>_<mode>): Refactor.
10912 (reduc_<maxmin_uns>_v4sf): Likewise.
10913 (reduc_<maxmin_uns>_v2si): Likewise.
10914 (aarch64_<fmaxmin><mode>: Remove.
10915 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10916 new builtin names.
10917 (vmin<q>_f<32,64>): Likewise.
10918 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10919 (FMAXMIN): New.
10920 (su): Add mappings for smax, smin, umax, umin.
10921 (maxmin): New.
10922 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10923 (FMAXMIN): Rename as...
10924 (FMAXMIN_UNS): ...This.
10925 (maxminv): Remove.
10926 (fmaxminv): Likewise.
10927 (fmaxmin): Likewise.
10928 (maxmin_uns): New.
10929 (maxmin_uns_op): Likewise.
10930
10931 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10932
10933 * config/aarch64/arm_neon.h
10934 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10935 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10936 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10937
10938 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10939
10940 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10941 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10942
10943 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10944
10945 * config/aarch64/aarch64-simd.md
10946 (vcond<mode>_internal): Handle special cases for constant masks.
10947 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10948 (vcondu<mode><mode>): Likewise.
10949 (vcond<v_cmp_result><mode>): New.
10950
10951 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10952
10953 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10954 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10955 * config/aarch64/aarch64-simd-builtins.def
10956 (cmeq): Update to BUILTIN_VALLDI.
10957 (cmgt): Likewise.
10958 (cmge): Likewise.
10959 (cmle): Likewise.
10960 (cmlt): Likewise.
10961 * config/aarch64/arm_neon.h
10962 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10963 to builtins or C as appropriate.
10964
10965 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10966
10967 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10968 (cmgeu): ...This.
10969 (cmhi): Rename to...
10970 (cmgtu): ...This.
10971 * config/aarch64/aarch64-simd.md
10972 (simd_mode): Add SF.
10973 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10974 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10975 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10976 (cstore<mode>_neg): ...This.
10977 * config/aarch64/iterators.md
10978 (VALLF): new.
10979 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10980 (COMPARISONS): New.
10981 (UCOMPARISONS): Likewise.
10982 (optab): Add missing comparisons.
10983 (n_optab): New.
10984 (cmp_1): Likewise.
10985 (cmp_2): Likewise.
10986 (CMP): Likewise.
10987 (cmp): Remove.
10988 (VCMP_S): Likewise.
10989 (VCMP_U): Likewise.
10990 (V_cmp_result): Add DF, SF modes.
10991 (v_cmp_result): Likewise.
10992 (v): Likewise.
10993 (vmtype): Likewise.
10994 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10995
10996 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10997
10998 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10999 define_insn to define_insn_and_split.
11000 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
11001 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
11002 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
11003 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
11004 (thumb2_negscc): Likewise.
11005
11006 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11007
11008 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
11009
11010 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
11011
11012 * config/arm/thumb2.md: Remove trailing whitespaces.
11013
11014 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11015
11016 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
11017 Use gen_int_mode rather than GEN_INT.
11018
11019 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
11020
11021 * value-prof.c (stream_in_histogram_value): Remove the strayed
11022 debug_gimple_stmt.
11023
11024 2013-04-30 Richard Biener <rguenther@suse.de>
11025
11026 PR middle-end/57122
11027 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11028
11029 2013-04-30 Richard Biener <rguenther@suse.de>
11030
11031 PR middle-end/57107
11032 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11033
11034 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11035
11036 PR rtl-optimization/56957
11037 PR rtl-optimization/57105
11038 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11039 variable. Use just INSN_UID for determining whether an insn
11040 should be only disconnected from the insn stream.
11041 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11042
11043 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11044
11045 PR tree-optimization/57104
11046 * tsan.c (instrument_expr): Don't instrument accesses to
11047 DECL_HARD_REGISTER VAR_DECLs.
11048
11049 2013-04-30 Richard Biener <rguenther@suse.de>
11050
11051 * function.h (loops_for_fn): New inline function.
11052 (set_loops_for_fn): Likewise.
11053 * cfgloop.h (place_new_loop): Add struct function parameter.
11054 (get_loop): Likewise.
11055 (get_loops): Likewise.
11056 (number_of_loops): Likewise.
11057 (fel_next): Adjust.
11058 (fel_init): Likewise.
11059 * cfg.c (get_loop_copy): Adjust.
11060 * cfgloop.c (flow_loops_dump): Likewise.
11061 (record_loop_exits): Likewise.
11062 (verify_loop_structure): Likewise.
11063 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11064 (estimate_reg_pressure_cost): Likewise.
11065 (mark_loop_exit_edges): Likewise.
11066 * cfgloopmanip.c (place_new_loop): Likewise.
11067 (add_loop): Likewise.
11068 (duplicate_loop): Likewise.
11069 * graph.c (draw_cfg_nodes): Likewise.
11070 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11071 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11072 (extract_affine_chrec): Likewise.
11073 (build_scop_iteration_domain): Likewise.
11074 * graphite.c (graphite_initialize): Likewise.
11075 * ira-build.c (create_loop_tree_nodes): Likewise.
11076 (more_one_region_p): Likewise.
11077 (rebuild_regno_allocno_maps): Likewise.
11078 (mark_loops_for_removal): Likewise.
11079 (mark_all_loops_for_removal): Likewise.
11080 (remove_unnecessary_regions): Likewise.
11081 (ira_build): Likewise.
11082 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11083 * loop-init.c (fix_loop_structure): Likewise.
11084 (gate_rtl_move_loop_invariants): Likewise.
11085 (gate_rtl_unswitch): Likewise.
11086 (gate_rtl_unroll_and_peel_loops): Likewise.
11087 (rtl_doloop): Likewise.
11088 * lto-streamer-in.c (input_cfg): Likewise.
11089 * lto-streamer-out.c (output_cfg): Likewise.
11090 * modulo-sched.c (sms_schedule): Likewise.
11091 * predict.c (tree_estimate_probability): Likewise.
11092 (tree_estimate_probability_driver): Likewise.
11093 (estimate_loops): Likewise.
11094 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11095 (move_sese_region_to_fn): Likewise.
11096 (debug_loop_num): Likewise.
11097 * tree-chrec.c (chrec_evaluate): Likewise.
11098 (hide_evolution_in_other_loops_than_loop): Likewise.
11099 (chrec_component_in_loop_num): Likewise.
11100 (reset_evolution_in_loop): Likewise.
11101 (evolution_function_is_invariant_rec_p): Likewise.
11102 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11103 * tree-inline.c (copy_loops): Likewise.
11104 (copy_cfg_body): Likewise.
11105 (tree_function_versioning): Likewise.
11106 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11107 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11108 Likewise.
11109 (add_to_evolution_1): Likewise.
11110 (scev_const_prop): Likewise.
11111 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11112 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11113 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11114 (tree_ssa_lim_initialize): Likewise.
11115 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11116 (verify_loop_closed_ssa): Likewise.
11117 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11118 (tree_ssa_loop_im): Likewise.
11119 (tree_ssa_loop_unswitch): Likewise.
11120 (tree_vectorize): Likewise.
11121 (check_data_deps): Likewise.
11122 (tree_ssa_loop_ivcanon): Likewise.
11123 (tree_ssa_loop_bounds): Likewise.
11124 (tree_complete_unroll): Likewise.
11125 (tree_complete_unroll_inner): Likewise.
11126 (tree_parallelize_loops): Likewise.
11127 (tree_ssa_loop_prefetch): Likewise.
11128 (tree_ssa_loop_ivopts): Likewise.
11129 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11130 * tree-vectorizer.c (vectorize_loops): Likewise.
11131
11132 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11133
11134 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11135 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11136 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11137 with EABI_LINK_SPEC.
11138
11139 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11140
11141 PR target/44578
11142 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11143 alternative.
11144
11145 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11146
11147 PR target/57097
11148 * lra-constraints.c (process_alt_operands): Discourage a bit more
11149 using memory for pseudos. Print cost dump for alternatives.
11150 Modify cost values for conflicts with early clobbers.
11151 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11152
11153 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11154
11155 PR target/57098
11156 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11157
11158 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11159
11160 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11161 from/to S register.
11162 (movdi_aarch64): Support LDR/STR from/to D register.
11163
11164 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11165
11166 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11167 or higher by default.
11168
11169 2013-04-29 Richard Biener <rguenther@suse.de>
11170
11171 PR middle-end/57075
11172 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11173 even if not adding abnormal edges for calls that can make
11174 abnormal gotos.
11175
11176 2013-04-29 Richard Biener <rguenther@suse.de>
11177
11178 PR middle-end/57103
11179 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11180 TREE_BLOCK.
11181 (move_stmt_r): Remove redundant checking.
11182
11183 2013-04-29 Teresa Johnson <tejohnson@google.com>
11184
11185 PR bootstrap/57077
11186 * basic-block.h (apply_scale): New function.
11187 (apply_probability): Use apply_scale.
11188 * gimple-streamer-in.c (input_bb): Ditto.
11189 * lto-streamer-in.c (input_cfg): Ditto.
11190 * lto-cgraph.c (merge_profile_summaries): Ditto.
11191 * tree-optimize.c (execute_fixup_cfg): Ditto.
11192 * tree-inline.c (copy_bb): Update comment to use apply_scale.
11193 (copy_edges_for_bb): Ditto.
11194 (copy_cfg_body): Ditto.
11195
11196 2013-04-29 Tom de Vries <tom@codesourcery.com>
11197
11198 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
11199 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
11200 (tail_merge_optimize): Handle current_loops == NULL.
11201
11202 2013-04-26 Jeff Law <law@redhat.com>
11203
11204 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
11205 (simplify_cond_using_ranges): Generalize code to simplify
11206 COND_EXPRs where one argument is a constant and the other
11207 is an SSA_NAME created by an integral type conversion.
11208
11209 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11210
11211 * config/arm/arm.md (store_minmaxsi): Use only when
11212 optimize_insn_for_size_p.
11213
11214 2013-04-29 Christian Bruel <christian.bruel@st.com>
11215
11216 PR target/57108
11217 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
11218
11219 2013-04-29 Richard Biener <rguenther@suse.de>
11220
11221 PR middle-end/57089
11222 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
11223 loop tree make sure to schedule a fixup for the child as well.
11224 (expand_omp_for_generic): Properly add loops.
11225 (expand_omp_for_static_nochunk): Likewise.
11226 (expand_omp_for_static_chunk): Likewise.
11227 (expand_omp_for): For the degenerate case fixup loops.
11228 (expand_omp_sections): Fix default bb placement in loops.
11229 (expand_omp_atomic_pipeline): Properly add loops.
11230
11231 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11232
11233 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
11234
11235 2013-04-29 Tom de Vries <tom@codesourcery.com>
11236
11237 * tree-ssa-tail-merge.c: Update header comment.
11238
11239 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11240
11241 * config/aarch64/arm_neon.h
11242 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
11243 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
11244 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
11245 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
11246 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
11247 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
11248 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
11249 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
11250
11251 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11252
11253 * config/aarch64/aarch64-simd.md
11254 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
11255 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
11256 fix_trunc, fixuns_trunc.
11257 (ftrunc<VDQF:mode>2): New.
11258 * config/aarch64/iterators.md (optab): Add fix, fixuns.
11259 (fix_trunc_optab): New.
11260
11261 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11262
11263 * config/aarch64/aarch64-builtins.c
11264 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
11265 iceilf, lround, iroundf.
11266
11267 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11268
11269 PR target/54349
11270 * config/i386/i386.h (enum ix86_tune_indices)
11271 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
11272 New, split from X86_TUNE_INTER_UNIT_MOVES.
11273 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
11274 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
11275 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
11276 (TARGET_INTER_UNIT_MOVES): Remove.
11277 * config/i386/i386.c (initial_ix86_tune_features): Update.
11278 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
11279 (ix86_expand_convert_uns_didf_sse): Use
11280 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11281 (ix86_expand_vector_init_one_nonzero): Ditto.
11282 (ix86_expand_vector_init_interleave): Ditto.
11283 (inline_secondary_memory_needed): Return true for moves from SSE class
11284 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
11285 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
11286 * config/i386/constraints.md (Yi, Ym): Depend on
11287 TARGET_INTER_UNIT_MOVES_TO_VEC.
11288 (Yj, Yn): New constraints.
11289 * config/i386/i386.md (*movdi_internal): Change constraints of
11290 operand 1 from Yi to Yj and from Ym to Yn.
11291 (*movsi_internal): Ditto.
11292 (*movdf_internal): Ditto.
11293 (*movsf_internal): Ditto.
11294 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
11295 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11296 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
11297 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
11298 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
11299 * config/i386/sse.md (movdi_to_sse): Ditto.
11300 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
11301 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
11302 TARGET_INTER_UNIT_MOVES.
11303 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
11304 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
11305 instead of TARGET_INTER_UNIT_MOVES.
11306 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
11307 operand 1 from Yi to Yj and from Ym to Yn.
11308
11309 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11310
11311 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
11312 (float_truncate_hi_): Likewise.
11313 (float_extend_lo_): Likewise.
11314 (float_truncate_lo_): Likewise.
11315 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
11316 (aarch64_float_extend_lo_v2df): Likewise.
11317 (vec_unpacks_hi_v4sf): Likewise.
11318 (aarch64_float_truncate_lo_v2sf): Likewise.
11319 (aarch64_float_truncate_hi_v4sf): Likewise.
11320 (vec_pack_trunc_v2df): Likewise.
11321 (vec_pack_trunc_df): Likewise.
11322
11323 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11324
11325 * config/aarch64/aarch64-builtins.c
11326 (aarch64_fold_builtin): Fold float conversions.
11327 * config/aarch64/aarch64-simd-builtins.def
11328 (floatv2si, floatv4si, floatv2di): New.
11329 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
11330 * config/aarch64/aarch64-simd.md
11331 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
11332 * config/aarch64/iterators.md (FLOATUORS): New.
11333 (optab): Add float, floatuns.
11334 (su_optab): Likewise.
11335
11336 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11337
11338 * config/aarch64/aarch64-builtins.c
11339 (aarch64_builtin_vectorized_function): Use new names for
11340 fcvt builtins.
11341 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
11342 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
11343 (fcvtzu): Split as...
11344 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
11345 (fcvtas): Split as...
11346 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
11347 (fcvtau): Split as...
11348 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
11349 (fcvtps): Split as...
11350 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
11351 (fcvtpu): Split as...
11352 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
11353 (fcvtms): Split as...
11354 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
11355 (fcvtmu): Split as...
11356 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
11357 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
11358 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
11359 (lfrintnusf, lfrintnudf): Likewise.
11360 * config/aarch64/aarch64-simd.md
11361 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
11362 define_insn.
11363 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
11364 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
11365 (fcvt_pattern): Likewise.
11366
11367 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11368
11369 * config/aarch64/aarch64-simd.md
11370 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
11371 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
11372
11373 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11374
11375 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
11376 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
11377 (vrnd<a,m,n,p>_f32): Implement using builtins.
11378 (vrnd<i,x><q>_f<32, 64>): New.
11379
11380 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11381
11382 * config/aarch64/aarch64-builtins.c
11383 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
11384 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
11385 (frintz): Rename to...
11386 (btrunc): ...this.
11387 (frintp): Rename to...
11388 (ceil): ...this.
11389 (frintm): Rename to...
11390 (floor): ...this.
11391 (frinti): Rename to...
11392 (nearbyint): ...this.
11393 (frintx): Rename to...
11394 (rint): ...this.
11395 (frinta): Rename to...
11396 (round): ...this.
11397 * config/aarch64/aarch64-simd.md
11398 (aarch64_frint<frint_suffix><mode>): Delete.
11399 (<frint_pattern><mode>2): Convert to insn.
11400 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
11401 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
11402 (frint_pattern): Likewise.
11403 (frint_suffix): Likewise.
11404
11405 2013-04-29 Richard Biener <rguenther@suse.de>
11406
11407 PR tree-optimization/57081
11408 * loop-init.c: Include tree-flow.h.
11409 (loop_optimizer_finalize): Free number of iteration estimates.
11410 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
11411
11412 2013-04-29 Jakub Jelinek <jakub@redhat.com>
11413
11414 PR tree-optimization/57083
11415 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
11416 non-singleton shift count range, zero extend low_bound for uns case.
11417
11418 * config/i386/predicates.md (general_vector_operand): New predicate.
11419 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
11420 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
11421 if they aren't nonimmediate operands. If their original values
11422 satisfy const_vector_equal_evenodd_p, don't shift them.
11423 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
11424 predicates. For the SSE4.1 case force operands[{1,2}] into registers
11425 if not nonimmediate_operand.
11426 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
11427 instead of register_operand.
11428 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
11429
11430 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
11431
11432 * stor-layout.c (finalize_size_functions): Allocate a structure and
11433 reset cfun before dumping the functions.
11434
11435 2013-04-27 Jakub Jelinek <jakub@redhat.com>
11436
11437 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
11438
11439 PR target/56866
11440 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
11441 use xop_pmacsdqh if uns_p.
11442 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
11443 the immediate rotate count.
11444
11445 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
11446
11447 * rtl.h (struct rtx_def): Add comment for field jump.
11448 (LRA_SUBREG_P): New macro.
11449 * recog.c (register_operand): Check LRA_SUBREG_P.
11450 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11451 stack frame.
11452 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11453 (lra_final_code_change): Skip subreg change for operators.
11454 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11455 if there are no operand changes.
11456 * lra-constraints.c (curr_insn_set): New.
11457 (match_reload): Set LRA_SUBREG_P.
11458 (emit_spill_move): Ditto.
11459 (check_and_process_move): Use curr_insn_set. Process only single
11460 set insns. Don't initialize sec_mem_p and change_p.
11461 (simplify_operand_subreg): Use LRA_SUBREG_P.
11462 (reg_in_class_p): New function.
11463 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11464 of #ifdef. Add code to remove cycling.
11465 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11466 non-null disp. Reload inner instead of disp when base and index
11467 are null. Try to put lo_sum into register.
11468 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11469 (check_and_process_move): Move code for move cost check to
11470 simple_move_p. Remove equiv_substitution.
11471 (simple_move_p): New function.
11472 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11473 curr_insn_set. Call check_and_process_move only for single set
11474 insns. Use the new function. Move call of check_and_process_move
11475 after operand equiv substitution and address process.
11476
11477 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11478
11479 PR go/57045
11480 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
11481 with nonlocal goto receivers or returns twice calls, ignore
11482 unininitialized values from abnormal edges to nl goto receiver
11483 or returns twice call.
11484
11485 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11486
11487 PR tree-optimization/57051
11488 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
11489 and VEC_RSHIFT_EXPR if shift count is a multiple of element
11490 bitsize.
11491
11492 2013-04-26 Richard Biener <rguenther@suse.de>
11493
11494 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
11495 (expand_omp_taskreg): Likewise. Mark loops for fixup.
11496 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
11497 (fixup_loop_arrays_after_move): New function.
11498 (move_sese_region_to_fn): Properly outline the loop tree parts
11499 of the SESE region.
11500
11501 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
11502
11503 * config/i386/i386.md (type, unit): Fix long lines.
11504
11505 2013-04-26 Richard Biener <rguenther@suse.de>
11506
11507 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
11508 (lto-streamer-out.o): Likewise.
11509 * cfgloop.c (init_loops_structure): Export, add struct function
11510 argument and adjust.
11511 (flow_loops_find): Adjust.
11512 * cfgloop.h (enum loop_estimation): Add EST_LAST.
11513 (init_loops_structure): Declare.
11514 * lto-streamer-in.c: Include cfgloop.h.
11515 (input_cfg): Input the loop tree.
11516 * lto-streamer-out.c: Include cfgloop.h.
11517 (output_cfg): Output the loop tree.
11518 (output_struct_function_base): Do not drop PROP_loops.
11519
11520 2013-03-26 Richard Biener <rguenther@suse.de>
11521
11522 * tree-cfg.c (execute_build_cfg): Build the loop tree.
11523 (pass_build_cfg): Provide PROP_loops.
11524 (move_sese_region_to_fn): Remove loops that are outlined into fn
11525 for now.
11526 * tree-inline.c: Include cfgloop.h.
11527 (initialize_cfun): Do not drop PROP_loops.
11528 (copy_loops): New function.
11529 (copy_cfg_body): Copy loop structure.
11530 (tree_function_versioning): Initialize destination loop tree.
11531 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
11532 (pass_parallelize_loops): Do IL verification.
11533 * loop-init.c (loop_optimizer_init): Fixup loops if required.
11534 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
11535 the CFG make sure we fixup loops as well.
11536 * tree-ssa-tail-merge.c: Include cfgloop.h.
11537 (replace_block_by): When merging loop latches mark loops for fixup.
11538 * lto-streamer-out.c (output_struct_function_base): Drop
11539 PROP_loops for now.
11540 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
11541 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
11542 * ipa-split.c: Include cfgloop.h.
11543 (split_function): Add the new return block to the loop tree root.
11544 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
11545 whether we have removed the forwarder block.
11546 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
11547 * cfgloop.h (place_new_loop): Declare.
11548 * cfgloopmanip.c (place_new_loop): Export.
11549 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
11550 (tree-switch-conversion.o): Likewise.
11551 (tree-complex.o): Likewise.
11552 (tree-inline.o): Likewise.
11553 (tree-ssa-tailmerge.o): Likewise.
11554 (ipa-split.o): Likewise.
11555 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
11556 (tree-ssa-copy.o): Likewise.
11557 * tree-switch-conversion.c: Include cfgloop.h
11558 (process_switch): If we emit a bit-test cascade, schedule loops
11559 for fixup.
11560 * tree-complex.c: Include cfgloop.h.
11561 (expand_complex_div_wide): Properly add new basic-blocks to loops.
11562 * asan.c: Include cfgloop.h.
11563 (create_cond_insert_point): Properly add new basic-blocks to
11564 loops, schedule loop fixup.
11565 * cfgloop.c (verify_loop_structure): Check that looks are not
11566 marked for fixup.
11567 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
11568 to loops.
11569 (expand_omp_for_generic): Likewise.
11570 (expand_omp_sections): Likewise.
11571 (expand_omp_atomic_pipeline): Schedule loops for fixup.
11572 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
11573 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
11574 is initialized, not when loops are present.
11575 * tree-parloops.c (parallelize_loops): Remove checking here.
11576 * passes.c (init_optimization_passes): Schedule a copy-propagation
11577 pass before complete unrolling of inner loops.
11578
11579 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11580
11581 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
11582 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
11583 (colorize_init): Add argument to _WIN32 version.
11584 * toplev.c: Include diagnostic-color.h.
11585 (process_options): Default to -fdiagnostics-color=auto if
11586 GCC_COLORS env var is in the environment.
11587 * common.opt (fdiagnostics-color=): Add Var and Init.
11588 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
11589 env var is in the environment, the default is auto rather than never.
11590
11591 * diagnostic.h (file_name_as_prefix): Add context argument.
11592 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
11593 the string as locus.
11594 * langhooks.c (lhd_print_error_function): Adjust caller.
11595
11596 2013-04-25 Lawrence Crowl <crowl@google.com>
11597
11598 * var-tracking.c (shared_hash_def::htab):
11599 Change type to hash_table. Update dependent calls and types.
11600
11601 2013-04-25 Lawrence Crowl <crowl@google.com>
11602
11603 * Makefile.in: Update as needed below.
11604
11605 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
11606 Move declaration to after the type's method definitons.
11607
11608 * attribs.c (htab_t scoped_attributes::attribute_hash):
11609 Change type to hash_table. Update dependent calls and types.
11610
11611 * bitmap.c (htab_t bitmap_desc_hash):
11612 Change type to hash_table. Update dependent calls and types.
11613
11614 * cselib.c (htab_t cselib_hash_table):
11615 Change type to hash_table. Update dependent calls and types.
11616
11617 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
11618 (hash_string_slot_node): Move implementation into lto-streamer.h
11619 struct string_slot_hasher.
11620 (eq_string_slot_node): Likewise.
11621
11622 * data-streamer-out.c: Update output_block::string_hash_table
11623 dependent calls and types.
11624
11625 * dwarf2cfi.c (htab_t trace_index):
11626 Change type to hash_table. Update dependent calls and types.
11627
11628 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
11629 Change type to hash_table. Update dependent calls and types.
11630 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
11631 (htab_t optimize_external_refs::map): Likewise.
11632 (htab_t output_comp_unit::extern_map): Likewise.
11633 (htab_t output_comdat_type_unit::extern_map): Likewise.
11634 (htab_t output_macinfo::macinfo_htab): Likewise.
11635 (htab_t optimize_location_lists::htab): Likewise.
11636 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
11637
11638 * except.c (htab_t ehspec_hash_type):
11639 Change type to hash_table. Update dependent calls and types.
11640 (assign_filter_values::ttypes): Likewise.
11641 (assign_filter_values::ehspec): Likewise.
11642 (sjlj_assign_call_site_values::ar_hash): Likewise.
11643 (convert_to_eh_region_ranges::ar_hash): Likewise.
11644
11645 * gcse.c (htab_t pre_ldst_table):
11646 Change type to hash_table. Update dependent calls and types.
11647
11648 * ggc-common.c (htab_t saving_htab):
11649 Change type to hash_table. Update dependent calls and types.
11650 (htab_t loc_hash): Likewise.
11651 (htab_t ptr_hash): Likewise.
11652 (call_count): Rename ggc_call_count.
11653 (call_alloc): Rename ggc_call_alloc.
11654 (loc_descriptor): Rename make_loc_descriptor.
11655 (add_statistics): Rename ggc_add_statistics.
11656
11657 * ggc-common.c (saving_htab):
11658 Change type to hash_table. Update dependent calls and types.
11659
11660 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
11661 (push_gimplify_context): Likewise.
11662 (pop_gimplify_context): Likewise.
11663 (struct gimple_temp_hash_elt): Added.
11664 (struct gimplify_hasher): Likewise.
11665 (struct gimplify_ctx.temp_htab):
11666 Change type to hash_table. Update dependent calls and types.
11667
11668 * gimple-fold.c: Include gimplify-ctx.h.
11669
11670 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
11671 Change type to hash_table. Update dependent calls and types.
11672 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
11673 avoid potential global name collision.
11674
11675 * gimplify.c: Include gimplify-ctx.h.
11676 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
11677 (htab_t gimplify_ctx::temp_htab):
11678 Update dependent calls and types for new type hash_table.
11679 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
11680 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
11681
11682 * gimplify-ctx.h: New.
11683 (struct gimple_temp_hash_elt): Move from gimplify.c.
11684 (class gimplify_hasher): New.
11685 (struct gimplify_ctx): Move from gimple.h.
11686 (htab_t gimplify_ctx::temp_htab):
11687 Change type to hash_table. Update dependent calls and types.
11688
11689 * graphite-clast-to-gimple.c: Include graphite-htab.h.
11690 (htab_t ivs_params::newivs_index):
11691 Change type to hash_table. Update dependent calls and types.
11692 (htab_t ivs_params::params_index): Likewise.
11693 (htab_t print_generated_program::params_index): Likewise.
11694 (htab_t gloog::newivs_index): Likewise.
11695 (htab_t gloog::params_index): Likewise.
11696
11697 * graphite.c: Include graphite-htab.h.
11698 4htab_t graphite_transform_loops::bb_pbb_mapping):
11699 Change type to hash_table. Update dependent calls and types.
11700
11701 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
11702 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
11703 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
11704
11705 * graphite-dependences.c: Include graphite-htab.h.
11706 (loop_is_parallel_p): Change hash table type of parameter.
11707
11708 * graphite-htab.h: New.
11709 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
11710 (extern find_pbb_via_hash): Move from graphite-poly.h.
11711 (extern loop_is_parallel_p): Move from graphite-poly.h.
11712 (extern get_loop_body_pbbs): Move from graphite-poly.h.
11713
11714 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
11715 (extern loop_is_parallel_p): Move to graphite-htab.h.
11716 (extern get_loop_body_pbbs): Move to graphite-htab.h.
11717
11718 * haifa-sched.c (htab_t delay_htab):
11719 Change type to hash_table. Update dependent calls and types.
11720 (htab_t delay_htab_i2): Likewise.
11721
11722 * ira-color.c (htab_t allocno_hard_regs_htab):
11723 Change type to hash_table. Update dependent calls and types.
11724
11725 * ira-costs.c (htab_t cost_classes_htab):
11726 Change type to hash_table. Update dependent calls and types.
11727
11728 * loop-invariant.c (htab_t merge_identical_invariants::eq):
11729 Change type to hash_table. Update dependent calls and types.
11730
11731 * loop-iv.c (htab_t bivs):
11732 Change type to hash_table. Update dependent calls and types.
11733
11734 * loop-unroll.c (htab_t opt_info::insns_to_split):
11735 Change type to hash_table. Update dependent calls and types.
11736 (htab_t opt_info::insns_with_var_to_expand): Likewise.
11737
11738 * lto-streamer.h (struct string_slot): Move from data-streamer.h
11739 (struct string_slot_hasher): New.
11740 (htab_t output_block::string_hash_table):
11741 Change type to hash_table. Update dependent calls and types.
11742
11743 * lto-streamer-in.c (freeing_string_slot_hasher): New.
11744 (htab_t file_name_hash_table):
11745 Change type to hash_table. Update dependent calls and types.
11746
11747 * lto-streamer-out.c: Update output_block::string_hash_table dependent
11748 calls and types.
11749
11750 * lto-streamer.c (htab_t tree_htab):
11751 Change type to hash_table. Update dependent calls and types.
11752
11753 * omp-low.c: Include gimplify-ctx.h.
11754
11755 * passes.c (htab_t name_to_pass_map):
11756 Change type to hash_table. Update dependent calls and types.
11757 (pass_traverse): Rename to passes_pass_traverse.
11758
11759 * plugin.c (htab_t event_tab):
11760 Change type to hash_table. Update dependent calls and types.
11761
11762 * postreload-gcse.c (htab_t expr_table):
11763 Change type to hash_table. Update dependent calls and types.
11764 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
11765
11766 * sese.c (debug_rename_map_1): Make extern.
11767 (htab_t copy_bb_and_scalar_dependences::rename_map):
11768 Change type to hash_table. Update dependent calls and types.
11769
11770 * sese.h (extern debug_rename_map): Move to .c file.
11771
11772 * store-motion.c (htab_t store_motion_mems_table):
11773 Change type to hash_table. Update dependent calls and types.
11774
11775 * trans-mem.c (htab_t tm_new_mem_hash):
11776 Change type to hash_table. Update dependent calls and types.
11777
11778 * tree-browser.c (htab_t TB_up_ht):
11779 Change type to hash_table. Update dependent calls and types.
11780
11781 * tree-cfg.c (htab_t discriminator_per_locus):
11782 Change type to hash_table. Update dependent calls and types.
11783
11784 * tree-complex.c: Include tree-hasher.h
11785 (htab_t complex_variable_components):
11786 Change type to hash_table. Update dependent calls and types.
11787
11788 * tree-eh.c (htab_t finally_tree):
11789 Change type to hash_table. Update dependent calls and types.
11790
11791 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
11792 struct int_tree_hasher.
11793 (extern int_tree_map_eq): Likewise.
11794 (uid_decl_map_hash): Removed.
11795 (extern decl_tree_map_eq): Likewise.
11796
11797 * tree-hasher.h: New.
11798 (struct int_tree_hasher): New.
11799 (typedef int_tree_htab_type): New.
11800
11801 * tree-inline.c: Include gimplify-ctx.h.
11802
11803 * tree-mudflap.c: Include gimplify-ctx.h.
11804
11805 * tree-parloops.c: Include tree-hasher.h.
11806 (htab_t eliminate_local_variables_stmt::decl_address):
11807 Change type to hash_table. Update dependent calls and types.
11808 (htab_t separate_decls_in_region::decl_copies): Likewise.
11809
11810 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11811 Change type to hash_table. Update dependent calls and types.
11812
11813 * tree-sra.c (candidates):
11814 Change type to hash_table. Update dependent calls and types.
11815
11816 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11817 in tree-flow.h.
11818 (int_tree_map_hash): Likewise.
11819
11820 * tree-ssa-dom.c (htab_t avail_exprs):
11821 Change type to hash_table. Update dependent calls and types.
11822
11823 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11824 Change type to hash_table. Update dependent calls and types.
11825
11826 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11827 Change type to hash_table. Update dependent calls and types.
11828
11829 * tree-ssa-phiopt.c (seen_ssa_names):
11830 Change type to hash_table. Update dependent calls and types.
11831
11832 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11833 Change type to hash_table. Update dependent calls and types.
11834
11835 * tree-ssa-uncprop.c (equiv):
11836 Change type to hash_table. Update dependent calls and types.
11837
11838 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11839
11840 PR rtl-optimization/57003
11841 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11842 call note_stores with kill_clobbered_value callback again after
11843 killing regs_invalidated_by_call.
11844
11845 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11846
11847 * config/aarch64/aarch64-simd.md
11848 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11849 (aarch64_simd_bsl<mode>): Likewise.
11850 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11851
11852 2013-04-25 Marek Polacek <polacek@redhat.com>
11853
11854 PR tree-optimization/57066
11855 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11856
11857 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11858
11859 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11860
11861 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11862
11863 * config/aarch64/aarch64-builtins.c
11864 (aarch64_fold_builtin): New.
11865 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11866 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11867 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11868 * config/aarch64/arm_neon.h
11869 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11870
11871 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11872 Tejas Belagod <tejas.belagod@arm.com>
11873
11874 * config/aarch64/aarch64-builtins.c
11875 (aarch64_gimple_fold_builtin): New.
11876 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11877 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11878 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11879 (addvv4sf): Update.
11880 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11881
11882 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11883
11884 * config/aarch64/aarch64.md
11885 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11886
11887 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11888
11889 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11890 (*ngcsi_uxtw): New pattern.
11891
11892 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11893 Julian Brown <julian@codesourcery.com>
11894
11895 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11896 (TB_DREG): Add T_V4HF.
11897 (v4hf_UP): New macro.
11898 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11899 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11900 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11901 built-ins.
11902 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11903 (arm_vector_mode_supported_p): Handle V4HF.
11904 (arm_mangle_map): Handle V4HFmode.
11905 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11906 * config/arm/arm_neon_builtins.def: Add entries for
11907 vcvtv4hfv4sf, vcvtv4sfv4hf.
11908 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11909 (neon_vcvtv4hfv4sf): Likewise.
11910 * config/arm/neon-gen.ml: Handle half-precision floating point
11911 features.
11912 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11913 * config/arm/arm_neon.h: Regenerate.
11914 * config/arm/neon.ml (type elts): Add F16.
11915 (type vectype): Add T_float16x4, T_floatHF.
11916 (type vecmode): Add V4HF.
11917 (type features): Add Requires_FP_bit feature.
11918 (elt_width): Handle F16.
11919 (elt_class): Likewise.
11920 (elt_of_class_width): Likewise.
11921 (mode_of_elt): Refactor.
11922 (type_for_elt): Handle F16, fix error messages.
11923 (vectype_size): Handle T_float16x4.
11924 (vcvt_sh): New function.
11925 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11926 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11927 (string_of_mode): Handle V4HF.
11928 * doc/arm-neon-intrinsics.texi: Regenerate.
11929
11930 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11931
11932 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11933 format specifier in 'X' case.
11934
11935 2013-04-25 Alan Modra <amodra@gmail.com>
11936
11937 PR target/57052
11938 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11939 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11940 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11941 Repeat for many other rotate/shift and mask patterns using subregs.
11942 Name lshiftrt insns.
11943 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11944 on WORDS_BIG_ENDIAN.
11945
11946 2013-04-25 Alan Modra <amodra@gmail.com>
11947
11948 * config.gcc: Support little-endian powerpc-linux targets.
11949 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11950 (LINK_OS_LINUX_SPEC): Define.
11951 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11952 Preserve MASK_LITTLE_ENDIAN.
11953 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11954 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11955 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11956 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11957 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11958 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11959 Correct fp word order for little-endian. Don't shift toc entries
11960 smaller than a word for little-endian.
11961 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11962 (bswapdi2 splits): Correct low-part subreg for little-endian.
11963 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11964 low/high where such is correct only for be.
11965 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11966 little-endian for -mcall-aixdesc.
11967
11968 2013-04-25 Alan Modra <amodra@gmail.com>
11969
11970 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11971 replace_equiv_address_nv.
11972
11973 2013-04-25 Alan Modra <amodra@gmail.com>
11974
11975 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11976
11977 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11978
11979 Revert:
11980 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11981 * rtl.h (struct rtx_def): ...
11982
11983 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11984
11985 PR rtl-optimizations/57046
11986 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11987 for multi-reg splits.
11988
11989 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11990
11991 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11992
11993 2013-04-24 Sterling Augustine <saugustine@google.com>
11994
11995 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11996 (comp_dir_string, debug_str_dwo_section): New.
11997 (DEBUG_STR_DWO_SECTION): Rename to ...
11998 (DEBUG_DWO_STR_SECTION): ... this.
11999 (DEBUG_NORM_STR_SECTION): Delete.
12000 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
12001 (DEBUG_STR_DWO_SECTION_FLAGS): New.
12002 (find_AT_string): Move most logic to ...
12003 (find_AT_string_in_table): ... here. New.
12004 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
12005 add_skeleton_AT_string. Delete logic.
12006 (output_skeleton_debug_sections): Remove call to
12007 add_top_level_skeleton_die_attrs.
12008 (add_comp_dir_attribute): Move logic to comp_dir_string.
12009 (dwarf2out_init): Initialize debug_str_dwo_section.
12010 (output_indirect_string): Call find_string_form.
12011 (output_indirect_strings): Rewrite.
12012 (prune_unused_types): Empty skeleton_debug_str_hash.
12013 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
12014 (dwarf2out_finish): Call output_indirect_strings.
12015
12016 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12017
12018 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
12019
12020 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
12021
12022 * rtl.h (struct rtx_def): Add comment for field jump.
12023 (LRA_SUBREG_P): New macro.
12024 * recog.c (register_operand): Check LRA_SUBREG_P.
12025 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12026 stack frame.
12027 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12028 (lra_final_code_change): Skip subreg change for operators.
12029 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12030 if there are no operand changes.
12031 * lra-constraints.c (curr_insn_set): New.
12032 (match_reload): Set LRA_SUBREG_P.
12033 (emit_spill_move): Ditto.
12034 (check_and_process_move): Use curr_insn_set. Process only single
12035 set insns. Don't initialize sec_mem_p and change_p.
12036 (simplify_operand_subreg): Use LRA_SUBREG_P.
12037 (reg_in_class_p): New function.
12038 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12039 of #ifdef. Add code to remove cycling.
12040 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12041 non-null disp. Reload inner instead of disp when base and index
12042 are null. Try to put lo_sum into register.
12043 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12044 (check_and_process_move): Move code for move cost check to
12045 simple_move_p. Remove equiv_substitution.
12046 (simple_move_p): New function.
12047 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12048 curr_insn_set. Call check_and_process_move only for single set
12049 insns. Use the new function. Move call of check_and_process_move
12050 after operand equiv substitution and address process.
12051
12052 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12053
12054 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12055 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12056 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12057
12058 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12059
12060 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12061
12062 2013-04-24 Marek Polacek <polacek@redhat.com>
12063
12064 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12065 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12066 (select_loops_exit_conditions): Likewise.
12067 (number_of_iterations_for_all_loops): Likewise.
12068 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12069 (scev_analysis): Likewise.
12070
12071 2013-04-02 Catherine Moore <clm@codesourcery.com>
12072 Chao-ying Fu <fu@mips.com>
12073
12074 * config/mips/micromips.md (jraddiusp): New pattern.
12075 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12076 instruction if possible.
12077
12078 2013-04-24 Alan Modra <amodra@gmail.com>
12079
12080 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12081
12082 2013-04-24 Julian Brown <julian@codesourcery.com>
12083 Chung-Lin Tang <cltang@codesourcery.com>
12084
12085 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12086 dependency behavior in enumeration type DIE generation. Add TODO note
12087 to comments about future DW_FORM_sdata/udata re-work of related code.
12088
12089 2013-04-23 Lawrence Crowl <crowl@google.com>
12090
12091 * Makefile.in: Update as needed below.
12092
12093 * hash-table.h (class hash_table):
12094 Correct many methods with parameter types compare_type to the correct
12095 value_type. (Correct code was unlikely to notice the change.)
12096 (hash_table::elements_with_deleted) New.
12097 (class hashtable::iterator): New.
12098 (hashtable::begin()): New.
12099 (hashtable::end()): New.
12100 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12101
12102 * statistics.c (statistics_hashes):
12103 Change type to hash_table. Update dependent calls and types.
12104
12105 * tree-into-ssa.c (var_infos):
12106 Change type to hash_table. Update dependent calls and types.
12107
12108 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12109 Change type to hash_table. Update dependent calls and types.
12110
12111 * tree-ssa-loop-im.c (struct mem_ref.refs):
12112 Change type to hash_table. Update dependent calls and types.
12113
12114 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12115 Change type to hash_table. Update dependent calls and types.
12116
12117 * tree-ssa-sccvn.c (vn_tables_s::nary):
12118 Change type to hash_table. Update dependent calls and types.
12119 (vn_tables_s::phis): Likewise.
12120 (vn_tables_s::references): Likewise.
12121
12122 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12123 (vn_reference_eq): Update parameter and return types.
12124
12125 * tree-ssa-structalias.c (pointer_equiv_class_table):
12126 Change type to hash_table. Update dependent calls and types.
12127 (location_equiv_class_table): Likewise.
12128
12129 * tree-vect-data-refs.c: Consequential changes for making
12130 peeling a hash_table.
12131
12132 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12133 (destroy_loop_vec_info): Dependent hash_table update.
12134
12135 * tree-vectorizer.h (peeling_htab):
12136 Change type to hash_table. Update dependent calls and types.
12137
12138 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12139
12140 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12141 to check the register content is equal or not.
12142 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12143 to assign register content record.
12144 * lra-eliminations.c (update_reg_eliminate): Use
12145 lra_update_reg_val_offset to update register content offset.
12146 * lra-int.h (struct lra_reg): Add offset member.
12147 (lra_reg_val_equal_p): New static inline function.
12148 (lra_update_reg_val_offset): New static inline function.
12149 (lra_assign_reg_val): New static inline function.
12150 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12151 to assign register content record.
12152 (initialize_lra_reg_info_element): Initial offset to zero.
12153
12154 2013-04-23 Catherine Moore <clm@codesourcery.com>
12155
12156 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12157 operands. Record compression.
12158
12159 2013-04-23 Xinliang David Li <davidxl@google.com>
12160
12161 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12162
12163 2013-04-23 Richard Biener <rguenther@suse.de>
12164
12165 PR middle-end/57036
12166 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12167 parameter, only add abnormal goto edges from the copied body
12168 if the call could perform abnormal gotos.
12169 (copy_cfg_body): Adjust.
12170
12171 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12172
12173 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12174
12175 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12176
12177 * coretypes.h (gimple_stmt_iterator): Add struct to make
12178 compatible with C.
12179
12180 2013-04-23 Richard Biener <rguenther@suse.de>
12181
12182 PR tree-optimization/57026
12183 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12184 from SSA names occuring in abnormal PHI nodes.
12185
12186 2013-04-22 Andi Kleen <ak@linux.intel.com>
12187
12188 * lto/lto.c (print_lto_report_1): Fix LTO report names.
12189
12190 2013-04-22 Andi Kleen <ak@linux.intel.com>
12191
12192 * lto/lto.c (print_lto_report_1): Declare early.
12193 (read_cgraph_and_symbols): Call print_lto_report_1 early.
12194
12195 2013-04-22 Andi Kleen <ak@linux.intel.com>
12196
12197 * common.opt (-flto-report-wpa): Add.
12198 * doc/invoke.texi (-flto-report-wpa): Add.
12199 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
12200 (lto_main): dito.
12201
12202 2013-04-22 Xinliang David Li <davidxl@google.com>
12203
12204 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
12205 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
12206 * Makefile.in: New dependency
12207
12208 David Daney <ddaney.cavm@gmail.com>
12209
12210 * configure.ac (gcc_cv_as_micromips_support): Use the
12211 --fatal-warnings option.
12212 * configure: Regenerate.
12213
12214 2013-04-22 Marek Polacek <polacek@redhat.com>
12215
12216 PR sanitizer/56990
12217 * tsan.c (instrument_expr): Don't instrument expression
12218 in case its size is zero.
12219
12220 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
12221
12222 PR target/57032
12223 Revert:
12224 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
12225
12226 * config/alpha/alpha.c (TARGET_LRA_P): New define.
12227
12228 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12229
12230 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
12231 (gimple_stmt_iterator): New typedef.
12232 * gimple.h (gimple_stmt_iterator): Rename to...
12233 (gimple_stmt_iterator_d): ... This.
12234 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
12235 trees be valid for GIMPLE and GENERIC.
12236 (TARGET_GIMPLE_FOLD_BUILTIN): New.
12237 * gimple-fold.c (gimple_fold_call): Call target hook
12238 gimple_fold_builtin.
12239 * hooks.c (hook_bool_gsiptr_false): New.
12240 * hooks.h (hook_bool_gsiptr_false): New.
12241 * target.def (fold_stmt): New.
12242 * doc/tm.texi: Regenerate.
12243
12244 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
12245
12246 PR target/57018
12247 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
12248 a set sp if no stack realignment.
12249
12250 2013-04-22 Nick Clifton <nickc@redhat.com>
12251
12252 * config.gcc (tilegx-linux): Extend extra_objs rather than
12253 overwriting it.
12254 (tilepro-linux): Likewise.
12255
12256 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12257
12258 * config/aarch64/aarch64-builtins.c
12259 (CF): Remove.
12260 (CF0, CF1, CF2, CF3, CF4, CF10): New.
12261 (VAR<1-12>): Add MAP parameter.
12262 (BUILTIN_*): Likewise.
12263 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
12264 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
12265 (aarch64_ushl_n<mode>): Likewise.
12266 (aarch64_sshr_n<mode>): Likewise.
12267 (aarch64_ushr_n<mode>): Likewise.
12268 (aarch64_<maxmin><mode>): Likewise.
12269 (aarch64_sqrt<mode>): Likewise.
12270 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
12271 (vshr<q>_n_*): Likewise.
12272
12273 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12274
12275 * config/aarch64/aarch64-builtins.c
12276 (aarch64_simd_builtin_type_mode): Handle SF types.
12277 (sf_UP): Define.
12278 (BUILTIN_GPF): Define.
12279 (aarch64_init_simd_builtins): Handle SF types.
12280 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
12281 (frecps): Likewise.
12282 (frecpx): Likewise.
12283 * config/aarch64/aarch64-simd.md
12284 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
12285 (aarch64_frecpe<mode>): New.
12286 (aarch64_frecps<mode>): Likewise.
12287 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
12288 (v8type): Add frecp<esx>.
12289 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
12290 (aarch64_frecps<mode>): Likewise.
12291 * config/aarch64/iterators.md (FRECP): New.
12292 (frecp_suffix): Likewise.
12293 * config/aarch64/arm_neon.h
12294 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
12295
12296 2013-04-22 Christian Bruel <christian.bruel@st.com>
12297
12298 PR target/56995
12299 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
12300 (REG_CLASS_NAMES): Idem.
12301 (REG_CLASS_CONTENTS): Idem.
12302 (REGCLASS_HAS_FP_REG): Idem.
12303 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
12304 (sh_conditional_register_usage): Idem.
12305
12306 2013-04-21 Jeff Law <law@redhat.com>
12307
12308 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
12309 (ssa_forward_propagate_and_combine): Use it.
12310
12311 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12312
12313 * lra.c: Update the flow chart diagram.
12314
12315 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12316
12317 PR rtl-optimization/56847
12318 * lra-constraints.c (process_alt_operands): Discourage alternative
12319 with non-matche doffsettable memory constraint fro memory with
12320 known offset.
12321
12322 2013-04-19 Richard Biener <rguenther@suse.de>
12323
12324 PR tree-optimization/56982
12325 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
12326 function.
12327 * gimplify.c (gimplify_call_expr): Notice special calls.
12328 (gimplify_modify_expr): Likewise.
12329 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
12330 abnormal control flow receivers.
12331 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
12332 in the same way as cfun->has_nonlocal_labels.
12333 (gimple_purge_dead_abnormal_call_edges): Likewise.
12334 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
12335 receivers start a basic-block.
12336
12337 2013-04-19 Richard Biener <rguenther@suse.de>
12338
12339 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
12340 member ...
12341 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
12342 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
12343 (SLP_TREE_LOAD_PERMUTATION): Add.
12344 (vect_transform_slp_perm_load): Adjust prototype.
12345 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
12346 (vect_free_slp_instance): Likewise.
12347 (vect_create_new_slp_node): Likewise.
12348 (vect_supported_slp_permutation_p): Remove.
12349 (vect_slp_rearrange_stmts): Adjust.
12350 (vect_supported_load_permutation_p): Likewise. Inline
12351 vect_supported_slp_permutation_p here.
12352 (vect_analyze_slp_instance): Compute load permutations per
12353 slp node instead of per instance.
12354 (vect_get_slp_defs): Adjust.
12355 (vect_transform_slp_perm_load): Likewise.
12356 (vect_schedule_slp_instance): Remove redundant code.
12357 (vect_schedule_slp): Remove hack for PR56270, add it ...
12358 * tree-vect-stmts.c (vectorizable_load): ... here, do not
12359 CSE loads for SLP. Adjust.
12360
12361 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12362
12363 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
12364 spelling in two comments.
12365
12366 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12367
12368 PR target/56797
12369 * config/arm/arm.c (load_multiple_sequence): Require SP
12370 as base register for loads if SP is in the register list.
12371
12372 2013-04-19 Martin Jambor <mjambor@suse.cz>
12373
12374 PR tree-optimization/56718
12375 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
12376 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
12377 and made public. Adjusted all callers.
12378 (ipa_intraprocedural_devirtualization): New function.
12379 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
12380 (ipa_intraprocedural_devirtualization): Likewise.
12381 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
12382
12383 2013-04-19 Richard Biener <rguenther@suse.de>
12384
12385 PR tree-optimization/57000
12386 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
12387
12388 2013-04-19 Terry Guo <terry.guo@arm.com>
12389
12390 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
12391 Replace with ...
12392 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
12393 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
12394 (cortex_m4_fmacs): Use new reservations.
12395 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
12396
12397 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
12398
12399 PR rtl-optimization/56999
12400 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
12401 related code.
12402 (lra_coalesce): Remove split_origin_bitmap and related code.
12403 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
12404 ranges if necessary.
12405
12406 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
12407
12408 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
12409 New array.
12410 (ix86_expand_call): Remove clobbered_registers array and use
12411 x86_64_ms_sysv_extra_clobbered_registers instead.
12412 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
12413 Declare here.
12414 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
12415 predicate.
12416 * config/i386/i386.md (*call_rex64_ms_sysv): Use
12417 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
12418 (*call_value_rex64_ms_sysv): Ditto.
12419
12420 2013-04-18 Cary Coutant <ccoutant@google.com>
12421
12422 * dwarf2out.c (output_pubnames): Check die_perennial_p of
12423 parent instead of die_mark.
12424
12425 2013-04-18 Diego Novillo <dnovillo@google.com>
12426
12427 * gimple.c (create_gimple_tmp): New.
12428 (get_expr_type): New.
12429 (build_assign): New.
12430 (build_type_cast): New.
12431 * gimple.h (enum ssa_mode): Define.
12432 (gimple_seq_set_location): New.
12433 * asan.c (build_check_stmt): Change some gimple_build_* calls
12434 to use build_assign and build_type_cast.
12435
12436 2013-04-18 Richard Biener <rguenther@suse.de>
12437
12438 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
12439 handle negative step. Remove redundant checks.
12440 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
12441 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
12442 for negative step and grouped loads fail to vectorize.
12443
12444 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12445
12446 * emit-rtl.c (reset_insn_used_flags): New function.
12447 (reset_all_used_flags): Use it.
12448 (verify_insn_sharing): New function.
12449 (verify_rtl_sharing): Fix verification for SEQUENCEs.
12450
12451 2013-04-18 Jakub Jelinek <jakub@redhat.com>
12452
12453 PR tree-optimization/56984
12454 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
12455 and (x >> M) >= N don't register any assertion if N << M is the
12456 minimum value.
12457
12458 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12459
12460 * lower-subreg.c (resolve_simple_move): If called self-recursive,
12461 do not delete_insn insns that have not yet been emitted, only
12462 unlink them with remove_insn.
12463 * df-scan.c (df_insn_delete): Revert r197492.
12464
12465 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
12466
12467 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
12468 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
12469
12470 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12471
12472 * config/arm/arm.md (movsicc_insn): Convert define_insn into
12473 define_insn_and_split.
12474 (and_scc,ior_scc,negscc): Likewise.
12475 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
12476
12477 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12478
12479 * config/arm/arm.c (use_return_insn): Return 0 for targets that
12480 can benefit from using a sequence of LDRD instructions in epilogue
12481 instead of a single LDM instruction.
12482
12483 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
12484
12485 PR 45688
12486 * doc/extend.texi: Fix typo.
12487
12488 2013-04-17 Richard Biener <rguenther@suse.de>
12489
12490 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
12491 (vect_build_slp_tree): ... here.
12492 (vect_build_slp_tree_1): Compute which stmts of the SLP group
12493 match. Remove special-casing of mismatched complex loads.
12494 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
12495 re-try the match with swapped commutative operands.
12496 (vect_supported_load_permutation_p): Remove special-casing of
12497 mismatched complex loads.
12498 (vect_analyze_slp_instance): Adjust.
12499
12500 2013-04-17 Richard Biener <rguenther@suse.de>
12501
12502 PR rtl-optimization/56921
12503 * cfgloop.h (struct loop): Add simple_loop_desc member.
12504 (struct niter_desc): Mark with GTY(()).
12505 (simple_loop_desc): Do not use aux field but simple_loop_desc.
12506 * loop-iv.c (get_simple_loop_desc): Likewise.
12507 (free_simple_loop_desc): Likewise.
12508
12509 Revert
12510 2013-04-16 Richard Biener <rguenther@suse.de>
12511
12512 PR rtl-optimization/56921
12513 * loop-init.c (pass_rtl_move_loop_invariants): Add
12514 TODO_do_not_ggc_collect to todo_flags_finish.
12515 (pass_rtl_unswitch): Same.
12516 (pass_rtl_unroll_and_peel_loops): Same.
12517 (pass_rtl_doloop): Same.
12518
12519 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
12520
12521 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
12522 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
12523 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
12524 references.
12525 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
12526 * tree-streamer.c (record_common_node): Adjust reference in comment.
12527
12528 2013-04-17 Terry Guo <terry.guo@arm.com>
12529
12530 * config/arm/cortex-m4.md: Add a new bypass.
12531
12532 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12533
12534 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
12535 New pattern.
12536 (*subs_<optab><mode>_multp2): New pattern.
12537 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12538 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12539
12540 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12541
12542 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
12543 (*subs_mul_imm_<mode>): New pattern.
12544
12545 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
12546
12547 PR target/56948
12548 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
12549 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
12550 (vsx_movti_32bit): Change j->wa to O->wa.
12551
12552 2013-04-16 Richard Biener <rguenther@suse.de>
12553
12554 PR rtl-optimization/56921
12555 * loop-init.c (pass_rtl_move_loop_invariants): Add
12556 TODO_do_not_ggc_collect to todo_flags_finish.
12557 (pass_rtl_unswitch): Same.
12558 (pass_rtl_unroll_and_peel_loops): Same.
12559 (pass_rtl_doloop): Same.
12560
12561 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
12562
12563 * config/arm/arm.c (emit_multi_reg_push): New declaration
12564 for an existing function.
12565 (arm_emit_strd_push): New function.
12566 (arm_expand_prologue): Used here.
12567 (arm_emit_ldrd_pop): New function.
12568 (arm_expand_epilogue): Used here.
12569 (arm_get_frame_offsets): Update condition.
12570 (arm_emit_multi_reg_pop): Add a special case for load of a single
12571 register with writeback.
12572
12573 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
12574
12575 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
12576 description.
12577
12578 2013-04-16 Richard Biener <rguenther@suse.de>
12579
12580 PR tree-optimization/56756
12581 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
12582 (first_mem_ref_loc): New.
12583 (execute_sm): Place the load temporarily before a previous
12584 access instead of in the latch edge to ensure its SSA dependencies
12585 are defined at points dominating the load.
12586
12587 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
12588
12589 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
12590 correct fix by moving header and footer insn to the footer of
12591 the merged basic block. Clear BB_END of the merged-away block.
12592
12593 PR middle-end/43631
12594 * emit-rtl.c (make_note_raw): New function.
12595 (link_insn_into_chain): New static inline function.
12596 (add_insn): Use it.
12597 (add_insn_before, add_insn_after): Factor insn chain linking code...
12598 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
12599 using link_insn_into_chain.
12600 (note_outside_basic_block_p): New helper function for emit_note_after
12601 and emit_note_before.
12602 (emit_note_after): Use nobb variant of add_insn_after if the note
12603 should not be contained in a basic block.
12604 (emit_note_before): Use nobb variant of add_insn_before if the note
12605 should not be contained in a basic block.
12606 (emit_note_copy): Use make_note_raw.
12607 (emit_note): Likewise.
12608 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
12609 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
12610 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
12611 the moved barrier the tail of the basic block it follows.
12612 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
12613
12614 2013-04-15 Jakub Jelinek <jakub@redhat.com>
12615
12616 PR tree-optimization/56962
12617 * gimple-ssa-strength-reduction.c (record_increment): Only set
12618 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
12619 either rhs1 or rhs2 is equal to c->base_expr.
12620
12621 2013-04-15 Richard Biener <rguenther@suse.de>
12622
12623 PR tree-optimization/56933
12624 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
12625 member.
12626 (GROUP_READ_WRITE_DEPENDENCE): Remove.
12627 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
12628 * tree-vect-data-refs.c (vect_analyze_group_access): Move
12629 dependence check ...
12630 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
12631 ... here.
12632 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
12633 GROUP_READ_WRITE_DEPENDENCE.
12634
12635 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12636
12637 * emit-rtl.c (reset_all_used_flags): New function.
12638 (verify_rtl_sharing): Call reset_all_used_flags before and after
12639 performing the checks.
12640
12641 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12642
12643 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
12644 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
12645 * config/arm/constraints.md (De): New constraint.
12646 * config/arm/neon.md (anddi3_neon): Delete.
12647 (neon_vand<mode>): Expand to standard anddi3 pattern.
12648 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
12649 Move earlier in the file.
12650 (neon_inv_logic_op2): Likewise.
12651 (arm_anddi_operand_neon): New predicate.
12652
12653 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12654
12655 * configure.ac (gcc_cv_ld_as_needed): Set
12656 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
12657 Use -z ignore, -z record on *-*-solaris2*.
12658 (HAVE_LD_AS_NEEDED): Update comment.
12659 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
12660 * configure: Regenerate.
12661 * config.in: Regenerate.
12662 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
12663 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
12664 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
12665 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
12666 equivalents. Fix markup.
12667 * doc/tm.texi: Regenerate.
12668
12669 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
12670
12671 * config/i386/i386.opt: New option mstack-protector-guard=.
12672 * config/i386/i386-opts.h: Add enum stack_protector_guard.
12673 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
12674 TARGET_SSP_TLS_GUARD.
12675 * config/i386/i386.c (ix86_option_override_internal): Set
12676 ix86_stack_protector_guard.
12677 * config/i386/i386.md (stack_protect_set): Enable for
12678 TARGET_SSP_TLS_GUARD only.
12679 (stack_protect_set_<mode>): Ditto.
12680 (stack_protect_test): Ditto.
12681 (stack_protect_test_<mode>): Ditto.
12682 * doc/invoke.texi (i386 Option): Document.
12683
12684 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
12685
12686 PR target/56890
12687 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
12688 (S_MODES): Set H_MODE bit.
12689 (SF_MODES): Set only S_MODE and SF_MODE bits.
12690 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
12691 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
12692 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
12693 <MODE_FLOAT>: Likewise.
12694
12695 2013-04-15 Joey Ye <joey.ye@arm.com>
12696
12697 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
12698
12699 2013-04-15 Joey Ye <joey.ye@arm.com>
12700
12701 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
12702 for real far jump.
12703 (thumb_far_jump_used_p): Count instruction size and set
12704 far_jump_used.
12705
12706 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
12707
12708 * reorg.c (fill_simple_delay_slots): Reindent block of code.
12709 * resource.c (mark_target_live_regs): Reformat conditional block.
12710
12711 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12712
12713 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
12714 notes, they are emitted only just before final.
12715 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
12716
12717 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12718
12719 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
12720 * cfgrtl.c (delete_insn): Call it here instead.
12721 * lra-spills.c (lra_final_code_change): Use delete_insn.
12722 * haifa-sched.c (sched_remove_insn): Likewise.
12723 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
12724 returning to the nop pool.
12725 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
12726 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
12727
12728 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12729
12730 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
12731 * doc/tm.texi: Regenerated.
12732
12733 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
12734
12735 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
12736 QImode checks.
12737
12738 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12739
12740 * df-core.c (df_find_def): Compare register numbers.
12741 (df_find_use): Likewise.
12742
12743 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
12744
12745 PR target/56903
12746 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
12747 lra_in_progress for return.
12748
12749 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12750
12751 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
12752 define_insn into define_insn_and_split and emit movsicc patterns.
12753
12754 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12755
12756 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
12757
12758 2013-04-12 Richard Biener <rguenther@suse.de>
12759
12760 * tree-pass.h (TODO_do_not_ggc_collect): New.
12761 * passes.c (execute_one_ipa_transform_pass): Honor
12762 TODO_do_not_ggc_collect.
12763 (execute_one_pass): Likewise.
12764
12765 Revert
12766 2013-04-10 Richard Biener <rguenther@suse.de>
12767
12768 * passes.c (init_optimization_passes): Remove reload pass.
12769 * ira.c (do_reload): Merge into ...
12770 (ira): ... this.
12771 (rest_of_handle_reload): Remove.
12772 (pass_reload): Likewise.
12773 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12774 of reload for vzeroupper pass placement.
12775
12776 2013-04-12 Jakub Jelinek <jakub@redhat.com>
12777
12778 PR tree-optimization/56918
12779 PR tree-optimization/56920
12780 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
12781 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
12782 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
12783 use wide_mul_with_sign method.
12784
12785 2013-04-12 Richard Biener <rguenther@suse.de>
12786
12787 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
12788 not be considered a gimple constant.
12789
12790 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
12791
12792 * fold-const.c (const_binop): Handle vector shifts by a scalar.
12793 (fold_binary_loc): Call const_binop also for mixed vector-scalar
12794 operations.
12795
12796 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
12797 Jakub Jelinek <jakub@redhat.com>
12798
12799 * opts.c: Include diagnostic-color.h.
12800 (common_handle_option): Handle OPT_fdiagnostics_color_.
12801 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12802 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12803 (diagnostic-color.o): New.
12804 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12805 (diagnostic_color_rule): New enum.
12806 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12807 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12808 the location string.
12809 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12810 either NULL, or color kind.
12811 * diagnostic-color.c: New file.
12812 * diagnostic-color.h: New file.
12813 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12814 arguments.
12815 * doc/invoke.texi (-fdiagnostics-color): Document.
12816 * pretty-print.h (pp_show_color): Define.
12817 (struct pretty_print_info): Add show_color field.
12818 * diagnostic.c: Include diagnostic-color.h.
12819 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12820 macros. Colorize error:, warning: etc. strings and also the location
12821 string.
12822 (diagnostic_show_locus): Colorize the caret line.
12823 * pretty-print.c: Include diagnostic-color.h.
12824 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12825 inside of %< %> quotes or quoted through q format modifier.
12826
12827 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12828
12829 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12830
12831 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12832
12833 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12834 code in CC_NZ mode.
12835 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12836 pattern.
12837
12838 2013-04-11 Marek Polacek <polacek@redhat.com>
12839
12840 PR tree-optimization/48184
12841 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12842
12843 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12844
12845 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12846 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12847 (skip_simple_arithmetic): Tidy up.
12848 * tree.h (skip_simple_constant_arithmetic): Declare.
12849
12850 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12851
12852 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12853
12854 2013-04-11 Richard Biener <rguenther@suse.de>
12855
12856 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12857 generate vector constants.
12858
12859 2013-04-11 Richard Biener <rguenther@suse.de>
12860
12861 PR tree-optimization/56878
12862 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12863 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12864 New function.
12865 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12866 Prefer to align the DR with the most invariant base address.
12867
12868 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12869
12870 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12871 comment.
12872
12873 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12874
12875 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12876 floating-point vector comparisons against 0.
12877
12878 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12879
12880 PR tree-optimization/56899
12881 * fold-const.c (extract_muldiv_1): Apply distributive law
12882 only if TYPE_OVERFLOW_WRAPS (ctype).
12883
12884 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12885
12886 PR target/56124
12887 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12888 loading has side effect.
12889
12890 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12891
12892 * config/sparc/sparc.c: Include tree-pass.h.
12893 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12894 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12895 head of file. Change return type. Split off gate function.
12896 (sparc_gate_work_around_errata): New function.
12897 (pass_work_around_errata): New pass definition.
12898 (insert_pass_work_around_errata) New pass insert definition to
12899 insert pass_work_around_errata just after delayed-branch scheduling.
12900 (sparc_option_override): Insert the pass.
12901 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12902
12903 2013-04-10 David S. Miller <davem@davemloft.net>
12904
12905 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12906 or -mcpu=hypersparc.
12907
12908 * target.def (cstore_mode): New hook.
12909 * target.h: Include insn-codes.h
12910 * targhooks.c: Likewise.
12911 (default_cstore_mode): New function.
12912 * targhooks.h: Declare it.
12913 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12914 * doc/tm.texi: Rebuild.
12915 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12916 target hook, rather than inspecting the insn_data.
12917 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12918 (TARGET_CSTORE_MODE): Redefine.
12919 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12920 result patterns.
12921 * config/sparc/predicates.md (cstore_result_operand): New special
12922 predicate.
12923 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12924 Use it for operand 0.
12925 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12926 (*snesi_special): Likewise.
12927 (*snesi_zero): Likewise.
12928 (*seqsi_zero): Likewise.
12929 (*sltu_insn): Likewise.
12930 (*sgeu_insn): Likewise.
12931 (*seqdi_special): Make operand 0 and comparison operation be of
12932 DImode.
12933 (*snedi_special): Likewise.
12934 (*snedi_special_vis3): Likewise.
12935 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12936 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12937 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12938 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12939 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12940 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12941 (*sltu_extend_sp64): Likewise.
12942 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12943 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12944 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12945 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12946 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12947
12948 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12949
12950 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12951 (aarch64_start_file): Use the new function.
12952
12953 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12954 Jason Merrill <jason@redhat.com>
12955
12956 * common.opt: Add -gdwarf.
12957 * opts.c (common_handle_option): Handle it.
12958 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12959
12960 2013-04-10 Richard Biener <rguenther@suse.de>
12961
12962 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12963 (execute_one_ipa_transform_pass): But unconditionally here.
12964 (execute_one_pass): And here.
12965 (init_optimization_passes): Remove reload pass.
12966 * tree-pass.h (TODO_ggc_collect): Remove.
12967 (pass_reload): Likewise.
12968 * ira.c (do_reload): Merge into ...
12969 (ira): ... this.
12970 (rest_of_handle_reload): Remove.
12971 (pass_reload): Likewise.
12972 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12973 of reload for vzeroupper pass placement.
12974 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12975 and todo_flags_finish of all passes.
12976
12977 2013-04-10 Richard Biener <rguenther@suse.de>
12978
12979 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12980 first_const_oprnd field, rename first_def_type to first_op_type.
12981 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12982 (vect_get_and_check_slp_defs): Always use the type of the
12983 operand. Allow mixed vect_external_def, vect_constant_def types.
12984 (vect_get_constant_vectors): Handle mixed vect_external_def,
12985 vect_constant_def types.
12986
12987 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12988
12989 PR tree-optimization/55524
12990 * tree-ssa-math-opts.c
12991 (convert_mult_to_fma): Don't use an fms construct
12992 when we don't have an fms operation, but fnma, and it looks
12993 likely that we'll be able to use the latter.
12994
12995 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12996
12997 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12998 function.
12999 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
13000 inline fail caused by overwritable functions.
13001
13002 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
13003
13004 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
13005 unnecessary bits in the constant power of two case.
13006
13007 2013-04-10 Richard Biener <rguenther@suse.de>
13008
13009 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
13010 broken code swapping operands.
13011 (vect_build_slp_tree): Do not compute load permutations here.
13012 (vect_analyze_slp_instance): Compute load permutations here,
13013 after building the SLP tree.
13014
13015 2013-04-09 Christian Bruel <christian.bruel@st.com>
13016
13017 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
13018 of next/prev_real_insn.
13019
13020 2013-04-09 Jan Hubicka <jh@suse.cz>
13021
13022 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13023 Drop aliased parameter.
13024 (function_and_variable_visibility): Do not handle alias pairs.
13025 * cgraph.c (varpool_externally_visible_p): Update prototype.
13026 * varpool.c (varpool_add_new_variable): Update.
13027
13028 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13029
13030 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13031
13032 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13033
13034 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13035
13036 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13037
13038 2013-04-09 Marek Polacek <polacek@redhat.com>
13039
13040 PR tree-optimization/48762
13041 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13042
13043 2013-04-09 Richard Biener <rguenther@suse.de>
13044
13045 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13046 dealing with cost.
13047 (vect_build_slp_tree): Likewise.
13048 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13049 calculating the cost of a SLP instance.
13050 (vect_analyze_slp_instance): Use it from here, after building
13051 the SLP tree.
13052
13053 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13054
13055 PR middle-end/56883
13056 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13057 expand_omp_for_static_chunk): Use simple_p = true in
13058 force_gimple_operand_gsi calls when assigning to addressable decls.
13059
13060 2013-04-09 Jeff Law <law@redhat.com>
13061
13062 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13063 when the boolean was created by converting a wider object which
13064 had a boolean range.
13065
13066 2013-04-09 Richard Biener <rguenther@suse.de>
13067
13068 * tree-vectorizer.h (slp_void_p): Remove.
13069 (slp_tree): Typedef before _slp_tree declaration.
13070 (struct _slp_tree): Use a vector of slp_tree as children.
13071 (vect_get_place_in_interleaving_chain): Remove.
13072 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13073 Move ...
13074 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13075 and make static.
13076 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13077 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13078 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13079 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13080 Use slp_node instead of slp_void_p and adjust.
13081
13082 2013-04-09 Richard Biener <rguenther@suse.de>
13083
13084 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13085 work that is not necessary.
13086
13087 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13088
13089 PR tree-optimization/56854
13090 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13091 forward into clobber stmts if it would change MEM_REF lhs into
13092 non-MEM_REF.
13093
13094 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13095
13096 * tree.c (type_hash_lookup, type_hash_add): Make static.
13097 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13098
13099 2013-04-09 Richard Biener <rguenther@suse.de>
13100
13101 * tree.h (unsave_expr_now): Remove.
13102 * tree-inline.c (mark_local_for_remap_r): Remove.
13103 (unsave_expr_1): Likewise.
13104 (unsave_r): Likewise.
13105 (unsave_expr_now): Likewise.
13106 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13107 (propagate_tree_value): Likewise.
13108
13109 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13110
13111 * doc/rtl.texi (sequence): Rewrite documentation to match the
13112 current use of SEQUENCE rtl objects.
13113 * rtl.def (SEQUENCE): Likewise.
13114
13115 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13116 Update documentation.
13117 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13118 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13119
13120 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13121
13122 2013-04-08 Teresa Johnson <tejohnson@google.com>
13123
13124 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13125 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13126 methods.
13127 (estimate_edge_size_and_time): Add comment to suggest using rounding
13128 methods.
13129 (estimate_node_size_and_time): Ditto.
13130 (remap_edge_change_prob): Use helper rounding divide methods.
13131 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13132 (gimple_mod_pow2_value_transform): Ditto.
13133 (gimple_mod_subtract_transform): Ditto.
13134 (gimple_ic_transform): Ditto.
13135 (gimple_stringops_transform): Ditto.
13136 * stmt.c (conditional_probability): Ditto.
13137 (emit_case_dispatch_table): Ditto.
13138 * lto-cgraph.c (merge_profile_summaries): Ditto.
13139 * tree-optimize.c (execute_fixup_cfg): Ditto.
13140 * cfgcleanup.c (try_forward_edges): Ditto.
13141 * cfgloopmanip.c (scale_loop_profile): Ditto.
13142 (loopify): Ditto.
13143 (duplicate_loop_to_header_edge): Ditto.
13144 (lv_adjust_loop_entry_edge): Ditto.
13145 * tree-vect-loop.c (vect_transform_loop): Ditto.
13146 * profile.c (compute_branch_probabilities): Ditto.
13147 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13148 * lto-streamer-in.c (input_cfg): Ditto.
13149 * gimple-streamer-in.c (input_bb): Ditto.
13150 * ipa-cp.c (update_profiling_info): Ditto.
13151 (update_specialized_profile): Ditto.
13152 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13153 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13154 rounding methods.
13155 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13156 (compute_trg_info): Ditto.
13157 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13158 (purge_dead_edges): Ditto.
13159 * loop-unswitch.c (unswitch_loop): Ditto.
13160 * cgraphclones.c (cgraph_clone_edge): Ditto.
13161 (cgraph_clone_node): Ditto.
13162 * tree-inline.c (copy_bb): Ditto.
13163 (copy_edges_for_bb): Ditto.
13164 (initialize_cfun): Ditto.
13165 (copy_cfg_body): Ditto.
13166 (expand_call_inline): Ditto.
13167
13168 2013-04-08 Kai Tietz <ktietz@redhat.com>
13169
13170 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13171 TARGET_CYGWIN64 by TARGET_64BIT.
13172
13173 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13174
13175 * config/epiphany/epiphany.md (GPR_1): New constant.
13176 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13177 * config/epiphany/epiphany.c (gen_compare_reg):
13178 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13179 is already in place.
13180 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13181 Don't require being called during rtl expansion; If y operlaps r0,
13182 return 0.
13183 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13184 (epiphany_expand_epilogue): Likewise.
13185
13186 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13187 Don't use CC_FPmode for ORDERED / UNORDERED.
13188 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
13189
13190 * config/epiphany/constraints.md (CnL): New constraint.
13191 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
13192 * config/epiphany/predicates.md (add_operand): Allow 1024.
13193
13194 * config/epiphany/epiphany.md (logical_op): New code iterator.
13195 (op_mnc): New code attribute.
13196 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
13197 (mov_f+1, mov_f+2): New peephole2 patterns.
13198
13199 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
13200 (cstoresi4): Also allow re-use of zero result when doing a NE
13201 comparison to a non-zero operand.
13202 Use (clobber (scratch)) for first insn if the gpr output is not needed.
13203
13204 * config/epiphany/epiphany.md (<insn_opname>v2si3):
13205 Use gen_addsi3_i / gen_subsi3_i.
13206
13207 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13208
13209 PR c++/34949
13210 PR c++/50243
13211 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
13212 contain anything but clobbers, at most one __builtin_stack_restore,
13213 optionally debug stmts and final resx, and if it has at least one
13214 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
13215 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
13216 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
13217 which isn't defaut definition, remove them.
13218 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
13219 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
13220 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
13221 with MEM_REF LHS with SSA_NAME address.
13222
13223 2013-04-08 Jeff Law <law@redhat.com>
13224
13225 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
13226
13227 2013-04-08 Richard Biener <rguenther@suse.de>
13228
13229 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
13230 extra newline.
13231 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
13232 determined vector type.
13233 (vect_analyze_data_refs): Likewise.
13234 (vect_get_new_vect_var): Adjust.
13235 (vect_create_destination_var): Preserve SSA name versions.
13236 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
13237 not dump anything here.
13238
13239 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13240
13241 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
13242 Add member lr_slot_known.
13243 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
13244 if necessary.
13245 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13246 Remove code that sets lr_slot_offset according to what a previous
13247 version of epiphany_emit_save_restore used to do.
13248 (epiphany_emit_save_restore): When doing an lr save or restore,
13249 set/verify lr_slot_known and lr_slot_offset.
13250
13251 2013-04-08 Xinyu Qi <xyqi@marvell.com>
13252
13253 PR target/54338
13254 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
13255 in ALL_REGS.
13256
13257 2013-04-08 Richard Biener <rguenther@suse.de>
13258
13259 * alias.c (find_base_term): Fix thinko in previous change.
13260
13261 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13262
13263 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
13264 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
13265 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
13266 if possible to compute val.
13267 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
13268 For QImode integers don't require anything about precision. Use
13269 const_with_all_bytes_same to find out if the constant doesn't have
13270 repeated bytes in it.
13271
13272 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13273
13274 * config/s390/s390.c (s390_expand_insv): Only accept insertions
13275 within mode size.
13276
13277 2013-04-08 Marek Polacek <polacek@redhat.com>
13278
13279 PR rtl-optimization/48182
13280 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
13281 value to 1.
13282
13283 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13284
13285 PR target/55487
13286 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
13287 nuses, make sure we have a label.
13288
13289 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13290
13291 PR target/56843
13292 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
13293 (rs6000_emit_swdiv_low_precision): Remove.
13294 (rs6000_emit_swdiv): Rewrite to handle between one and four
13295 iterations of Newton-Raphson generally; modify required number of
13296 iterations for some cases.
13297 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
13298
13299 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
13300
13301 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
13302 set-but-unused variable.
13303
13304 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
13305 basic blocks of released function bodies garbage-collectable.
13306
13307 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
13308 (struct rtl_opt_pass): Add TODO_df_finish.
13309
13310 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
13311
13312 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13313
13314 * config/arm/constraints.md (q): New constraint.
13315 * config/arm/ldrdstrd.md: New file.
13316 * config/arm/arm.md (ldrdstrd.md) New include.
13317 (arm_movdi): Use "q" instead of "r" constraint
13318 for double-word memory access.
13319 (movdf_soft_insn): Likewise.
13320 * config/arm/vfp.md (movdi_vfp): Likewise.
13321 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
13322 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
13323 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
13324 (mem_ok_for_ldrd_strd): Likewise.
13325 (output_move_double): Update assertion.
13326
13327 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13328
13329 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
13330
13331 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13332
13333 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
13334 define_insn_and_split.
13335 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
13336
13337 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13338
13339 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
13340 define_insn_and_split.
13341 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
13342 (shiftsi3_compare): New pattern.
13343 (rrx): New pattern.
13344 * config/arm/unspecs.md (UNSPEC_RRX): New.
13345
13346 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13347
13348 * config/arm/arm.md (negdi_extendsidi): New pattern.
13349 (negdi_zero_extendsidi): Likewise.
13350
13351 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13352
13353 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
13354 define_insn_and_split.
13355 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
13356 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
13357
13358 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13359
13360 * config/arm/arm.md (arm_subdi3): Convert define_insn into
13361 define_insn_and_split.
13362 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
13363 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
13364
13365 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13366
13367 * config/arm/arm.md (subsi3_carryin): New pattern.
13368 (subsi3_carryin_const): Likewise.
13369 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
13370 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
13371
13372 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13373
13374 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
13375
13376 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13377
13378 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
13379 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
13380
13381 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13382
13383 * config/arm/arm.c (arm_expand_builtin): Change fcode
13384 type to unsigned int.
13385
13386 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13387
13388 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
13389
13390 2013-04-04 Ian Lance Taylor <iant@google.com>
13391
13392 * doc/standards.texi (Standards): The Go frontend supports the Go 1
13393 language standard.
13394
13395 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
13396
13397 PR middle-end/56729
13398 * df-scan.c (df_insn_delete): Disable failing assert.
13399
13400 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13401
13402 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
13403 New function prototype.
13404 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
13405 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
13406 (arm_builtin_vectorized_function): New function.
13407
13408 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13409
13410 * config/arm/arm_neon_builtins.def: New file.
13411 * config/arm/arm.c (neon_builtin_data): Move contents to
13412 arm_neon_builtins.def.
13413 (enum arm_builtins): Include neon builtin definitions.
13414 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
13415 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
13416
13417 2013-04-04 Marek Polacek <polacek@redhat.com>
13418
13419 PR tree-optimization/48186
13420 * predict.c (maybe_hot_frequency_p): Return false if
13421 HOT_BB_FREQUENCY_FRACTION is 0.
13422 (cgraph_maybe_hot_edge_p): Likewise.
13423
13424 2013-04-04 Richard Biener <rguenther@suse.de>
13425
13426 PR tree-optimization/56826
13427 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
13428 more accurately.
13429
13430 2013-04-04 Richard Biener <rguenther@suse.de>
13431
13432 PR tree-optimization/56213
13433 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
13434 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
13435
13436 2013-04-04 Richard Biener <rguenther@suse.de>
13437
13438 PR tree-optimization/56837
13439 * tree-loop-distribution.c (classify_partition): For non-zero
13440 values require that the value has the same precision as its
13441 mode to be useful as memset value.
13442
13443 2013-04-03 Nick Clifton <nickc@redhat.com>
13444
13445 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
13446 (fmssf4): Use fmsf.s on E3V5 architectures.
13447 (fnmasf4): Use fnmaf.s on E3V5 architectures.
13448 (fnmssf4): Use fnmsf.s on E3V5 architectures.
13449
13450 2013-04-03 Jeff Law <law@redhat.com>
13451
13452 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
13453 (lra-eliminations.o): Likewise.
13454
13455 2013-04-03 Teresa Johnson <tejohnson@google.com>
13456
13457 * gcov-io.c (compute_working_sets): Moved most of body of old
13458 compute_working_sets here from profile.c.
13459 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
13460 (gcov_working_set_t): Moved typedef here from basic-block.h
13461 (compute_working_set): Declare.
13462 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
13463 (get_working_sets): Renamed from compute_working_set,
13464 replace most of body with call to new compute_working_sets.
13465 (get_exec_counts): Replace call to compute_working_sets
13466 to get_working_sets.
13467 * profile.h (get_working_sets): Renamed from compute_working_set.
13468 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
13469 to get_working_sets.
13470 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
13471 * gcov-dump.c (dump_working_sets): New function.
13472
13473 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
13474
13475 * hwint.c (sext_hwi, zext_hwi): New functions.
13476 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
13477 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
13478 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
13479 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
13480 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
13481 (sext_hwi, zext_hwi): New functions.
13482
13483 2013-04-03 Jeff Law <law@redhat.com>
13484
13485 PR tree-optimization/56799
13486 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
13487 back test for widening conversion erroneously dropped in prior change.
13488
13489 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13490
13491 PR target/56809
13492 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
13493 instead of next_real_insn.
13494
13495 2013-04-03 Marek Polacek <polacek@redhat.com>
13496
13497 PR sanitizer/55702
13498 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
13499
13500 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13501
13502 PR target/56809
13503 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
13504 next_real_insn.
13505 (thumb1_output_casesi): Likewise.
13506 (thumb2_output_casesi): Likewise.
13507
13508 2013-04-03 Richard Biener <rguenther@suse.de>
13509
13510 PR tree-optimization/56817
13511 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
13512 Split out ...
13513 (tree_unroll_loops_completely_1): ... new function to manually
13514 walk the loop tree, properly defering outer loops of unrolled
13515 loops to later iterations.
13516
13517 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13518
13519 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
13520 (vectorizable_load): Likewise.
13521 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
13522 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
13523
13524 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13525
13526 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
13527 BIT_FIELD_REF.
13528
13529 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13530
13531 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
13532
13533 2013-04-03 Bin Cheng <bin.cheng@arm.com>
13534
13535 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
13536
13537 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13538
13539 PR tree-optimization/56790
13540 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
13541 folding.
13542
13543 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13544
13545 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
13546 Handle VEC_MERGE.
13547 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
13548 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
13549 equal arguments.
13550
13551 2013-04-03 Jakub Jelinek <jakub@redhat.com>
13552
13553 PR c/19449
13554 * tree.h (force_folding_builtin_constant_p): New decl.
13555 * builtins.c (force_folding_builtin_constant_p): New variable.
13556 (fold_builtin_constant_p): Fold immediately also if
13557 force_folding_builtin_constant_p.
13558
13559 2013-04-03 Richard Biener <rguenther@suse.de>
13560
13561 PR tree-optimization/56812
13562 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
13563 DRs of the same interleaving chain are independent.
13564
13565 2013-04-02 Jason Merrill <jason@redhat.com>
13566
13567 * gdbinit.in (pbb): Use debug fn.
13568
13569 2013-04-02 Lawrence Crowl <crowl@google.com>
13570
13571 * sese.h (struct ivtype_map_elt_s): Remove unused.
13572 (extern debug_ivtype_map): Remove unused.
13573 (extern eq_ivtype_map_elts): Remove unused.
13574 * sese.c (debug_ivtype_map): Removed unused.
13575 (debug_ivtype_map_1): Removed unused.
13576 (debug_ivtype_elt): Remove unused.
13577 (eq_ivtype_map_elts): Remove unused.
13578
13579
13580 2013-04-02 Kai Tietz <ktietz@redhat.com>
13581
13582 PR target/52790
13583 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
13584 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
13585 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
13586 function.
13587 (legitimize_pe_coff_symbol): Likewise.
13588 (is_imported_p): New helper-function.
13589 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
13590 for Windows x64 targets.
13591 (ix86_expand_prologue): Optimize for pe-coff targets.
13592 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
13593 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
13594 medium/large code-model.
13595 (legitimize_pic_address): Likewise.
13596 (legitimize_tls_address): Likewise.
13597 (ix86_expand_call): Likewise.
13598 (x86_output_mi_thunk): Likewise.
13599 (get_dllimport_decl): Add new beimport argument.
13600 (construct_plt_address): Don't assert for x64 pe-coff targets.
13601 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
13602 targets.
13603 (SYMBOL_FLAG_STUBVAR): New macro.
13604 (SYMBOL_REF_STUBVAR_P): Likewise.
13605 * config/i386/winnt.c (stub_list): New structure.
13606 (stub_head): New local variable.
13607 (i386_pe_record_stub): New function.
13608 (i386_pe_file_end): Emit refptr-stubs.
13609
13610 2013-04-02 Jakub Jelinek <jakub@redhat.com>
13611
13612 PR rtl-optimization/56745
13613 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
13614 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
13615
13616 PR c++/34949
13617 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
13618 and both of them are MEM_REFs, just compare first argument for
13619 equality and attempt to deal even with differing offsets.
13620
13621 PR c++/34949
13622 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
13623 of gimple_clobber_p to be MEM_REF.
13624 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
13625 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
13626 after gimplification.
13627 * asan.c (get_mem_ref_of_assignment): Don't instrument
13628 gimple_clobber_p stmts.
13629 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
13630 gimple_clobber_p stmt if they have MEM_REF lhs and
13631 are dead because of another gimple_clobber_p stmt.
13632 * tree-ssa-live.c (clear_unused_block_pointer): Treat
13633 gimple_clobber_p stmts like debug stmts.
13634 (remove_unused_locals): Remove clobbers with MEM_REF lhs
13635 that refer to unused VAR_DECLs or uninitialized values.
13636 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
13637 gimple_clobber_p stmts if they refer to removed parameters.
13638 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
13639 formatting.
13640
13641 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
13642
13643 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
13644 using SWI48 mode attribute.
13645
13646 2013-04-02 Wei Mi <wmi@google.com>
13647
13648 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
13649 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
13650 *<rotate_insn><mode>3_mask in i386.md.
13651
13652 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
13653
13654 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
13655
13656 2013-04-02 Richard Biener <rguenther@suse.de>
13657
13658 PR tree-optimization/56778
13659 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
13660 Runtime alias tests are not supported for gather loads.
13661 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
13662 stmts referenced from SSA operands before updating SSA form.
13663
13664 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
13665 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13666
13667 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
13668 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
13669 * config/arm/cortex-a53.md: New file.
13670 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
13671 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
13672 * config/arm/arm.c (arm_issue_rate): Likewise.
13673 * config/arm/arm-tune.md: Regenerate
13674 * config/arm/arm-tables.opt: Regenerate.
13675 * config/arm/arm-cores.def: Add cortex-a53.
13676
13677 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
13678
13679 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
13680 non-static link.
13681
13682 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13683
13684 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
13685 scalar load/store operations using B/H registers.
13686 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
13687
13688 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13689
13690 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
13691 scalar move.
13692 * config/aarch64/aarch64.c
13693 (aarch64_simd_scalar_immediate_valid_for_move): New.
13694 * config/aarch64/aarch64-protos.h
13695 (aarch64_simd_scalar_immediate_valid_for_move): New.
13696 * config/aarch64/constraints.md (Dh, Dq): New.
13697 * config/aarch64/iterators.md (hq): New.
13698
13699 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
13700
13701 * reorg.c (get_branch_condition): Deal with conditional returns.
13702 (fill_simple_delay_slots): Remove dead code dealing with jumps.
13703
13704 2013-04-01 Wei Mi <wmi@google.com>
13705
13706 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
13707 Truncate operand 2 using %b asm operand modifier.
13708 (*<shift_insn><mode>3_mask): Ditto.
13709 (*<rotate_insn><mode>3_mask): Ditto.
13710
13711 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
13712
13713 PR middle-end/56798
13714 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
13715
13716 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
13717
13718 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
13719 of next_real_insn.
13720 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
13721
13722 2013-03-30 Lawrence Crowl <crowl@google.com>
13723
13724 * dse.c (clear_alias_sets): Remove never set.
13725 (disqualified_clear_alias_sets): Remove never set.
13726 (clear_alias_mode_pool): Remove never set.
13727 (dse_step0): Remove condition that is never true.
13728 (canon_address): Remove condition that is never true.
13729 (dse_step7): Remove condition that is never true.
13730 (rest_of_handle_dse): Remove condition that is never true.
13731 (rest_of_handle_dse::did_global): Remove never read from above.
13732 (dse_step2_spill): Remove never called from above.
13733 (dse_step5_spill): Remove never called from above.
13734
13735 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
13736
13737 * doc/md.texi (Standard Names) <casesi>: Update documentation for
13738 JUMP_TABLE_DATA changes.
13739 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
13740 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
13741 (Insns) <jump_table_data>: New entry.
13742 * doc/tm.texi: Regenerate.
13743
13744 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
13745
13746 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
13747 for table jump at the end of a basic block using tablejump_p.
13748 * targhooks.c (default_invalid_within_doloop): Likewise.
13749 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
13750 target hook implementation that is identical to the default hook.
13751 (rs6000_invalid_within_doloop): Remove.
13752
13753 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
13754 unused variable from tablejump_p call.
13755
13756 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
13757 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
13758 (INSN_DELETED_P): Likewise.
13759 (emit_jump_table_data): New prototype.
13760 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
13761 after 4th as unused.
13762 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
13763 * sched-vis.c (print_insn): Likewise.
13764 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
13765 insn for compatibility with back ends that use next_active_insn to
13766 identify jump table data.
13767 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
13768 (remove_insn): Likewise.
13769 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
13770 to be emitted.
13771 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
13772 (emit_jump_table_data): New function.
13773
13774 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
13775 basic block, a JUMP_TABLE_DATA never is.
13776 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
13777 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
13778 off from code handling real insns.
13779 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
13780 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
13781 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
13782 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
13783 is not a NONDEBUG_INSN_P.
13784 * ira-costs.c (scan_one_insn): Likewise.
13785 * jump.c (mark_all_labels): Likewise.
13786 (mark_jump_label_1): Likewise.
13787 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13788 * lra.c (get_insn_freq): Expect all insns reaching here to be in
13789 a basic block.
13790 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
13791 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
13792 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
13793 JUMP_TABLE_DATA_P insns.
13794 (calculate_elim_costs_all_insns): Likewise.
13795 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
13796 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
13797 (delete_output_reload): Code style fixups.
13798 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
13799 insn flags on this non-insn.
13800 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13801 as scheduling barriers, for pre-change compatibility.
13802 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13803 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13804
13805 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13806 redundant JUMP_TABLE_DATA_P test.
13807 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13808 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13809 (frv_for_each_packet): Likewise.
13810 * config/i386/i386.c (min_insn_size): Likewise.
13811 (ix86_avoid_jump_mispredicts): Likewise.
13812 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13813 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13814 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13815 (mips16_insn_length): Robustify.
13816 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13817 (mips16_split_long_branches): Likewise.
13818 * config/pa/pa.c (pa_combine_instructions): Likewise.
13819 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13820 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13821 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13822 as contributing to pool range lengths.
13823 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13824 Remove redundant JUMP_TABLE_DATA_P test.
13825 (sh_loop_align): Likewise.
13826 (split_branches): Likewise.
13827 (sh_insn_length_adjustment): Likewise.
13828 * config/spu/spu.c (get_branch_target): Likewise.
13829
13830 2013-03-29 Jan Hubicka <jh@suse.cz>
13831
13832 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13833 gcov streaming; stream hot bb threshold to ltrans.
13834 * predict.c (get_hot_bb_threshold): Break out from ....
13835 (maybe_hot_count_p): ... here.
13836 (set_hot_bb_threshold): New function.
13837 * lto-section-in.c (lto_section_name): Add profile.
13838 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13839 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13840 and data-streamer.h
13841 (histogram_entry): New structure.
13842 (histogram, histogram_pool): New global vars.
13843 (histogram_hash): New structure.
13844 (histogram_hash::hash): New method.
13845 (histogram_hash::equal): Likewise.
13846 (account_time_size): New function.
13847 (cmp_counts): New function.
13848 (dump_histogram): New function.
13849 (ipa_profile_generate_summary): New function.
13850 (ipa_profile_write_summary): New function.
13851 (ipa_profile_read_summary): New function.
13852 (ipa_profile): Decide on threshold.
13853 (pass_ipa_profile): Add ipa_profile_write_summary and
13854 ipa_profile_read_summary.
13855 * Makefile.in (ipa.o): Update dependencies.
13856 * lto-streamer.h (LTO_section_ipa_profile): New section.
13857
13858 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13859
13860 * tree.h (VAR_P): New.
13861
13862 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13863
13864 PR lto/56777
13865 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13866
13867 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13868
13869 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13870 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13871 (control_flow_insn_p): Likewise.
13872 * cfgrtl.c (duplicate_insn_chain): Likewise.
13873 * final.c (get_attr_length_1): Likewise.
13874 (shorten_branches): Likewise.
13875 (final_scan_insn): Likewise.
13876 * function.c (instantiate_virtual_regs): Likewise.
13877 * gcse.c (insert_insn_end_basic_block): Likewise.
13878 * ira-costs.c (scan_one_insn): Likewise.
13879 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13880 * lra.c (check_rtl): Likewise.
13881 * reload1.c (elimination_costs_in_insn): Likewise.
13882 * reorg.c (follow_jumps): Likewise.
13883
13884 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13885 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13886 (thumb_far_jump_used_p): Likewise.
13887 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13888 (workaround_speculation): Likewise.
13889 (add_sched_insns_for_speculation): Likewise.
13890 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13891 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13892 (frv_for_each_packet): Likewise.
13893 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13894 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13895 (final_emit_insn_group_barriers): Likewise.
13896 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13897 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13898 (mips16_insn_length): Likewise.
13899 * config/pa/pa.c (pa_reorg): Likewise.
13900 (pa_combine_instructions): Likewise.
13901 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13902 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13903 (sh_reorg): Likewise.
13904 (split_branches): Likewise.
13905 * config/spu/spu.c (get_branch_target): Likewise.
13906
13907 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13908 JUMP_TABLE_DATA_P.
13909
13910 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13911
13912 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13913 Fix declaration name.
13914
13915 2013-03-28 Lawrence Crowl <crowl@google.com>
13916
13917 * graphds.h (struct graph.indicies): Remove unused.
13918 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13919 (SCOP_ORIGINAL_PDDRS): Remove unused.
13920 * sese.h (extern insert_loop_close_phis): Removed unused.
13921 (extern insert_guard_phis): Removed unused.
13922 (extern ivtype_map_elt_info): Removed unused.
13923 (new_ivtype_map_elt): Removed unused.
13924 * sese.c (ivtype_map_elt_info): Removed unused.
13925
13926 2013-03-28 Lawrence Crowl <crowl@google.com>
13927
13928 * Makefile.in: Add several missing include dependences.
13929 (DUMPFILE_H): New.
13930 (test-dump.o): New. This object is not added to any executable,
13931 but is present for ad-hoc testing.
13932 * bitmap.c
13933 (debug (const bitmap_head_def &)): New.
13934 (debug (const bitmap_head_def *)): New.
13935 * bitmap.h
13936 (extern debug (const bitmap_head_def &)): New.
13937 (extern debug (const bitmap_head_def *)): New.
13938 * cfg.c
13939 (debug (edge_def &)): New.
13940 (debug (edge_def *)): New.
13941 * cfghooks.c
13942 (debug (basic_block_def &)): New.
13943 (debug (basic_block_def *)): New.
13944 * dumpfile.h
13945 (dump_node (const_tree, int, FILE *)): Correct source file.
13946 * dwarf2out.c
13947 (debug (die_struct &)): New.
13948 (debug (die_struct *)): New.
13949 * dwarf2out.h
13950 (extern debug (die_struct &)): New.
13951 (extern debug (die_struct *)): New.
13952 * gimple-pretty-print.c
13953 (debug (gimple_statement_d &)): New.
13954 (debug (gimple_statement_d *)): New.
13955 * gimple-pretty-print.h
13956 (extern debug (gimple_statement_d &)): New.
13957 (extern debug (gimple_statement_d *)): New.
13958 * ira-build.c
13959 (debug (ira_allocno_copy &)): New.
13960 (debug (ira_allocno_copy *)): New.
13961 (debug (ira_allocno &)): New.
13962 (debug (ira_allocno *)): New.
13963 * ira-int.h
13964 (extern debug (ira_allocno_copy &)): New.
13965 (extern debug (ira_allocno_copy *)): New.
13966 (extern debug (ira_allocno &)): New.
13967 (extern debug (ira_allocno *)): New.
13968 * ira-lives.c
13969 (debug (live_range &)): New.
13970 (debug (live_range *)): New.
13971 * lra-int.h
13972 (debug (lra_live_range &)): New.
13973 (debug (lra_live_range *)): New.
13974 * lra-lives.c
13975 (debug (lra_live_range &)): New.
13976 (debug (lra_live_range *)): New.
13977 * omega.c
13978 (debug (omega_pb_d &)): New.
13979 (debug (omega_pb_d *)): New.
13980 * omega.h
13981 (extern debug (omega_pb_d &)): New.
13982 (extern debug (omega_pb_d *)): New.
13983 * print-rtl.c
13984 (debug (const rtx_def &)): New.
13985 (debug (const rtx_def *)): New.
13986 * print-tree.c
13987 (debug_tree (tree): Move within file.
13988 (debug_raw (const tree_node &)): New.
13989 (debug_raw (const tree_node *)): New.
13990 (dump_tree_via_hooks (const tree_node *, int)): New.
13991 (debug (const tree_node &)): New.
13992 (debug (const tree_node *)): New.
13993 (debug_verbose (const tree_node &)): New.
13994 (debug_verbose (const tree_node *)): New.
13995 (debug_head (const tree_node &)): New.
13996 (debug_head (const tree_node *)): New.
13997 (debug_body (const tree_node &)): New.
13998 (debug_body (const tree_node *)): New.
13999 (debug_vec_tree (tree): Move and reimplement in terms of dump.
14000 (debug (vec<tree, va_gc> &)): New.
14001 (debug (vec<tree, va_gc> *)): New.
14002 * rtl.h
14003 (extern debug (const rtx_def &)): New.
14004 (extern debug (const rtx_def *)): New.
14005 * sbitmap.c
14006 (debug_raw (simple_bitmap_def &)): New.
14007 (debug_raw (simple_bitmap_def *)): New.
14008 (debug (simple_bitmap_def &)): New.
14009 (debug (simple_bitmap_def *)): New.
14010 * sbitmap.h
14011 (extern debug (simple_bitmap_def &)): New.
14012 (extern debug (simple_bitmap_def *)): New.
14013 (extern debug_raw (simple_bitmap_def &)): New.
14014 (extern debug_raw (simple_bitmap_def *)): New.
14015 * sel-sched-dump.c
14016 (debug (vinsn_def &)): New.
14017 (debug (vinsn_def *)): New.
14018 (debug_verbose (vinsn_def &)): New.
14019 (debug_verbose (vinsn_def *)): New.
14020 (debug (expr_def &)): New.
14021 (debug (expr_def *)): New.
14022 (debug_verbose (expr_def &)): New.
14023 (debug_verbose (expr_def *)): New.
14024 (debug (vec<rtx> &)): New.
14025 (debug (vec<rtx> *)): New.
14026 * sel-sched-dump.h
14027 (extern debug (vinsn_def &)): New.
14028 (extern debug (vinsn_def *)): New.
14029 (extern debug_verbose (vinsn_def &)): New.
14030 (extern debug_verbose (vinsn_def *)): New.
14031 (extern debug (expr_def &)): New.
14032 (extern debug (expr_def *)): New.
14033 (extern debug_verbose (expr_def &)): New.
14034 (extern debug_verbose (expr_def *)): New.
14035 (extern debug (vec<rtx> &)): New.
14036 (extern debug (vec<rtx> *)): New.
14037 * sel-sched-ir.h
14038 (_list_iter_cond_expr): Make inline instead of static.
14039 * sreal.c
14040 (debug (sreal &)): New.
14041 (debug (sreal *)): New.
14042 * sreal.h
14043 (extern debug (sreal &)): New.
14044 (extern debug (sreal *)): New.
14045 * tree.h
14046 (extern debug_raw (const tree_node &)): New.
14047 (extern debug_raw (const tree_node *)): New.
14048 (extern debug (const tree_node &)): New.
14049 (extern debug (const tree_node *)): New.
14050 (extern debug_verbose (const tree_node &)): New.
14051 (extern debug_verbose (const tree_node *)): New.
14052 (extern debug_head (const tree_node &)): New.
14053 (extern debug_head (const tree_node *)): New.
14054 (extern debug_body (const tree_node &)): New.
14055 (extern debug_body (const tree_node *)): New.
14056 (extern debug (vec<tree, va_gc> &)): New.
14057 (extern debug (vec<tree, va_gc> *)): New.
14058 * tree-cfg.c
14059 (debug (struct loop &)): New.
14060 (debug (struct loop *)): New.
14061 (debug_verbose (struct loop &)): New.
14062 (debug_verbose (struct loop *)): New.
14063 * tree-dump.c: Add header dependence.
14064 * tree-flow.h
14065 (extern debug (struct loop &)): New.
14066 (extern debug (struct loop *)): New.
14067 (extern debug_verbose (struct loop &)): New.
14068 (extern debug_verbose (struct loop *)): New.
14069 * tree-data-ref.c
14070 (debug (data_reference &)): New.
14071 (debug (data_reference *)): New.
14072 (debug (vec<data_reference_p> &)): New.
14073 (debug (vec<data_reference_p> *)): New.
14074 (debug (vec<ddr_p> &)): New.
14075 (debug (vec<ddr_p> *)): New.
14076 * tree-data-ref.h
14077 (extern debug (data_reference &)): New.
14078 (extern debug (data_reference *)): New.
14079 (extern debug (vec<data_reference_p> &)): New.
14080 (extern debug (vec<data_reference_p> *)): New.
14081 (extern debug (vec<ddr_p> &)): New.
14082 (extern debug (vec<ddr_p> *)): New.
14083 * tree-ssa-alias.c
14084 (debug (pt_solution &)): New.
14085 (debug (pt_solution *)): New.
14086 * tree-ssa-alias.h
14087 (extern debug (pt_solution &)): New.
14088 (extern debug (pt_solution *)): New.
14089 * tree-ssa-alias.c
14090 (debug (_var_map &)): New.
14091 (debug (_var_map *)): New.
14092 (debug (tree_live_info_d &)): New.
14093 (debug (tree_live_info_d *)): New.
14094 * tree-ssa-alias.h
14095 (extern debug (_var_map &)): New.
14096 (extern debug (_var_map *)): New.
14097 (extern debug (tree_live_info_d &)): New.
14098 (extern debug (tree_live_info_d *)): New.
14099
14100 2013-03-28 Jan Hubicka <jh@suse.cz>
14101
14102 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14103
14104 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14105
14106 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14107 record only when desired or required.
14108
14109 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14110
14111 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14112 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14113
14114 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14115
14116 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14117 (*andsi3_compare0_uxtw): New pattern.
14118 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14119 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14120
14121 2013-03-28 Jan Hubicka <jh@suse.cz>
14122
14123 * data-streamer-in.c (streamer_read_gcov_count): New function.
14124 * gimple-streamer-out.c: Include value-prof.h.
14125 (output_gimple_stmt): Output histogram.
14126 (output_bb): Use streamer_write_gcov_count.
14127 * value-prof.c: Include data-streamer.h
14128 (dump_histogram_value): Add HIST_TYPE_MAX.
14129 (stream_out_histogram_value): New function.
14130 (stream_in_histogram_value): New function.
14131 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14132 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14133 * data-streamer-out.c (streamer_write_gcov_count): New function.
14134 (streamer_write_gcov_count_stream): New function.
14135 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14136 (lto_output_node): Likewise.
14137 (input_node, input_edge): Likewise.
14138 * lto-streamer-out.c (output_cfg): Update streaming.
14139 * lto-streamer-in.c (input_cfg): Likewise.
14140 * data-streamer.h (streamer_write_gcov_count,
14141 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14142 * gimple-streamer-in.c: Include value-prof.h
14143 (input_gimple_stmt): Input histograms.
14144 (input_bb): Update profile streaming.
14145
14146 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14147
14148 * genmodes.c (emit_max_int): New function.
14149 (emit_insn_modes_h): Added call to emit_max_function.
14150 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14151 Added doc.
14152 * machmode.def: Fixed comment.
14153
14154 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14155
14156 * combine.c (try_combine): Removed useless assert.
14157 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14158
14159 2013-03-28 Marek Polacek <polacek@redhat.com>
14160 Richard Biener <rguenther@suse.de>
14161
14162 PR tree-optimization/56695
14163 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14164 build signed result of a vector comparison.
14165 * tree-cfg.c (verify_gimple_comparison): Check that a result
14166 of a vector comparison has signed type.
14167
14168 2013-03-28 Richard Biener <rguenther@suse.de>
14169
14170 PR tree-optimization/37021
14171 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14172 do not restrict gaps between groups.
14173 * tree-vect-stmts.c (vectorizable_load): Properly account for
14174 a gap between groups.
14175
14176 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14177
14178 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14179 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14180 is not enabled.
14181
14182 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14183
14184 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14185 * doc/extend.texi (Named Address Spaces): Ditto.
14186 (Variable Attributes): Ditto.
14187
14188 2013-03-27 Kai Tietz <ktietz@redhat.com>
14189
14190 * config.build: Add support for cygwin x64 target.
14191 * config.gcc: Likewise.
14192 * config.host: Likewise.
14193 * configure.ac: Likewise
14194 * configure: Regenerated.
14195
14196 2013-03-27 Kai Tietz <ktietz@redhat.com>
14197
14198 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
14199 * config/i386/t-cygwin-w64: New file.
14200 * config/i386/cygwin-w64.h: New file.
14201 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
14202 and add support for x64-cygwin target.
14203 (CPP_SPEC): Likewise.
14204 (CXX_WRAP_SPEC_LIST): Undefine before define.
14205 (LIBGCJ_SONAME): Use 15 as version.
14206
14207 2013-03-27 Richard Biener <rguenther@suse.de>
14208
14209 PR tree-optimization/56716
14210 * tree-ssa-structalias.c (perform_var_substitution): Adjust
14211 dumping for ref nodes.
14212
14213 2013-03-27 Martin Jambor <mjambor@suse.cz>
14214
14215 PR tree-optimization/55334
14216 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
14217 restricted pointers to arrays.
14218
14219 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
14220
14221 * Makefile.in (.SUFFIXES): Add .cc.
14222 (.c.o): Apply same recipe for implicit rule .cc.o.
14223
14224 2013-03-27 Richard Biener <rguenther@suse.de>
14225
14226 PR tree-optimization/37021
14227 * tree-vect-data-refs.c (vect_check_strided_load): Allow
14228 REALPART/IMAGPART_EXPRs around the supported refs.
14229 * tree-ssa-structalias.c (find_func_aliases): Assume that
14230 floating-point values are not used to transfer pointers.
14231
14232 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
14233
14234 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
14235 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
14236 * gcc/doc/tm.texi: Regenerate.
14237 * gcc/targhooks.h (default_has_ifunc_p): New.
14238 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
14239 * gcc/config/linux-protos.h: New file.
14240 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
14241 this hook for linux which disables support of indirect functions in
14242 android.
14243 * gcc/config/linux-android.c: New file.
14244 * gcc/config/t-linux-android.c: Ditto.
14245 * gcc/config.gcc: Added new object file linux-android.o.
14246 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
14247 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
14248 * gcc/varasm.c (do_assemble_alias): Likewise.
14249 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
14250 doesn't support indirect functions.
14251 * configure: Regenerate.
14252
14253 2013-03-27 Bin Cheng <bin.cheng@arm.com>
14254
14255 PR target/56102
14256 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
14257 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
14258 mult-word mode.
14259
14260 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14261
14262 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
14263
14264 2013-03-27 Terry Guo <terry.guo@arm.com>
14265
14266 * config/arm/arm-cores.def: Added core cortex-r7.
14267 * config/arm/arm-tune.md: Regenerated.
14268 * config/arm/arm-tables.opt: Regenerated.
14269 * doc/invoke.texi: Added entry for core cortex-r7.
14270
14271 2013-03-27 Walter Lee <walt@tilera.com>
14272
14273 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
14274 double-decrement of next_scratch_regno.
14275
14276 2013-03-27 Walter Lee <walt@tilera.com>
14277
14278 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
14279 input operands.
14280 (insn_v1mulus): Ditto.
14281 (insn_v2muls): Ditto.
14282
14283 2013-03-27 Walter Lee <walt@tilera.com>
14284
14285 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
14286 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
14287
14288 2013-03-27 Walter Lee <walt@tilera.com>
14289
14290 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
14291 (*sibcall_value): Ditto.
14292
14293 2013-03-27 Walter Lee <walt@tilera.com>
14294
14295 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
14296 (insn_mnz_v8qi): ... this ...
14297 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
14298 vector equivalent.
14299 (insn_v<n>mnz): Replaced by ...
14300 (insn_v1mnz): ... this ...
14301 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
14302 equivalent.
14303 (insn_mz_<mode>): Replaced by ...
14304 (insn_mz_v8qi): ... this ...
14305 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
14306 vector equivalent.
14307 (insn_v<n>mz): Replaced by ...
14308 (insn_v1mz): ... this ...
14309 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
14310 equivalent.
14311
14312 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
14313
14314 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
14315
14316 2013-03-26 Roland McGrath <mcgrathr@google.com>
14317
14318 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
14319 than fprintf with a non-constant, non-format string.
14320
14321 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14322
14323 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
14324 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
14325 operand 0 predicate.
14326 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
14327 attribute. Use general_x64nomem_operand as operand 1 predicate.
14328 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
14329 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14330 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
14331 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14332 (mov<mode>_insv_1): Remove expander. Merge insn with
14333 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
14334 Use general_x64nomem_operand as operand 1 predicate.
14335 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
14336 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
14337 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14338 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
14339 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
14340 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
14341 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14342 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
14343 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
14344 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
14345 (general_x64nomem_operand): Ditto.
14346
14347 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
14348
14349 * config/rtems.opt: Add -pthread option.
14350
14351 2013-03-26 Richard Biener <rguenther@suse.de>
14352
14353 * alias.c (find_base_term): Avoid redundant and not used recursion.
14354 (base_alias_check): Get the initial base term from the caller.
14355 (true_dependence_1): Compute and pass base terms to base_alias_check.
14356 (write_dependence_p): Likewise.
14357 (may_alias_p): Likewise.
14358
14359 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
14360
14361 * config/aarch64/aarch64.c (aarch64_classify_address): Support
14362 PC-relative load in SI modes and above only.
14363
14364 2013-03-26 Xinyu Qi <xyqi@marvell.com>
14365
14366 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
14367 * config/arm/iwmmxt.md (WCGR0): Update.
14368 (WCGR1, WCGR2, WCGR3): Likewise.
14369
14370 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14371
14372 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
14373 Use x64 and nox64 isa attributes.
14374
14375 2013-03-26 Richard Biener <rguenther@suse.de>
14376
14377 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
14378 alignment computations and rely on get_object_alignment_1
14379 for the !TYPE_P case.
14380 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
14381
14382 2013-03-26 Walter Lee <walt@tilera.com>
14383
14384 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
14385 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
14386
14387 2013-03-25 Jeff Law <law@redhat.com>
14388
14389 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
14390 check for INTEGRAL_TYPE_P that was missing due to checking in
14391 wrong version of prior patch.
14392
14393 2013-03-25 Walter Lee <walt@tilera.com>
14394
14395 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
14396 TILEGX_INSN_SHUFFLEBYTES1.
14397 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
14398 shufflebytes1.
14399 (tilegx_builtins): Ditto.
14400 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
14401
14402 2013-03-25 Walter Lee <walt@tilera.com>
14403
14404 * config/tilegx/tilegx.md (floatsisf2): New pattern.
14405 (floatunssisf2): New pattern.
14406 (floatsidf2): New pattern.
14407 (floatunssidf2): New pattern.
14408
14409 2013-03-25 Walter Lee <walt@tilera.com>
14410
14411 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
14412 tests for constraint J, K, N, P.
14413
14414 2013-03-25 Walter Lee <walt@tilera.com>
14415
14416 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
14417 Use indirect/pcrel encoding.
14418 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
14419 Ditto.
14420
14421 2013-03-25 Steve Ellcey <sellcey@mips.com>
14422
14423 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
14424 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
14425 * config/mips/mips.c (mips_option_override): Set IMADD default.
14426 * config/mips/mips.h (PTF_AVOID_IMADD): New.
14427 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
14428 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
14429 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
14430 * doc/invoke.texi (-mimadd/-mno-imadd): New.
14431
14432 2013-03-25 Jeff Law <law@redhat.com>
14433
14434 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
14435 slightly to avoid creating and folding useless trees. Simplify
14436 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
14437
14438 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
14439
14440 * config/i386/i386.md (*zero_extendsidi2): Merge with
14441 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
14442 * config/i386/predicates.md (x86_64_zext_operand): Rename from
14443 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
14444 targets. Clarify comment.
14445
14446 2013-03-25 Martin Jambor <mjambor@suse.cz>
14447
14448 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
14449 pass-through jump functions differently.
14450 (ipa_read_jump_function): Likewise. Also use setter functions to set
14451 up jump functions.
14452
14453 2013-03-25 Martin Jambor <mjambor@suse.cz>
14454
14455 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
14456 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
14457 process it.
14458 (ipa_get_indirect_edge_target): New function.
14459 (devirtualization_time_bonus): New parameter known_aggs, pass it to
14460 ipa_get_indirect_edge_target. Update all callers.
14461 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
14462 ipa_get_indirect_edge_target_1 instead of calling
14463 ipa_get_indirect_edge_target.
14464 (create_specialized_node): Pass aggvlas to
14465 ipcp_discover_new_direct_edges.
14466
14467 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14468
14469 * config/arm/arm.md (f_sels, f_seld): New types.
14470 (*cmov<mode>): New pattern.
14471 * config/arm/predicates.md (arm_vsel_comparison_operator): New
14472 predicate.
14473
14474 2013-03-25 Kai Tietz <ktietz@redhat.com>
14475
14476 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
14477 POSIX-printf for mingw-hosted builds.
14478
14479 2013-03-25 Richard Biener <rguenther@suse.de>
14480
14481 PR middle-end/56694
14482 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
14483 must-not-throw stmt location.
14484
14485 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14486
14487 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
14488 Emit load-acquire versions when acq is true.
14489 (arm_emit_store_exclusive): Add rel parameter.
14490 Emit store-release versions when rel is true.
14491 (arm_split_compare_and_swap): Use acquire-release instructions
14492 instead.
14493 of barriers when appropriate.
14494 (arm_split_atomic_op): Likewise.
14495 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
14496 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
14497 (VUNSPEC_SLX): Likewise.
14498 (VUNSPEC_LDA): Likewise.
14499 (VUNSPEC_STL): Likewise.
14500 * config/arm/sync.md (atomic_load<mode>): New pattern.
14501 (atomic_store<mode>): Likewise.
14502 (arm_load_acquire_exclusive<mode>): Likewise.
14503 (arm_load_acquire_exclusivesi): Likewise.
14504 (arm_load_acquire_exclusivedi): Likewise.
14505 (arm_store_release_exclusive<mode>): Likewise.
14506
14507 2013-03-25 Catherine Moore <clm@codesourcery.com>
14508
14509 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
14510 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
14511 * config/mip/predicates.md (lwsp_swsp_operand,
14512 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
14513 sb16_operand, db4_operand, db7_operand, ib3_operand,
14514 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
14515 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
14516 andi16_operand): New predicates.
14517 * config/mips/mips.md (compression): New attribute.
14518 (enabled): New attribute.
14519 (length): Consider compression in computing length.
14520 (shift_compression): New code attribute.
14521 (*add<mode>3): New operands. Record compression.
14522 (sub<mode>3): Likewise.
14523 (one_cmpl<mode>2): Likewise.
14524 (*and<mode>3): Likewise.
14525 (*ior<mode>3): Likewise.
14526 (unnamed pattern for xor): Likewise.
14527 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
14528 (*<optab><mode>3): Likewise.
14529 (*mov<mode>_internal: Likewise.
14530 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
14531 (mips_unsigned_immediate_p): New.
14532 (umips_lwsp_swsp_address_p): New.
14533 (m16_based_address_p): New.
14534 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
14535 (mips_unsigned_immediate_p): New prototype.
14536 (lwsp_swsp_address_p): New prototype.
14537 (m16_based_address_p): New prototype.
14538 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
14539 (mips_signed_immediate_p): New function.
14540 (m16_based_address_p): New function.
14541 (lwsp_swsp_address_p): New function.
14542 (mips_print_operand_punctuation): Recognize short delay slot insns
14543 for microMIPS.add<mode>3"
14544
14545 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14546
14547 PR target/56720
14548 * config/arm/iterators.md (v_cmp_result): New mode attribute.
14549 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
14550
14551 2013-03-25 Richard Biener <rguenther@suse.de>
14552
14553 PR tree-optimization/56689
14554 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
14555 any edge.
14556
14557 2013-03-25 Richard Biener <rguenther@suse.de>
14558
14559 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
14560 of bitmap.
14561 (memory_references): Likewise.
14562 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
14563 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
14564 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
14565 (gather_mem_refs_in_loops): Fold into ...
14566 (analyze_memory_references): ... this. Move initialization
14567 to tree_ssa_lim_initialize.
14568 (fill_always_executed_in): Rename to ...
14569 (fill_always_executed_in_1): ... this.
14570 (fill_always_executed_in): Move contains_call computation to
14571 this new function from ...
14572 (tree_ssa_lim_initialize): ... here.
14573 (tree_ssa_lim): Call fill_always_executed_in.
14574
14575 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14576
14577 * postreload.c (reload_combine): Fix code detecting returns.
14578
14579 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14580
14581 * function.c (emit_use_return_register_into_block): On cc0 targets,
14582 do not emit the sequence between cc0 setter and user.
14583
14584 2013-03-25 Kai Tietz <ktietz@redhat.com>
14585
14586 * config/i386/predicates.md (local_symbolic_operand): Interpret
14587 dll-imported symbols as none-local.
14588
14589 2013-03-25 Richard Biener <rguenther@suse.de>
14590
14591 * tree-ssa-loop-im.c (struct depend): Remove.
14592 (struct lim_aux_data): Make depends a vec of gimples.
14593 (free_lim_aux_data): Adjust.
14594 (add_dependency): Likewise.
14595 (set_level): Likewise.
14596
14597 2013-03-25 Richard Biener <rguenther@suse.de>
14598
14599 PR middle-end/56434
14600 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
14601 the pointer returned by calls with ECF_MALLOC set.
14602
14603 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14604
14605 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
14606
14607 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14608
14609 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
14610 using MMXMODE mode iterator.
14611 (*move<mode>_internal): Merge with *movv2sf_internal and
14612 *movv2sf_internal_rex64 using MMXMODE mode iterator.
14613
14614 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14615
14616 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
14617 (record_last_mem_set_info): Likewise.
14618
14619 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
14620 of XNEWVEC followed by memset.
14621 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
14622
14623 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14624
14625 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
14626 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
14627 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
14628 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
14629 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
14630 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
14631 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
14632 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
14633 BARRIER_P instead of GET_CODE.
14634
14635 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14636
14637 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
14638 inaccuracy in the probing code.
14639
14640 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
14641 (ctrapdi4): Likewise.
14642
14643 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14644
14645 * calls.c (expand_call): Add missing guard to code handling return
14646 of non-BLKmode structures in MSB.
14647 * function.c (expand_function_end): Likewise.
14648
14649 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14650
14651 * combine.c (try_combine): Adjust comment. Do not add the set of
14652 insn #0 if the destination indirectly is set or dies in insn #2.
14653 Tidy up code to distribute a new note.
14654
14655 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14656
14657 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
14658 also for alternatives 16 and 17.
14659
14660 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14661
14662 * config/i386/sse.md (*mov<mode>_internal): Merge with
14663 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
14664 Emit insn template depending on type attribute. Use
14665 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14666 movd instead of movq mnemonic for interunit moves. Rewrite mode
14667 attribute calculation. Remove unit attribute calculation.
14668 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
14669 Set prefix_data16 attribute for DImode ssemov types.
14670 Use Ym instead of y for SSE-MMX conversion alternatives.
14671 Reorder operand constraints.
14672
14673 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
14674
14675 * df.h (df_insn_delete): Adjust prototype.
14676 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
14677 and let it decide whether mark the basic block dirty.
14678 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
14679 * df-scan.c (df_insn_info_delete): New helper function, split
14680 off from df_insn_delete.
14681 (df_scan_free_bb_info): Use it.
14682 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
14683 Likewise.
14684 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
14685 that the insn is actually an insn and it has a non-NULL basic block.
14686 Do not mark basic block dirty if only deleting a DEBUG_INSN.
14687
14688 2013-03-22 Richard Biener <rguenther@suse.de>
14689
14690 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
14691 dep_ref members.
14692 (mem_ref_alloc): Do not allocate them.
14693 (refs_independent_p): Do not query or maintain a cache.
14694
14695 2013-03-22 Richard Biener <rguenther@suse.de>
14696
14697 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
14698 (gather_mem_refs_in_loops): Do not compute it.
14699 (analyze_memory_references): Do not allocate it.
14700 (tree_ssa_lim_finalize): Do not free it.
14701 (for_all_locs_in_loop): Do not query all_refs_in_loop.
14702
14703 2013-03-22 Richard Biener <rguenther@suse.de>
14704
14705 * is-a.h (as_a): Use gcc_checking_assert.
14706
14707 2013-03-22 Ian Bolton <ian.bolton@arm.com>
14708
14709 * config/aarch64/aarch64.c (aarch64_print_operand): New
14710 format specifier for printing a constant in hex.
14711 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
14712 format specifier for printing second operand.
14713
14714 2013-03-22 Richard Biener <rguenther@suse.de>
14715
14716 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
14717 bitmaps.
14718 (gather_mem_refs_in_loops): Perform store accumulation here.
14719 (create_vop_ref_mapping_loop): Remove.
14720 (create_vop_ref_mapping): Likewise.
14721 (analyze_memory_references): Initialize refs_stored_in_loop.
14722 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
14723 (record_indep_loop): Remove.
14724 (record_dep_loop): New function.
14725 (ref_indep_loop_p_1): Adjust to only walk over references
14726 in the loop, not its subloops.
14727 (ref_indep_loop_p): Rename to ...
14728 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
14729 maintaining a more fine-grained cache.
14730 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
14731 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
14732
14733 2013-03-22 Richard Biener <rguenther@suse.de>
14734
14735 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
14736 (struct mem_ref): Make accesses_in_loop a vec of a vec of
14737 aggregate mem_ref_loc.
14738 (free_mem_ref_locs): Inline into ...
14739 (memref_free): ... this and adjust.
14740 (mem_ref_alloc): Adjust.
14741 (mem_ref_locs_alloc): Remove.
14742 (record_mem_ref_loc): Adjust.
14743 (get_all_locs_in_loop): Rewrite into ...
14744 (for_all_locs_in_loop): ... this iterator.
14745 (rewrite_mem_ref_loc): New functor.
14746 (rewrite_mem_refs): Use for_all_locs_in_loop.
14747 (sm_set_flag_if_changed): New functor.
14748 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
14749 (ref_always_accessed): New functor.
14750 (ref_always_accessed_p): Use for_all_locs_in_loop.
14751
14752 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14753
14754 * tree-pass.h (PROP_gimple_lvec): New.
14755 * passes.c (dump_properties): Handle PROP_gimple_lvec.
14756 (init_optimization_passes): Move pass_lower_vector.
14757 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
14758 PROP_gimple_lvec.
14759 (pass_lower_vector): Provide PROP_gimple_lvec.
14760 (pass_lower_vector_ssa): Likewise.
14761 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
14762
14763 2013-03-21 Mark Wielaard <mjw@redhat.com>
14764
14765 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
14766
14767 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14768
14769 * config/i386/i386.md (*movdi_internal): Disparage slightly
14770 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
14771 conversion alternatives.
14772
14773 2013-03-21 Jakub Jelinek <jakub@redhat.com>
14774
14775 PR middle-end/48087
14776 * diagnostic.def (DK_WERROR): New kind.
14777 * diagnostic.h (werrorcount): Define.
14778 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
14779 promoted to DK_ERROR, increment DK_WERROR counter instead of
14780 DK_ERROR counter.
14781 * toplev.c (toplev_main): Call print_ignored_options even if
14782 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
14783 even if just werrorcount is non-zero.
14784
14785 PR debug/55608
14786 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
14787 on failure.
14788 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
14789 (string_cst_pool_decl): New function.
14790 (optimize_one_addr_into_implicit_ptr): New function.
14791 (resolve_addr_in_expr): Optimize DWARF location expression
14792 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
14793 which doesn't live in memory, but has DW_AT_location or
14794 DW_AT_const_value, or refers to a string literal, into
14795 DW_OP_GNU_implicit_pointer.
14796 (optimize_location_into_implicit_ptr): New function.
14797 (resolve_addr): If removing DW_AT_location of a variable because
14798 it was DW_OP_addr of address of the variable, but the variable doesn't
14799 live in memory, try to emit const value attribute for the initializer.
14800
14801 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14802
14803 * tree.h (VECTOR_TYPE_P): New macro.
14804 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14805 TYPE_MODE): Use it.
14806 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14807 VEC_COND_EXPR cannot be lvalues.
14808 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14809
14810 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14811
14812 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14813 Restrict the transformation to equal modes.
14814
14815 2013-03-21 Richard Biener <rguenther@suse.de>
14816
14817 PR tree-optimization/39326
14818 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14819 (MEM_ANALYZABLE): Adjust.
14820 (record_mem_ref_loc): Move bitmap ops ...
14821 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14822 unanalyzable refs, do not record locations for it.
14823 (analyze_memory_references): Allocate ref zero as shared
14824 unanalyzable ref.
14825 (refs_independent_p): Do not test for unanalyzed mems here.
14826 (ref_indep_loop_p_1): Special-case disambiguation against
14827 the unanalyzed ref.
14828 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14829
14830 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14831
14832 * config/arm/arm-protos.h (tune_params): Add
14833 prefer_neon_for_64bits field.
14834 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14835 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14836 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14837 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14838 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14839 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14840 (arm_option_override): Handle -mneon-for-64bits new option.
14841 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14842 (prefer_neon_for_64bits): Declare new variable.
14843 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14844 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14845 (arch_enabled): Handle new arch types. Remove support for onlya8
14846 and nota8.
14847 (one_cmpldi2): Use new arch names.
14848 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14849 * config/arm/arm.opt (mneon-for-64bits): Add option.
14850 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14851 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14852 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14853 of onlya8.
14854 * doc/invoke.texi (-mneon-for-64bits): Document.
14855
14856 2013-03-21 Richard Biener <rguenther@suse.de>
14857
14858 PR tree-optimization/39326
14859 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14860 (sort_bbs_in_loop_postorder_cmp): New function.
14861 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14862
14863 2013-03-21 Richard Biener <rguenther@suse.de>
14864
14865 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14866 (vect_insert_into_interleaving_chain): Likewise.
14867 (vect_drs_dependent_in_basic_block): Inline ...
14868 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14869 split out from ...
14870 (vect_analyze_data_ref_dependence): ... here. Simplify.
14871 (vect_check_interleaving): Simplify.
14872 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14873 (vect_slp_analyze_data_ref_dependences): ... this new function.
14874 (dr_group_sort_cmp): New function.
14875 (vect_analyze_data_ref_accesses): Compute data-reference groups
14876 here instead of in vect_analyze_data_ref_dependence. Use
14877 a more efficient algorithm.
14878 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14879 vect_slp_analyze_data_ref_dependences. Call
14880 vect_analyze_data_ref_accesses earlier.
14881 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14882 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14883 (vect_slp_analyze_data_ref_dependences): New prototype.
14884
14885 2013-03-21 Richard Biener <rguenther@suse.de>
14886
14887 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14888 ref is stored in the loop.
14889 (find_refs_for_sm): Walk only over all stores.
14890 (store_motion_loop): Allocate from lim_bitmap_obstack.
14891 (store_motion): Likewise.
14892
14893 2013-03-21 Richard Biener <rguenther@suse.de>
14894
14895 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14896 Update virtual SSA form.
14897
14898 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14899
14900 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14901 * configure: Regenerate.
14902 * config.in: Regenerate.
14903 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14904 if !HAVE_LD_EH_FRAME_CIEV3.
14905
14906 2013-03-21 Richard Biener <rguenther@suse.de>
14907
14908 * tree-cfg.c (verify_expr_no_block): New function.
14909 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14910 nor DECL_VALUE_EXPR have locations with associated blocks.
14911 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14912 (clear_unused_block_pointer): Remove code dealing with
14913 blocks in DECL_DEBUG_EXPR locations.
14914
14915 2013-03-21 Richard Biener <rguenther@suse.de>
14916
14917 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14918 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14919 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14920 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14921 instead of DECL_DEBUG_EXPR_IS_FROM.
14922 * gimplify.c (gimplify_modify_expr): Likewise.
14923 * tree-cfg.c (verify_expr_location_1): Likewise.
14924 * tree-complex.c (create_one_component_var): Likewise.
14925 * tree-sra.c (create_access_replacement): Likewise.
14926 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14927 (clear_unused_block_pointer): Likewise.
14928 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14929 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14930 * var-tracking.c (var_debug_decl): Likewise.
14931 (track_expr_p): Likewise.
14932 * tree-inline.c (add_local_variables): Likewise. Set
14933 DECL_HAS_DEBUG_EXPR_P after copying it.
14934 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14935 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14936
14937 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14938
14939 PR bootstrap/56656
14940 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14941 * configure: Regenerate.
14942 * config.in: Regenerate.
14943 * config/i386/i386.md (*movdf_internal): Use
14944 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14945 movd instead of movq mnemonic for interunit moves.
14946 (*movdi_internal): Ditto.
14947
14948 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14949
14950 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14951 (abd<mode>_3): New pattern.
14952 (aba<mode>_3): New pattern.
14953 (fabd<mode>_3): New pattern.
14954
14955 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14956
14957 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14958 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14959 occurrence of REGISTER_PREFIX as its empty string.
14960
14961 2013-03-20 Jeff Law <law@redhat.com>
14962
14963 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14964 addititional equivalences for equality comparisons between an SSA_NAME
14965 and a constant where the SSA_NAME was set from a widening conversion.
14966
14967 2013-03-20 Walter Lee <walt@tilera.com>
14968
14969 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14970
14971 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14972
14973 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14974 depending on type attribute.
14975 (*movti_internal): Ditto.
14976 (*movtf_internal): Ditto.
14977 (*movxf_internal): Ditto.
14978 (*movdf_internal): Ditto.
14979 (*movsf_internal): Ditto.
14980
14981 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14982
14983 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14984 maybe_vex for sselog1 and ssemov types.
14985 (*movdi_internal): Reorder operand constraints.
14986 (*movsi_internal): Ditto. Set prefix attribute to
14987 maybe_vex for sselog1 and ssemov types.
14988 (*movtf_internal): Set prefix attribute to maybe_vex
14989 for sselog1 and ssemov types.
14990 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14991 DImode ssemov types. Reorder operand constraints.
14992 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14993 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14994 attribute for SImode ssemov types. Reorder operand constraints.
14995
14996 2013-03-20 Martin Jambor <mjambor@suse.cz>
14997
14998 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14999 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
15000
15001 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
15002
15003 * config/rs6000/predicates.md (indexed_address, update_address_mem
15004 update_indexed_address_mem): New predicates.
15005 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
15006 attribute for load/store instructions.
15007 * config/rs6000/dfp.md (movsd_store): Likewise.
15008 (movsd_load): Likewise.
15009 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
15010 (unnamed HI->DI extend define_insn): Likewise.
15011 (unnamed SI->DI extend define_insn): Likewise.
15012 (unnamed QI->SI extend define_insn): Likewise.
15013 (unnamed QI->HI extend define_insn): Likewise.
15014 (unnamed HI->SI extend define_insn): Likewise.
15015 (unnamed HI->SI extend define_insn): Likewise.
15016 (extendsfdf2_fpr): Likewise.
15017 (movsi_internal1): Likewise.
15018 (movsi_internal1_single): Likewise.
15019 (movhi_internal): Likewise.
15020 (movqi_internal): Likewise.
15021 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
15022 attribute for load/store instructions.
15023 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15024 instructions.
15025 (mov<mode>_softfloat): Likewise.
15026 (mov<mode>_hardfloat32): Likewise.
15027 (mov<mode>_hardfloat64): Likewise.
15028 (mov<mode>_softfloat64): Likewise.
15029 (movdi_internal32): Likewise.
15030 (movdi_internal64): Likewise.
15031 (probe_stack_<mode>): Likewise.
15032
15033 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15034
15035 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15036 floating point, and decimal floating point to reload iterator.
15037
15038 * config/rs6000/constraints.md (wl constraint): New constraints to
15039 return FLOAT_REGS if certain options are used to reduce the number
15040 of separate patterns that exist in the file.
15041 (wx constraint): Likewise.
15042 (wz constraint): Likewise.
15043
15044 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15045 -mdebug=reg, print wg, wl, wx, and wz constraints.
15046 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15047 Initialize the reload functions for 64-bit binary/decimal floating
15048 point types.
15049 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15050 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15051 create the buffer on the stack to overcome not having a 32-bit
15052 load and store.
15053 (rs6000_emit_move): Likewise.
15054 (rs6000_secondary_memory_needed_rtx): Likewise.
15055 (rs6000_alloc_sdmode_stack_slot): Likewise.
15056 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15057 via xxlxor, just like DFmode 0.0.
15058
15059 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15060 define as 1 if we are running on a power7 or newer.
15061 (enum r6000_reg_class_enum): Add new constraints.
15062
15063 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15064 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15065 with other moves by using conditional constraits (wg). Use LFIWZX
15066 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15067 (movsd splitter): Likewise.
15068 (movsd_hardfloat): Likewise.
15069 (movsd_softfloat): Likewise.
15070
15071 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15072 binary and decimal floating point moves.
15073 (fmove_ok): New attributes to combine binary and decimal floating
15074 point moves, and to combine power6x (mfpgpr) moves along normal
15075 floating moves.
15076 (real_value_to_target): Likewise.
15077 (f32_lr): Likewise.
15078 (f32_lm): Likewise.
15079 (f32_li): Likewise.
15080 (f32_sr): Likewise.
15081 (f32_sm): Likewise.
15082 (f32_si): Likewise.
15083 (movsf): Combine binary and decimal floating point moves. Combine
15084 power6x (mfpgpr) moves with other moves by using conditional
15085 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15086 (mov<mode> for SFmode/SDmode); Likewise.
15087 (SFmode/SDmode splitters): Likewise.
15088 (movsf_hardfloat): Likewise.
15089 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15090 (movsf_softfloat): Likewise.
15091 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15092
15093 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15094 wx and wz constraints.
15095
15096 * config/rs6000/constraints.md (wg constraint): New constraint to
15097 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15098
15099 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15100 constraint.
15101
15102 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15103 -mdebug=reg, print wg, wl, wx, and wz constraints.
15104 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15105 Initialize the reload functions for 64-bit binary/decimal floating
15106 point types.
15107 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15108 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15109 create the buffer on the stack to overcome not having a 32-bit
15110 load and store.
15111 (rs6000_emit_move): Likewise.
15112 (rs6000_secondary_memory_needed_rtx): Likewise.
15113 (rs6000_alloc_sdmode_stack_slot): Likewise.
15114 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15115 via xxlxor, just like DFmode 0.0.
15116
15117 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15118 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15119 with other moves by using conditional constraits (wg). Use LFIWZX
15120 and STFIWX for loading SDmode on power7.
15121 (movdd splitters): Likewise.
15122 (movdd_hardfloat32): Likewise.
15123 (movdd_softfloat32): Likewise.
15124 (movdd_hardfloat64_mfpgpr): Likewise.
15125 (movdd_hardfloat64): Likewise.
15126 (movdd_softfloat64): Likewise.
15127
15128 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15129 64-bit binary and decimal floating point moves.
15130 (FMOVE64X): Likewise.
15131 (movdf): Combine 64-bit binary and decimal floating point moves.
15132 Combine power6x (mfpgpr) moves with other moves by using
15133 conditional constraits (wg).
15134 (mov<mode> for DFmode/DDmode): Likewise.
15135 (DFmode/DDmode splitters): Likewise.
15136 (movdf_hardfloat32): Likewise.
15137 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15138 (movdf_softfloat32): Likewise.
15139 (movdf_hardfloat64_mfpgpr): Likewise.
15140 (movdf_hardfloat64): Likewise.
15141 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15142 (movdf_softfloat64): Likewise.
15143 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15144 (reload_<mode>_load): Move to later in the file so they aren't in
15145 the middle of the floating point move insns.
15146 (reload_<mode>_store): Likewise.
15147
15148 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15149 constraint.
15150
15151 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15152 constraint if -mdebug=reg.
15153 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15154 Enable using dd reload support if needed.
15155
15156 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15157 binary and decimal floating point moves in rs6000.md.
15158 (movtd_internal): Likewise.
15159
15160 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15161 decimal floating point moves.
15162 (movtf): Likewise.
15163 (movtf_internal): Likewise.
15164 (mov<mode>_internal, TDmode/TFmode): Likewise.
15165 (movtf_softfloat): Likewise.
15166 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15167
15168 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15169 movdi_internal64, using wg constraint for move direct operations.
15170 (movdi_internal64): Likewise.
15171
15172 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15173 MODES_TIEABLE_P for selected modes. Print the numerical value of
15174 the various virtual registers. Use GPR/FPR first/last values,
15175 instead of hard coding the register numbers. Print which modes
15176 have reload functions registered.
15177 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15178 settings before/after setting cpu, target and subtarget settings.
15179 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15180 and for secondary reload failures in rs6000_secondary_reload_inner.
15181 (rs6000_secondary_reload_fail): Likewise.
15182 (rs6000_secondary_reload_inner): Likewise.
15183
15184 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15185 macros for first/last GPR and FPR registers.
15186 (LAST_GPR_REGNO): Likewise.
15187 (FIRST_FPR_REGNO): Likewise.
15188 (LAST_FPR_REGNO): Likewise.
15189
15190 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
15191 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
15192 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
15193 (vcond<mode><mode>): Likewise.
15194 (vcondu<mode><mode>): Likewise.
15195 (vector_gtu<mode>): Likewise.
15196 (vector_gte<mode>): Likewise.
15197 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
15198 to prevent the compiler from converting DImode operations to TImode.
15199 (ior<mode>3): Likewise.
15200 (and<mode>3): Likewise.
15201 (one_cmpl<mode>2): Likewise.
15202 (nor<mode>3): Likewise.
15203 (andc<mode>3): Likewise.
15204
15205 * config/rs6000/constraints.md (wt constraint): New constraint
15206 that returns VSX_REGS if TImode is allowed in VSX registers.
15207
15208 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
15209 constant under VSX.
15210
15211 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
15212 similar to TImode, but it is restricted to being in the GPRs.
15213
15214 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
15215 TImode to occupy a single VSX register.
15216
15217 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
15218 -mvsx-timode for power7/power8.
15219 (power7 cpu): Likewise.
15220 (power8 cpu): Likewise.
15221
15222 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
15223 sure that TFmode/TDmode take up two registers if they are ever
15224 allowed in the upper VSX registers.
15225 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
15226 registers.
15227 (rs6000_init_hard_regno_mode_ok): Likewise.
15228 (rs6000_debug_reg_global): Add debugging for PTImode and wt
15229 constraint. Print if LRA is turned on.
15230 (rs6000_option_override_internal): Give an error if -mvsx-timode
15231 and VSX is not enabled.
15232 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
15233 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
15234 to reg+offset addressing. Use PTImode when checking offset
15235 addresses for validity.
15236 (reg_offset_addressing_ok_p): Likewise.
15237 (rs6000_legitimate_offset_address_p): Likewise.
15238 (rs6000_legitimize_address): Likewise.
15239 (rs6000_legitimize_reload_address): Likewise.
15240 (rs6000_legitimate_address_p): Likewise.
15241 (rs6000_eliminate_indexed_memrefs): Likewise.
15242 (rs6000_emit_move): Likewise.
15243 (rs6000_secondary_reload): Likewise.
15244 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
15245 reloads to fpr registers to continue to use reg+offset addressing,
15246 but 64-bit reloads to altivec registers need reg+reg addressing.
15247 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
15248 it. Treat LO_SUM like a PLUS operation.
15249 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
15250 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
15251 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
15252 registers to share a register with a smaller sized type, since VSX
15253 puts scalars in the upper 64-bits.
15254 (print_operand): Add support for PTImode.
15255 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
15256 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
15257 registers, but don't have arithmetic support.
15258 (rs6000_memory_move_cost): Add test for VSX.
15259 (rs6000_opt_masks): Add -mvsx-timode.
15260
15261 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
15262 for TImode.
15263 (VSs): Likewise.
15264 (VSr): Use wt constraint for TImode.
15265 (VSv): Drop TImode support.
15266 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
15267 (vsx_movti_64bit): Likewise.
15268 (vsx_movti_32bit): Likewise.
15269 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
15270 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
15271 one '?' on the appropriate output constraint. Do not allow TImode
15272 logical operations on 32-bit systems.
15273 (vsx_ior<mode>3): Likewise.
15274 (vsx_xor<mode>3): Likewise.
15275 (vsx_one_cmpl<mode>2): Likewise.
15276 (vsx_nor<mode>3): Likewise.
15277 (vsx_andc<mode>3): Likewise.
15278 (vsx_concat_<mode>): Likewise.
15279 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
15280
15281 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
15282 OPTION_MASK_VSX_TIMODE.
15283 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
15284 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
15285
15286 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
15287 (TI2 iterator): New iterator for TImode, PTImode.
15288 (wd mode attribute): Add values for vector types.
15289 (movti_string): Replace TI move operations with operations for TImode
15290 and PTImode. Add support for TImode being allowed in VSX registers.
15291 (mov<mode>_string, TImode/PTImode): Likewise.
15292 (movti_ppc64): Likewise.
15293 (mov<mode>_ppc64, TImode/PTImode): Likewise.
15294 (TI mode splitters): Likewise.
15295
15296 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
15297 constraint.
15298
15299 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
15300
15301 PR tree-optimization/56355
15302 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
15303 Also handle integers with undefined overflow.
15304
15305 2013-03-20 Catherine Moore <clm@codesourcery.com>
15306 Maciej W. Rozycki <macro@codesourcery.com>
15307 Tom de Vries <tom@codesourcery.com>
15308 Nathan Sidwell <nathan@codesourcery.com>
15309 Iain Sandoe <iain@codesourcery.com>
15310 Nathan Froyd <froydnj@codesourcery.com>
15311 Chao-ying Fu <fu@mips.com>
15312
15313 * doc/extend.texi: (micromips, nomicromips, nocompression):
15314 Document new function attributes.
15315 * doc/invoke.texi (minterlink-compressed, mmicromips,
15316 m14k, m14ke, m14kec): Document new options.
15317 (minterlink-mips16): Update documentation.
15318 * doc/md.texi (ZC, ZD): Document new constraints.
15319 * configure.ac (gcc_cv_as_micromips): Check if linker
15320 supports the .set micromips directive.
15321 * configure: Regenerate.
15322 * config.in: Regenerate.
15323 * config/mips/mips-tables.opt: Regenerate.
15324 * config/mips/micromips.md: New file.
15325 * constraints.md (ZC, ZD): New constraints.
15326 * config/mips/predicates.md (movep_src_register): New predicate.
15327 (movep_src_operand): New predicate.
15328 (non_volatile_mem_operand): New predicate.
15329 * config/mips/mips.md (multimem): New type.
15330 (length): Differentiate between 17-bit and 18-bit branch offsets.
15331 (MOVEP1, MOVEP2): New mode iterator.
15332 (mov_<load>l): Use ZC constraint.
15333 (mov_<load>r): Likewise.
15334 (mov_<store>l): Likewise.
15335 (mov_<store>r): Likewise.
15336 (*branch_equality<mode>_inverted): Add microMIPS support.
15337 (*branch_equality<mode>): Likewise.
15338 (*jump_absolute): Likewise.
15339 (indirect_jump_<mode>): Likewise.
15340 (tablejump_<mode>): Likewise.
15341 (<optab>_internal): Likewise.
15342 (sibcall_internal): Likewise.
15343 (sibcall_value_internal): Likewise.
15344 (prefetch): Use constraint ZD.
15345 * config/mips/mips.opt (minterlink-compressed): New option.
15346 (minterlink-mips16): Now an alias for minterlink-compressed.
15347 (mmicromips): New option.
15348 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15349 (compare_and_swap_12): Likewise.
15350 (sync_add<mode>): Likewise.
15351 (sync_<optab>_12): Likewise.
15352 (sync_old_<optab>_12): Likewise.
15353 (sync_new_<optab>_12): Likewise.
15354 (sync_nand_12): Likewise.
15355 (sync_old_nand_12): Likewise.
15356 (sync_new_nand_12): Likewise.
15357 (sync_sub<mode>): Likewise.
15358 (sync_old_add<mode>): Likewise.
15359 (sync_old_sub<mode>): Likewise.
15360 (sync_new_add<mode>): Likewise.
15361 (sync_new_sub<mode>): Likewise.
15362 (sync_<optab><mode>): Likewise.
15363 (sync_old_<optab><mode>): Likewise.
15364 (sync_new_<optab><mode>): Likewise.
15365 (sync_nand<mode>): Likewise.
15366 (sync_old_nand<mode>): Likewise.
15367 (sync_new_nand<mode>): Likewise.
15368 (sync_lock_test_and_set<mode>): Likewise.
15369 (test_and_set_12): Likewise.
15370 (atomic_compare_and_swap<mode>): Likewise.
15371 (atomic_exchange<mode>_llsc): Likewise.
15372 (atomic_fetch_add<mode>_llsc): Likewise.
15373 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15374 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15375 (umips_save_restore_pattern_p): Likewise.
15376 (umips_load_store_pair_p): Likewise.
15377 (umips_output_load_store_pair): Likewise.
15378 (umips_movep_target_p): Likewise.
15379 (umips_12bit_offset_address_p): Likewise.
15380 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15381 (mips_base_mips16): Rename this...
15382 (mips_base_compression_flags): ...to this. Update all uses.
15383 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15384 (mips_mips16_decl_p): Delete.
15385 (mips_nomips16_decl_p): Delete.
15386 (mips_get_compress_on_flags): New function.
15387 (mips_get_compress_off_flags): New function.
15388 (mips_get_compress_mode): New function.
15389 (mips_get_compress_on_name): New function.
15390 (mips_get_compress_off_name): New function.
15391 (mips_insert_attributes): Support multiple compression types.
15392 (mips_merge_decl_attributes): Likewise.
15393 (umips_12bit_offset_address_p): New function.
15394 (mips_start_function_definition): Emit .set micromips directive.
15395 (mips_call_may_need_jalx_p): New function.
15396 (mips_function_ok_for_sibcall): Add microMIPS support.
15397 (mips_print_operand_punctuation): Support short delay slots and
15398 compact jumps.
15399 (umips_swm_mask, umips_swm_encoding): New.
15400 (umips_build_save_restore): New function.
15401 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15402 (was_mips16_p): Remove.
15403 (old_compression_mode): New.
15404 (mips_set_compression_mode): New function.
15405 (mips_set_current_function): Add microMIPS support.
15406 (mips_option_override): Likewise.
15407 (umips_save_restore_pattern_p): New function.
15408 (umips_output_save_restore): New function.
15409 (umips_load_store_pair_p_1): New function.
15410 (umips_load_store_pair_p): New function.
15411 (umips_output_load_store_pair_1): New function.
15412 (umips_output_load_store_pair): New function.
15413 (umips_movep_target_p) New function.
15414 (mips_prepare_pch_save): Add microMIPS support.
15415 * config/mips/mips.h (TARGET_COMPRESSION): New.
15416 (TARGET_CPU_CPP_BUILTINS): Update macro
15417 to use new compression flags and to support microMIPS.
15418 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15419 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15420 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15421 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15422 (ASM_SPEC): Support mmicromips and mno-micromips.
15423 (M16STORE_REG_P): New macro.
15424 (MIPS_CALL): Support TARGET_MICROMIPS.
15425 (MICROMIPS_J): New macro.
15426 (mips_base_mips16): Rename this...
15427 (mips_base_compression_flags): ...to this.
15428 (UMIPS_12BIT_OFFSET_P): New macro.
15429 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15430 (MULTILIB_DIRNAMES): Likewise.
15431 2013-03-20 Richard Biener <rguenther@suse.de>
15432
15433 PR tree-optimization/56661
15434 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
15435 the result does not have to be distinct.
15436
15437 2013-03-20 Richard Biener <rguenther@suse.de>
15438
15439 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
15440 remap_gimple_op_r.
15441
15442 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15443 Steven Bosscher <steven@gcc.gnu.org>
15444
15445 PR rtl-optimization/56605
15446 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
15447
15448 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15449
15450 PR bootstrap/56656
15451 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
15452 that require movd instead of movq.
15453
15454 2013-03-20 Richard Biener <rguenther@suse.de>
15455
15456 * tree-ssa-structalias.c (struct variable_info): Add pointer
15457 to the first field of an aggregate with sub-vars. Make
15458 this and the pointer to the next subfield its ID.
15459 (vi_next): New function.
15460 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
15461 storedanything_id, integer_id): Increment by one.
15462 (new_var_info, get_call_vi, lookup_call_clobber_vi,
15463 get_call_clobber_vi): Adjust.
15464 (solution_set_expand): Simplify and speedup.
15465 (solution_set_add): Inline into ...
15466 (set_union_with_increment): ... this. Adjust accordingly.
15467 (do_sd_constraint): Likewise.
15468 (do_ds_constraint): Likewise.
15469 (do_complex_constraint): Simplify.
15470 (build_pred_graph): Adjust.
15471 (solve_graph): Likewise. Simplify and speedup.
15472 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
15473 get_constraint_for_component_ref, get_constraint_for_1,
15474 first_vi_for_offset, first_or_preceding_vi_for_offset,
15475 create_function_info_for, create_variable_info_for_1,
15476 create_variable_info_for, intra_create_variable_infos): Adjust.
15477 (init_base_vars): Push NULL for ID zero.
15478 (compute_points_to_sets): Adjust.
15479
15480 2013-03-20 Richard Biener <rguenther@suse.de>
15481
15482 * cfgloop.c (verify_loop_structure): Streamline and avoid
15483 ICEing on corrupt loop tree.
15484 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
15485 loop tree.
15486
15487 2013-03-20 Richard Biener <rguenther@suse.de>
15488
15489 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
15490 check whether an SSA update is needed.
15491
15492 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
15493
15494 * config/mips/constraints.md (T): Rename to...
15495 (Yf): ...this.
15496 (U): Rename to...
15497 (Yd): ...this.
15498 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
15499 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
15500
15501 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15502
15503 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
15504 (*subsi3_carryin_uxtw): Likewise.
15505
15506 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15507
15508 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
15509 (*rorsi3_insn_uxtw): Likewise.
15510
15511 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15512
15513 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
15514 (*extrsi5_insn_uxtw): Likewise.
15515
15516 2013-03-19 Richard Biener <rguenther@suse.de>
15517
15518 PR tree-optimization/56273
15519 * passes.c (init_optimization_passes): Move second VRP after DOM.
15520
15521 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
15522
15523 * config/i386/i386.md (*movti_internal): Merge from
15524 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
15525 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
15526 nox64 isa attributes.
15527
15528 2013-03-18 Richard Biener <rguenther@suse.de>
15529
15530 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
15531 (unite): Likewise.
15532 (merge_node_constraints): Likewise.
15533 (build_succ_graph): Likewise.
15534 (valid_graph_edge): Inline into single caller.
15535 (unify_nodes): Likewise. Use bitmap_set_bit return value
15536 and cache varinfo.
15537 (scc_visit): Fix formatting and variable use.
15538 (do_sd_constraint): Use gcc_checking_assert.
15539 (do_ds_constraint): Likewise.
15540 (do_complex_constraint): Likewise.
15541 (condense_visit): Likewise. Cleanup.
15542 (dump_pred_graph): New function.
15543 (perform_var_substitution): Dump the pred-graph before
15544 variable substitution.
15545 (find_equivalent_node): Use gcc_checking_assert.
15546 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
15547
15548 2013-03-18 Richard Biener <rguenther@suse.de>
15549
15550 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
15551 Remove cond_expr_stmt_list argument and do not gimplify the
15552 built expression.
15553 (vect_loop_versioning): Adjust.
15554 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
15555 Cleanup to use less temporaries.
15556 (vect_create_data_ref_ptr): Cleanup.
15557
15558 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15559
15560 PR tree-optimization/56635
15561 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
15562 require types_compatible_p types.
15563
15564 2013-03-18 Nick Clifton <nickc@redhat.com>
15565
15566 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
15567 spurious backslash.
15568
15569 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
15570 Add missing line to comment describing function.
15571
15572 2013-03-18 Richard Biener <rguenther@suse.de>
15573
15574 PR tree-optimization/56210
15575 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
15576 Handle string / character search functions.
15577 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
15578
15579 2013-03-18 Richard Biener <rguenther@suse.de>
15580
15581 PR middle-end/56483
15582 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
15583 and implement properly.
15584 * gimple.h (gimple_cond_single_var_p): Remove.
15585
15586 2013-03-18 Richard Biener <rguenther@suse.de>
15587
15588 * tree-data-ref.h (find_data_references_in_loop): Declare.
15589 * tree-data-ref.c (get_references_in_stmt): Use a stack
15590 vector pre-allocated in the callers.
15591 (find_data_references_in_stmt): Adjust.
15592 (graphite_find_data_references_in_stmt): Likewise.
15593 (create_rdg_vertices): Likewise.
15594 (find_data_references_in_loop): Export.
15595 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
15596 Compute dependences here...
15597 (vect_analyze_data_refs): ...not here. When we encounter
15598 a non-vectorizable data reference in basic-block vectorization
15599 truncate the data reference vector. Do not bother to
15600 fixup data-dependence information for gather loads.
15601 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
15602 of data references, as reported.
15603
15604 2013-03-18 Richard Biener <rguenther@suse.de>
15605
15606 PR tree-optimization/3713
15607 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
15608 has_constants and expr.
15609 (stmt_has_constants): Properly valueize SSA names when deciding
15610 whether the stmt has constants.
15611
15612 2013-03-18 Richard Biener <rguenther@suse.de>
15613
15614 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
15615 whole function when there is nothing to do.
15616 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
15617 * tree-vectorizer.c (vectorize_loops): Update virtual and
15618 loop-closed SSA once.
15619 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
15620
15621 2013-03-18 Richard Biener <rguenther@suse.de>
15622
15623 PR middle-end/56113
15624 * domwalk.c (bb_postorder): New global static.
15625 (cmp_bb_postorder): New function.
15626 (walk_dominator_tree): Replace scheme imposing an order for
15627 visiting dominator sons by one sorting them at the time they
15628 are pushed on the stack.
15629
15630 2013-03-18 Richard Biener <rguenther@suse.de>
15631
15632 PR tree-optimization/39326
15633 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
15634 (struct mem_ref): Replace mem member with ao_ref typed member.
15635 (MEM_ANALYZABLE): Adjust.
15636 (memref_eq): Likewise.
15637 (mem_ref_alloc): Likewise.
15638 (gather_mem_refs_stmt): Likewise.
15639 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
15640 (execute_sm_if_changed_flag_set): Adjust.
15641 (execute_sm): Likewise.
15642 (ref_always_accessed_p): Likewise.
15643 (refs_independent_p): Likewise.
15644 (can_sm_ref_p): Likewise.
15645
15646 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15647
15648 PR c/56566
15649 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
15650 return 1 even for !unsignedp.
15651
15652 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15653
15654 * config/i386/i386.md (isa): Add x64 and nox64.
15655 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
15656 (*pushtf): Enable *roF alternative for x64 isa only.
15657 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
15658 mode attribute of integer alternatives to DImode for TARGET_64BIT.
15659 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
15660 (*movtf_internal): Merge from *movtf_internal_rex64 and
15661 *movtf_internal_sse. Use x64 and nox64 isa attributes.
15662 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
15663 nox64 isa attributes.
15664 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
15665 nox64 isa attributes.
15666 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
15667
15668 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15669
15670 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15671
15672 2013-03-17 Jakub Jelinek <jakub@redhat.com>
15673
15674 PR target/56640
15675 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
15676 class names. Remove trailing comma after "ALL_REGS".
15677
15678 2013-03-16 Jan Hubicka <jh@suse.cz>
15679
15680 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
15681 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
15682 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
15683 of cgraph_get_create_node.
15684 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
15685
15686 2013-03-16 Jason Merrill <jason@redhat.com>
15687
15688 PR debug/49090
15689 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
15690 with DW_AT_default_value.
15691
15692 2013-03-16 Jakub Jelinek <jakub@redhat.com>
15693
15694 * BASE-VER: Set to 4.9.0.
15695
15696 2013-03-14 Andi Kleen <ak@linux.intel.com>
15697
15698 PR target/56619
15699 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
15700 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
15701 Document _x* TSX intrinsics.
15702
15703 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
15704 David Holsgrove <david.holsgrove@xilinx.com>
15705
15706 * configure.ac: Add MicroBlaze TLS support detection.
15707 * configure: Regenerate.
15708 * config/microblaze/microblaze-protos.h
15709 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
15710 symbol_mentioned_p, label_mentioned_p): Add prototypes.
15711 * config/microblaze/microblaze.c (microblaze_address_type): Add
15712 ADDRESS_TLS and tls_reloc address types.
15713 (microblaze_address_info): Add tls_reloc.
15714 (TARGET_HAVE_TLS): Define.
15715 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
15716 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
15717 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
15718 load_tls_operand, microblaze_call_tls_get_addr,
15719 microblaze_legitimize_tls_address): New functions.
15720 (microblaze_classify_unspec): Handle UNSPEC_TLS.
15721 (get_base_reg): Use microblaze_tls_symbol_p.
15722 (microblaze_classify_address): Handle TLS.
15723 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
15724 label_mentioned_p and microblaze_tls_referenced_p.
15725 (microblaze_legitimize_address): Handle TLS.
15726 (microblaze_address_insns): Handle ADDRESS_TLS.
15727 (pic_address_needs_scratch): Handle TLS.
15728 (print_operand_address): Handle TLS.
15729 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
15730 (microblaze_expand_move): Handle TLS.
15731 (microblaze_legitimate_constant_p): Check
15732 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
15733 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
15734 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
15735 (PIC_OFFSET_TABLE_REGNUM): Set.
15736 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
15737 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
15738 (addsi3, movsi_internal2, movdf_internal): Update constraints
15739 * config/microblaze/predicates.md (arith_plus_operand): Define
15740 (move_operand): Redefine as move_src_operand,
15741 check microblaze_tls_referenced_p.
15742
15743 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15744
15745 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
15746 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
15747
15748 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15749
15750 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
15751 CC mode for AND.
15752
15753 2013-03-14 Jakub Jelinek <jakub@redhat.com>
15754
15755 PR tree-optimization/53265
15756 * common.opt (Waggressive-loop-optimizations): New option.
15757 * tree-ssa-loop-niter.c: Include tree-pass.h.
15758 (do_warn_aggressive_loop_optimizations): New function.
15759 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
15760 if number_of_latch_executions returned constant.
15761 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
15762 early. If number_of_latch_executions returned constant, set
15763 nb_iterations_upper_bound back to it.
15764 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
15765 field.
15766 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
15767 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
15768
15769 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
15770 (MULTILIB_OSDIRNAMES): Set.
15771 * genmultilib: If defaultosdirname doesn't start with :: , set
15772 defaultosdirname2 instead, clear it and emit two . multilib_raw
15773 entries instead of just one.
15774
15775 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
15776
15777 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15778 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
15779 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15780 (SUBTARGET_OVERRIDE_OPTIONS): New.
15781
15782 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
15783
15784 PR target/49880
15785 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
15786 (musermode): Convert to Var(TARGET_USERMODE).
15787 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
15788 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
15789 * config/sh/sh.c (sh_option_override): Use
15790 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
15791 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
15792 condition.
15793 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
15794 TARGET_SH4.
15795 (udivsi3_i4_single, divsi3_i4_single): Use
15796 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
15797
15798 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15799
15800 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15801 default setting.
15802
15803 2013-03-13 Richard Biener <rguenther@suse.de>
15804
15805 PR tree-optimization/56608
15806 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15807 calls when vectorizing basic-blocks.
15808
15809 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15810
15811 PR plugins/45078
15812 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15813 tm_file.
15814
15815 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15816
15817 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15818
15819 2013-03-11 Jan Hubicka <jh@suse.cz>
15820
15821 PR lto/56557
15822 * lto-streamer-out.c (output_symbol_p): Skip references from
15823 constructors of external variables.
15824
15825 2013-03-11 Jan Hubicka <jh@suse.cz>
15826
15827 PR middle-end/56571
15828 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15829 from pseudos.
15830 * emit-rtl.c (verify_rtx_sharing): Likewise.
15831 (copy_insn_1): Likewise.
15832 * rtl.c (copy_rtx): Likewise.
15833
15834 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15835
15836 PR target/56591
15837 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15838 output_operand_lossage message.
15839
15840 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15841
15842 PR target/56470
15843 * arm.c (shift_op): Validate RTL pattern on the fly.
15844 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15845 the RTL.
15846
15847 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15848
15849 PR target/56347
15850 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15851 new call patterns that clobber %fr12.
15852 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15853 split and postreload patterns.
15854 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15855 registers %fr12 and %fr12R as call used.
15856
15857 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15858
15859 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15860 (canon_address, record_store, replace_read, check_mem_read_rtx,
15861 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15862 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15863 rest_of_handle_dse): Likewise.
15864
15865 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15866
15867 PR middle-end/56524
15868 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15869 Add base_optabs.
15870 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15871 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15872 (save_optabs_if_changed): Replace with...
15873 (init_tree_optimization_optabs): ...this.
15874 * optabs.c (save_optabs_if_changed): Rename to...
15875 (init_tree_optimization_optabs): ...this. Take the optimization node
15876 as argument. Do nothing if the base optabs are already correct.
15877 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15878 to recompute optabs.
15879 * function.h (function): Remove optabs field.
15880 * function.c (invoke_set_current_function_hook): Call
15881 init_tree_optimization_optabs. Use the result to initialize
15882 this_fn_optabs.
15883
15884 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15885
15886 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15887 if GTMA_HAS_NO_INSTRUMENTATION.
15888 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15889 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15890 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15891 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15892 GTMA_HAS_NO_INSTRUMENTATION.
15893
15894 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15895
15896 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15897 libasan_preinit.o.
15898
15899 2013-03-08 Marek Polacek <polacek@redhat.com>
15900 Jakub Jelinek <jakub@redhat.com>
15901
15902 PR tree-optimization/56478
15903 * predict.c (is_comparison_with_loop_invariant_p): Change the
15904 type of loop_step to tree.
15905 (predict_loops): Adjust.
15906 (predict_iv_comparison): Perform the computations on double_ints.
15907
15908 2013-03-08 Richard Biener <rguenther@suse.de>
15909
15910 PR tree-optimization/56570
15911 * tree-cfg.c (verify_expr_location_1): Verify locations for
15912 DECL_DEBUG_EXPR.
15913 * tree-sra.c (create_access_replacement): Strip locations
15914 from DECL_DEBUG_EXPRs.
15915
15916 2013-03-08 Richard Biener <rguenther@suse.de>
15917
15918 * tree-inline.c (expand_call_inline): Do not associate
15919 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15920 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15921
15922 2013-03-08 Richard Biener <rguenther@suse.de>
15923
15924 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15925 or block changes with -Og. Fix for location / block encoding
15926 changes and PHI arguments with locations.
15927
15928 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15929
15930 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15931 for all counters.
15932 (struct output_info): Likewise.
15933 (register_overhead): Remove bad gcc_assert.
15934 (bitmap_find_bit): If there is only a single bitmap element, do not
15935 count a miss as a search.
15936 (print_statistics): Update for counter type changes.
15937 (dump_bitmap_statistics): Likewise. Print headers such that they
15938 are properly lined up with the printed counters.
15939
15940 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15941
15942 PR tree-optimization/56559
15943 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15944 check that it has only a single use.
15945
15946 2013-03-07 Richard Biener <rguenther@suse.de>
15947
15948 * doc/invoke.texi (fwhole-program): Discourage use in combination
15949 with -flto.
15950
15951 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15952
15953 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15954
15955 PR tree-optimization/56539
15956 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15957 instead of GSI_CONTINUE_LINKING as last argument to
15958 force_gimple_operand_gsi. Adjust function comment.
15959
15960 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15961 aarch64-cores.def.
15962
15963 PR middle-end/56548
15964 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15965 promoted mode, convert the result back to the original mode.
15966
15967 2013-03-06 Richard Biener <rguenther@suse.de>
15968
15969 PR middle-end/56294
15970 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15971 (insert_updated_phi_nodes_compare_uids): New function.
15972 (update_ssa): Sort symbols_to_rename after UID before
15973 traversing it to insert PHI nodes.
15974
15975 2013-03-06 Richard Biener <rguenther@suse.de>
15976
15977 PR middle-end/50494
15978 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15979 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15980
15981 Revert
15982 2013-02-13 Richard Biener <rguenther@suse.de>
15983
15984 PR lto/50494
15985 * varasm.c (output_constant_def_1): Get the decl representing
15986 the constant as argument.
15987 (output_constant_def): Wrap output_constant_def_1.
15988 (make_decl_rtl): Use output_constant_def_1 with the decl
15989 representing the constant.
15990 (build_constant_desc): Optionally re-use a decl already
15991 representing the constant.
15992 (tree_output_constant_def): Adjust.
15993
15994 2013-03-06 Joey Ye <joey.ye@arm.com>
15995
15996 PR lto/50293
15997 * gcc.c (convert_white_space): New function.
15998 (main): Handles white space in function name.
15999
16000 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
16001
16002 PR target/56529
16003 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
16004 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
16005 to SH_DIV_CALL_TABLE for TARGET_SH2.
16006 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
16007 list.
16008 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
16009 call-table options.
16010
16011 2013-03-05 Sterling Augustine <saugustine@google.com>
16012 Cary Coutant <ccoutant@google.com>
16013
16014 PR debug/55364
16015 * dwarf2out.c (resolve_addr): Don't call
16016 remove_loc_list_addr_table_entries a second time for the same
16017 expression.
16018
16019 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16020
16021 PR debug/56510
16022 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
16023 (avoid_complex_debug_insns): New function.
16024 (expand_debug_locations): Call it.
16025
16026 PR rtl-optimization/56484
16027 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16028 lifetimes of hard registers on small register class machines.
16029
16030 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16031
16032 * config/microblaze/microblaze-protos.h: Rename
16033 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16034 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16035 fast_interrupt.
16036 (microblaze_fast_interrupt_function_p): New function.
16037 (microblaze_is_interrupt_handler): Rename to
16038 microblaze_is_interrupt_variant and add fast_interrupt check.
16039 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16040 (save_restore_insns): Likewise.
16041 (compute_frame_size): Likewise.
16042 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16043 (microblaze_globalize_label): Likewise.
16044 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16045 * config/microblaze/microblaze.md: Use wrapper
16046 microblaze_is_interrupt_variant.
16047
16048 2013-03-05 Kai Tietz <ktietz@redhat.com>
16049
16050 * sdbout.c (sdbout_one_type): Switch to current function's section
16051 supporting cold/hot.
16052
16053 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16054
16055 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16056 -mxl-reorder.
16057
16058 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16059
16060 PR middle-end/56461
16061 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16062 if VALGRIND_GET_VBITS is defined, temporarily make object
16063 memory all defined, and restore previous valgrind addressability
16064 and definability afterwards. Free this_object at the end.
16065
16066 PR middle-end/56461
16067 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16068 right before calling lra_create_live_ranges, also call it
16069 when clearing live_p. Only call lra_clear_live_ranges
16070 at the end if live_p.
16071
16072 PR middle-end/56461
16073 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16074
16075 2013-03-05 Richard Biener <rguenther@suse.de>
16076
16077 PR tree-optimization/56521
16078 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16079 value-id.
16080
16081 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16082
16083 PR c++/55135
16084 * except.h (remove_unreachable_eh_regions): New prototype.
16085 * except.c (remove_eh_handler_splicer): New function, split out
16086 of remove_eh_handler.
16087 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16088 warning about running it on many EH regions one at a time.
16089 (remove_unreachable_eh_regions_worker): New function, walk the
16090 EH tree in depth-first order and remove non-marked regions.
16091 (remove_unreachable_eh_regions): New function.
16092 * tree-eh.c (mark_reachable_handlers): New function, split out
16093 from remove_unreachable_handlers.
16094 (remove_unreachable_handlers): Use mark_reachable_handlers and
16095 remove_unreachable_eh_regions.
16096 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16097 and remove_unreachable_eh_regions.
16098
16099 2013-03-05 Richard Biener <rguenther@suse.de>
16100
16101 PR middle-end/56525
16102 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16103 not freeing them until the end.
16104
16105 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16106
16107 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16108
16109 2013-03-05 Richard Biener <rguenther@suse.de>
16110
16111 PR tree-optimization/56270
16112 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16113 of loads after scheduling an SLP instance.
16114
16115 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16116
16117 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16118 tic6x.exp.
16119 (check_gcc_parallelize): Run guality.exp as a separate job from
16120 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16121 struct-layout-1.exp with stackalign.exp.
16122
16123 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16124
16125 PR middle-end/56461
16126 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16127 load_index sbitmap even if some bit in it isn't set.
16128
16129 PR middle-end/56461
16130 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16131 (discover_iteration_bound_by_body_walk): Change queues to
16132 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16133 spelling in comment. Call safe_push on queues[bound_index] directly.
16134 Release queues[queue_index] in every iteration unconditionally.
16135 Release bounds vector.
16136
16137 PR middle-end/56461
16138 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16139 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16140 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16141 inner_phis vector.
16142
16143 2013-03-05 Richard Biener <rguenther@suse.de>
16144
16145 PR lto/56515
16146 * tree-inline.c (remap_blocks_to_null): New function.
16147 (expand_call_inline): When expanding a call stmt without
16148 an associated block inline remap all callee blocks to NULL.
16149
16150 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16151
16152 PR rtl-optimization/56494
16153 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16154 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16155 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16156
16157 PR middle-end/56461
16158 * sel-sched-ir.c (free_sched_pools): Release
16159 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16160 if succs_info_pool.max_top isn't -1.
16161
16162 PR bootstrap/56509
16163 * opts.c (opts_obstack, opts_concat): Moved to...
16164 * opts-common.c (opts_obstack, opts_concat): ... here.
16165
16166 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16167
16168 PR middle-end/56461
16169 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16170
16171 2013-03-04 Martin Jambor <mjambor@suse.cz>
16172
16173 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16174 all appropriate places.
16175
16176 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16177
16178 PR tree-optimization/56424
16179 * ipa-split.c (split_function): Do not set the RSO flag if result is
16180 not by reference and its type is a register type.
16181
16182 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16183
16184 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16185 (microblaze_legitimate_pic_operand): Likewise
16186 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16187 new function microblaze_legitimate_pic_operand
16188 * config/microblaze/microblaze-protos.h
16189 (microblaze_legitimate_pic_operand): Declare.
16190
16191 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16192
16193 * config/microblaze/predicates.md (call_insn_simple_operand):
16194 New predicate for supported rtx code types.
16195 * config/microblaze/microblaze.md (call_internal1): Use
16196 call_insn_simple_operand predicate.
16197
16198 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16199
16200 PR middle-end/56461
16201 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
16202 partitions.ordered_remove.
16203
16204 PR middle-end/56461
16205 * tree-vect-stmts.c (vectorizable_conversion): Don't call
16206 vec_oprnds0.create (1) for modifier == NONE.
16207
16208 PR middle-end/56461
16209 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
16210 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
16211 vec_oprnds1 right before pushing anything to it for
16212 scalar_shift_arg.
16213
16214 PR middle-end/56461
16215 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
16216 set nbbs to 0 instead of having separate code path.
16217 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
16218 instead of false as last argument if returning NULL.
16219
16220 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
16221
16222 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
16223 the attribute is now called "target" instead of "option".
16224 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
16225 * doc/tm.texi.in (Target Attributes): Likewise document the correct
16226 attribute/pragma name for TARGET_OPTION_VALID_P and
16227 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
16228 * doc/tm.texi: Regenerated.
16229
16230 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
16231
16232 * config/microblaze/microblaze.c:
16233 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
16234 * config/microblaze/microblaze.h: Add -mxl-reorder to
16235 DRIVER_SELF_SPECS.
16236 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
16237 instructions emitted if TARGET_REORDER.
16238 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
16239 or 0 for -m/-mno case, but initialises as 2 to detect default use case
16240 separately.
16241
16242 2013-03-01 Xinliang David Li <davidxl@google.com>
16243
16244 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
16245 walk length.
16246
16247 2013-03-01 Jakub Jelinek <jakub@redhat.com>
16248
16249 PR middle-end/56461
16250 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
16251 vector even when returning true. Fix up function comment formatting.
16252
16253 PR middle-end/56461
16254 * ira-build.c (ira_loop_nodes_count): New variable.
16255 (create_loop_tree_nodes): Initialize it.
16256 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
16257
16258 PR middle-end/56461
16259 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
16260 method on dr_chain and result_chain.
16261 * tree-vect-stmts.c (vectorizable_store): Only call
16262 result_chain.create if j == 0.
16263
16264 PR middle-end/56461
16265 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
16266 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
16267 before overwriting it.
16268
16269 2013-03-01 Tobias Burnus <burnus@net-b.de>
16270
16271 * doc/extended.texi (C Extensions): Change order in @menu
16272 to match @node.
16273 (Other MIPS Built-in Functions): Move last MIPS entry before
16274 "picoChip Built-in Functions".
16275 (SH Built-in Functions): Move after RX Built-in Functions.
16276 * doc/gcc.texi (Introduction): Change order in @menu
16277 to match @node.
16278 * doc/md.texi (Constraints): Ditto.
16279 * gty.texi (Type Information): Ditto.
16280 (User-provided marking routines for template types): Make
16281 subsection.
16282 * doc/invoke.texi (AArch64 Options): Move before
16283 "Adapteva Epiphany Options".
16284
16285 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
16286 Jakub Jelinek <jakub@redhat.com>
16287
16288 PR sanitizer/56454
16289 * asan.c (gate_asan): Lookup no_sanitize_address instead of
16290 no_address_safety_analysis attribute.
16291 * doc/extend.texi (no_address_safety_attribute): Rename to
16292 no_sanitize_address attribute, mention no_address_safety_analysis
16293 attribute as deprecated alias.
16294
16295 2013-02-28 Jakub Jelinek <jakub@redhat.com>
16296
16297 PR middle-end/56461
16298 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
16299 type to vec<vec<tree> > *.
16300 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
16301 to be vec<tree> instead of vec<tree> *, set vec_defs
16302 to vNULL and call vec_defs.create (number_of_vects), adjust other
16303 uses of vec_defs.
16304 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
16305 vectorizable_condition): Adjust vect_get_slp_defs callers.
16306
16307 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16308
16309 * config/aarch64/aarch64.c
16310 (aarch64_float_const_representable): Remove unused variable.
16311
16312 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16313
16314 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
16315
16316 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16317
16318 * config/aarch64/aarch64-builtins.c
16319 (aarch64_init_simd_builtins): Make static.
16320
16321 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16322
16323 * config/aarch64/aarch64.c
16324 (aarch64_simd_make_constant): Make static.
16325
16326 2013-02-28 Martin Jambor <mjambor@suse.cz>
16327
16328 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
16329 with no initialization to the RHS of debug statements.
16330
16331 2013-02-28 Martin Jambor <mjambor@suse.cz>
16332
16333 PR tree-optimization/56294
16334 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
16335 Adjust dumping.
16336 (get_access_replacement): Do not call create_access_replacement.
16337 Assert a replacement exists.
16338 (get_repl_default_def_ssa_name): Create the replacement declaration
16339 itself.
16340
16341 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16342
16343 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
16344 final_end_function.
16345
16346 2013-02-28 Marek Polacek <polacek@redhat.com>
16347
16348 PR rtl-optimization/56466
16349 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
16350 if we're changing a loop.
16351 (peel_loops_completely): Likewise.
16352
16353 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
16354
16355 PR c++/55813
16356 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
16357
16358 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16359
16360 PR target/56445
16361 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
16362 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
16363 INTX_FTYPE_FX, FX_FTYPE_INTX.
16364 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
16365
16366 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16367
16368 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
16369 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
16370 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
16371 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
16372 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
16373 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
16374 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
16375 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
16376 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
16377 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
16378 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
16379 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
16380 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
16381 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
16382 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
16383 (avrxmega6): Increase max flash segments from 5 to 6.
16384 * config/avr/t-multilib: Regenerate.
16385 * config/avr/avr-tables.opt: Regenerate.
16386 * doc/avr-mmcu.texi: Regenerate.
16387
16388 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16389
16390 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
16391 (avr_device_to_arch): Rename to avr_device_to_ld.
16392 (avr_device_to_as): New prototype.
16393 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
16394 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
16395 * config/avr/driver-avr.c (avr_device_to_as): New.
16396 (avr_device_to_arch): Rename to avr_device_to_ld.
16397
16398 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16399
16400 PR middle-end/56461
16401 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
16402 method on dr_chain and result_chain.
16403
16404 PR middle-end/56461
16405 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
16406 pointer_set_destroy on not_executed_last_iteration.
16407
16408 PR middle-end/56461
16409 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
16410
16411 PR middle-end/56461
16412 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
16413 FOR_EACH_DEFINED_FUNCTION when freeing state.
16414
16415 PR middle-end/56461
16416 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
16417 pool_free.
16418 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
16419 overwriting it.
16420
16421 PR middle-end/56461
16422 * ipa-cp.c (decide_whether_version_node): Call vec_free on
16423 known_aggs[i].items and release known_aggs vector.
16424
16425 PR middle-end/56461
16426 * ipa-reference.c (propagate): Free node_info even for alias nodes.
16427
16428 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16429
16430 * config/microblaze/microblaze.c (microblaze_emit_compare):
16431 Use xor for EQ/NE comparisions.
16432 * config/microblaze/microblaze.md (cstoresf4): Add constraints
16433 (cbranchsf4): Adjust operator to comparison_operator.
16434
16435 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16436
16437 PR middle-end/56461
16438 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
16439 vector.
16440 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
16441 vec_safe_push, always update *slot.
16442 (redirect_edge_var_map_clear): Use vec_free.
16443 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
16444 (free_var_map_entry): Use vec_free.
16445 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
16446 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
16447
16448 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
16449
16450 PR middle-end/45472
16451 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
16452 when the may_trap_p bit of the exprs being merged differs.
16453 Reorder tests for speculativeness in the logical and operator.
16454
16455 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16456
16457 * incpath.c (add_standard_paths): Use reconcat instead of concat
16458 where appropriate and avoid leaking memory.
16459
16460 * opts.h: Include obstack.h.
16461 (opts_concat): New prototype.
16462 (opts_obstack): New declaration.
16463 * opts.c (opts_concat): New function.
16464 (opts_obstack): New variable.
16465 (init_options_struct): Call gcc_init_obstack on opts_obstack.
16466 (finish_options): Use opts_concat instead of concat
16467 and XOBNEWVEC instead of XNEWVEC.
16468 * opts-common.c (generate_canonical_option, decode_cmdline_option,
16469 generate_option): Likewise.
16470 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
16471 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
16472
16473 PR target/56455
16474 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
16475 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
16476
16477 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16478
16479 PR middle-end/56461
16480 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
16481
16482 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16483
16484 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
16485 (arm_block_move_unaligned_straight): Likewise.
16486 (arm_adjust_block_mem): Likewise.
16487
16488 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16489
16490 PR target/48901
16491 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
16492 temp, cond and label.
16493 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
16494
16495 PR target/52500
16496 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
16497 * config/c6x/c6x.h (dbx_register_map): Update declaration.
16498
16499 PR target/52501
16500 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
16501 of prologue/epilogue functions.
16502
16503 PR target/52550
16504 * config/tilegx/tilegx.c (tilegx_expand_prologue):
16505 Remove unused variable cfa_offset.
16506 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
16507
16508 PR target/54639
16509 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
16510 type promotion to unsigned.
16511
16512 PR target/54640
16513 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
16514 for HOST_WIDE_INT of 32 bit / same size as int.
16515 (arm_block_move_unaligned_straight): Likewise.
16516 (arm_adjust_block_mem): Likewise.
16517
16518 PR target/54662
16519 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
16520 ALL_CFLAGS.
16521
16522 2013-02-26 Marek Polacek <polacek@redhat.com>
16523
16524 PR tree-optimization/56426
16525 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
16526
16527 2013-02-26 Richard Biener <rguenther@suse.de>
16528
16529 PR target/56444
16530 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
16531 unused variable loops.
16532
16533 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16534
16535 PR tree-optimization/56448
16536 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
16537 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
16538 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
16539 later operands of the references, or even first operand for
16540 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
16541
16542 PR tree-optimization/56443
16543 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
16544 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
16545 to type_for_mode langhook.
16546
16547 2013-02-25 Matt Turner <mattst88@gmail.com>
16548
16549 * doc/invoke.texi: Document r4700.
16550
16551 2013-02-25 Richard Biener <rguenther@suse.de>
16552
16553 PR tree-optimization/56175
16554 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
16555 split out from ...
16556 (simplify_bitwise_binary): ... here. Also guard the conversion
16557 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
16558
16559 2013-02-25 Catherine Moore <clm@codesourcery.com>
16560
16561 Revert:
16562 2013-02-24 Catherine Moore <clm@codesourcery.com>
16563 Maciej W. Rozycki <macro@codesourcery.com>
16564 Tom de Vries <tom@codesourcery.com>
16565 Nathan Sidwell <nathan@codesourcery.com>
16566 Iain Sandoe <iain@codesourcery.com>
16567 Nathan Froyd <froydnj@codesourcery.com>
16568 Chao-ying Fu <fu@mips.com>
16569
16570 * doc/extend.texi: (micromips, nomicromips, nocompression):
16571 Document new function attributes.
16572 * doc/invoke.texi (minterlink-compressed, mmicromips,
16573 m14k, m14ke, m14kec): Document new options.
16574 (minterlink-mips16): Update documentation.
16575 * doc/md.texi (ZC, ZD): Document new constraints.
16576 * configure.ac (gcc_cv_as_micromips): Check if linker
16577 supports the .set micromips directive.
16578 * configure: Regenerate.
16579 * config.in: Regenerate.
16580 * config/mips/mips-tables.opt: Regenerate.
16581 * config/mips/micromips.md: New file.
16582 * constraints.md (ZC, AD): New constraints.
16583 * config/mips/predicates.md (movep_src_register): New predicate.
16584 (movep_src_operand): New predicate.
16585 (non_volatile_mem_operand): New predicate.
16586 * config/mips/mips.md (multimem): New type.
16587 (length): Differentiate between 17-bit and 18-bit branch offsets.
16588 (MOVEP1, MOVEP2): New mode iterator.
16589 (mov_<load>l): Use ZC constraint.
16590 (mov_<load>r): Likewise.
16591 (mov_<store>l): Likewise.
16592 (mov_<store>r): Likewise.
16593 (*branch_equality<mode>_inverted): Add microMIPS support.
16594 (*branch_equality<mode>): Likewise.
16595 (*jump_absolute): Likewise.
16596 (indirect_jump_<mode>): Likewise.
16597 (tablejump_<mode>): Likewise.
16598 (<optab>_internal): Likewise.
16599 (sibcall_internal): Likewise.
16600 (sibcall_value_internal): Likewise.
16601 (prefetch): Use constraint ZD.
16602 * config/mips/mips.opt (minterlink-compressed): New option.
16603 (minterlink-mips16): Now an alias for minterlink-compressed.
16604 (mmicromips): New option.
16605 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16606 (compare_and_swap_12): Likewise.
16607 (sync_add<mode>): Likewise.
16608 (sync_<optab>_12): Likewise.
16609 (sync_old_<optab>_12): Likewise.
16610 (sync_new_<optab>_12): Likewise.
16611 (sync_nand_12): Likewise.
16612 (sync_old_nand_12): Likewise.
16613 (sync_new_nand_12): Likewise.
16614 (sync_sub<mode>): Likewise.
16615 (sync_old_add<mode>): Likewise.
16616 (sync_old_sub<mode>): Likewise.
16617 (sync_new_add<mode>): Likewise.
16618 (sync_new_sub<mode>): Likewise.
16619 (sync_<optab><mode>): Likewise.
16620 (sync_old_<optab><mode>): Likewise.
16621 (sync_new_<optab><mode>): Likewise.
16622 (sync_nand<mode>): Likewise.
16623 (sync_old_nand<mode>): Likewise.
16624 (sync_new_nand<mode>): Likewise.
16625 (sync_lock_test_and_set<mode>): Likewise.
16626 (test_and_set_12): Likewise.
16627 (atomic_compare_and_swap<mode>): Likewise.
16628 (atomic_exchange<mode>_llsc): Likewise.
16629 (atomic_fetch_add<mode>_llsc): Likewise.
16630 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16631 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16632 (umips_save_restore_pattern_p): Likewise.
16633 (umips_load_store_pair_p): Likewise.
16634 (umips_output_load_store_pair): Likewise.
16635 (umips_movep_target_p): Likewise.
16636 (umips_12bit_offset_address_p): Likewise.
16637 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16638 (mips_base_mips16): Rename this...
16639 (mips_base_compression_flags): ...to this. Update all uses.
16640 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16641 (mips_mips16_decl_p): Delete.
16642 (mips_nomips16_decl_p): Delete.
16643 (mips_get_compress_on_flags): New function.
16644 (mips_get_compress_off_flags): New function.
16645 (mips_get_compress_mode): New function.
16646 (mips_get_compress_on_name): New function.
16647 (mips_get_compress_off_name): New function.
16648 (mips_insert_attributes): Support multiple compression types.
16649 (mips_merge_decl_attributes): Likewise.
16650 (umips_12bit_offset_address_p): New function.
16651 (mips_start_function_definition): Emit .set micromips directive.
16652 (mips_call_may_need_jalx_p): New function.
16653 (mips_function_ok_for_sibcall): Add microMIPS support.
16654 (mips_print_operand_punctuation): Support short delay slots and
16655 compact jumps.
16656 (umips_swm_mask, umips_swm_encoding): New.
16657 (umips_build_save_restore): New function.
16658 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16659 (was_mips16_p): Remove.
16660 (old_compression_mode): New.
16661 (mips_set_compression_mode): New function.
16662 (mips_set_current_function): Add microMIPS support.
16663 (mips_option_override): Likewise.
16664 (umips_save_restore_pattern_p): New function.
16665 (umips_output_save_restore): New function.
16666 (umips_load_store_pair_p_1): New function.
16667 (umips_load_store_pair_p): New function.
16668 (umips_output_load_store_pair_1): New function.
16669 (umips_output_load_store_pair): New function.
16670 (umips_movep_target_p) New function.
16671 (mips_prepare_pch_save): Add microMIPS support.
16672 * config/mips/mips.h (TARGET_COMPRESSION): New.
16673 (TARGET_CPU_CPP_BUILTINS): Update macro
16674 to use new compression flags and to support microMIPS.
16675 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16676 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16677 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16678 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16679 (ASM_SPEC): Support mmicromips and mno-micromips.
16680 (M16STORE_REG_P): New macro.
16681 (MIPS_CALL): Support TARGET_MICROMIPS.
16682 (MICROMIPS_J): New macro.
16683 (mips_base_mips16): Rename this...
16684 (mips_base_compression_flags): ...to this.
16685 (UMIPS_12BIT_OFFSET_P): New macro.
16686 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16687 (MULTILIB_DIRNAMES): Likewise.
16688
16689 2013-02-25 Tom de Vries <tom@codesourcery.com>
16690
16691 PR rtl-optimization/56131
16692 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
16693 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
16694 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
16695
16696 2013-02-25 Tobias Burnus <burnus@net-b.de>
16697
16698 * doc/invoke.texi (-fsanitize=): Move from optimization
16699 to debugging options.
16700
16701 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16702
16703 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
16704
16705 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16706 Alexander Monakov <amonakov@ispras.ru>
16707
16708 PR middle-end/56077
16709 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
16710 flush pending lists also on non-jumps. Adjust comment.
16711
16712 2013-02-24 Catherine Moore <clm@codesourcery.com>
16713 Maciej W. Rozycki <macro@codesourcery.com>
16714 Tom de Vries <tom@codesourcery.com>
16715 Nathan Sidwell <nathan@codesourcery.com>
16716 Iain Sandoe <iain@codesourcery.com>
16717 Nathan Froyd <froydnj@codesourcery.com>
16718 Chao-ying Fu <fu@mips.com>
16719
16720 * doc/extend.texi: (micromips, nomicromips, nocompression):
16721 Document new function attributes.
16722 * doc/invoke.texi (minterlink-compressed, mmicromips,
16723 m14k, m14ke, m14kec): Document new options.
16724 (minterlink-mips16): Update documentation.
16725 * doc/md.texi (ZC, ZD): Document new constraints.
16726 * configure.ac (gcc_cv_as_micromips): Check if linker
16727 supports the .set micromips directive.
16728 * configure: Regenerate.
16729 * config.in: Regenerate.
16730 * config/mips/mips-tables.opt: Regenerate.
16731 * config/mips/micromips.md: New file.
16732 * constraints.md (ZC, AD): New constraints.
16733 * config/mips/predicates.md (movep_src_register): New predicate.
16734 (movep_src_operand): New predicate.
16735 (non_volatile_mem_operand): New predicate.
16736 * config/mips/mips.md (multimem): New type.
16737 (length): Differentiate between 17-bit and 18-bit branch offsets.
16738 (MOVEP1, MOVEP2): New mode iterator.
16739 (mov_<load>l): Use ZC constraint.
16740 (mov_<load>r): Likewise.
16741 (mov_<store>l): Likewise.
16742 (mov_<store>r): Likewise.
16743 (*branch_equality<mode>_inverted): Add microMIPS support.
16744 (*branch_equality<mode>): Likewise.
16745 (*jump_absolute): Likewise.
16746 (indirect_jump_<mode>): Likewise.
16747 (tablejump_<mode>): Likewise.
16748 (<optab>_internal): Likewise.
16749 (sibcall_internal): Likewise.
16750 (sibcall_value_internal): Likewise.
16751 (prefetch): Use constraint ZD.
16752 * config/mips/mips.opt (minterlink-compressed): New option.
16753 (minterlink-mips16): Now an alias for minterlink-compressed.
16754 (mmicromips): New option.
16755 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16756 (compare_and_swap_12): Likewise.
16757 (sync_add<mode>): Likewise.
16758 (sync_<optab>_12): Likewise.
16759 (sync_old_<optab>_12): Likewise.
16760 (sync_new_<optab>_12): Likewise.
16761 (sync_nand_12): Likewise.
16762 (sync_old_nand_12): Likewise.
16763 (sync_new_nand_12): Likewise.
16764 (sync_sub<mode>): Likewise.
16765 (sync_old_add<mode>): Likewise.
16766 (sync_old_sub<mode>): Likewise.
16767 (sync_new_add<mode>): Likewise.
16768 (sync_new_sub<mode>): Likewise.
16769 (sync_<optab><mode>): Likewise.
16770 (sync_old_<optab><mode>): Likewise.
16771 (sync_new_<optab><mode>): Likewise.
16772 (sync_nand<mode>): Likewise.
16773 (sync_old_nand<mode>): Likewise.
16774 (sync_new_nand<mode>): Likewise.
16775 (sync_lock_test_and_set<mode>): Likewise.
16776 (test_and_set_12): Likewise.
16777 (atomic_compare_and_swap<mode>): Likewise.
16778 (atomic_exchange<mode>_llsc): Likewise.
16779 (atomic_fetch_add<mode>_llsc): Likewise.
16780 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16781 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16782 (umips_save_restore_pattern_p): Likewise.
16783 (umips_load_store_pair_p): Likewise.
16784 (umips_output_load_store_pair): Likewise.
16785 (umips_movep_target_p): Likewise.
16786 (umips_12bit_offset_address_p): Likewise.
16787 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16788 (mips_base_mips16): Rename this...
16789 (mips_base_compression_flags): ...to this. Update all uses.
16790 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16791 (mips_mips16_decl_p): Delete.
16792 (mips_nomips16_decl_p): Delete.
16793 (mips_get_compress_on_flags): New function.
16794 (mips_get_compress_off_flags): New function.
16795 (mips_get_compress_mode): New function.
16796 (mips_get_compress_on_name): New function.
16797 (mips_get_compress_off_name): New function.
16798 (mips_insert_attributes): Support multiple compression types.
16799 (mips_merge_decl_attributes): Likewise.
16800 (umips_12bit_offset_address_p): New function.
16801 (mips_start_function_definition): Emit .set micromips directive.
16802 (mips_call_may_need_jalx_p): New function.
16803 (mips_function_ok_for_sibcall): Add microMIPS support.
16804 (mips_print_operand_punctuation): Support short delay slots and
16805 compact jumps.
16806 (umips_swm_mask, umips_swm_encoding): New.
16807 (umips_build_save_restore): New function.
16808 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16809 (was_mips16_p): Remove.
16810 (old_compression_mode): New.
16811 (mips_set_compression_mode): New function.
16812 (mips_set_current_function): Add microMIPS support.
16813 (mips_option_override): Likewise.
16814 (umips_save_restore_pattern_p): New function.
16815 (umips_output_save_restore): New function.
16816 (umips_load_store_pair_p_1): New function.
16817 (umips_load_store_pair_p): New function.
16818 (umips_output_load_store_pair_1): New function.
16819 (umips_output_load_store_pair): New function.
16820 (umips_movep_target_p) New function.
16821 (mips_prepare_pch_save): Add microMIPS support.
16822 * config/mips/mips.h (TARGET_COMPRESSION): New.
16823 (TARGET_CPU_CPP_BUILTINS): Update macro
16824 to use new compression flags and to support microMIPS.
16825 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16826 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16827 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16828 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16829 (ASM_SPEC): Support mmicromips and mno-micromips.
16830 (M16STORE_REG_P): New macro.
16831 (MIPS_CALL): Support TARGET_MICROMIPS.
16832 (MICROMIPS_J): New macro.
16833 (mips_base_mips16): Rename this...
16834 (mips_base_compression_flags): ...to this.
16835 (UMIPS_12BIT_OFFSET_P): New macro.
16836 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16837 (MULTILIB_DIRNAMES): Likewise.
16838
16839 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16840
16841 PR target/52555
16842 * target-globals.c (save_target_globals): For init_reg_sets and
16843 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16844
16845 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16846
16847 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16848 * config/aarch64/t-aarch64
16849 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16850
16851 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16852
16853 PR inline-asm/56148
16854 * lra-constraints.c (process_alt_operands): Reload operand
16855 conflicting with earlier clobber only if no more other conflicting
16856 operands.
16857
16858 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16859
16860 PR sanitizer/56393
16861 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16862 if not linking a shared library.
16863
16864 2013-02-22 Seth LaForge <sethml@google.com>
16865
16866 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16867
16868 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16869
16870 * config/arm/arm.md (split for extendsidi): Update condition.
16871 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16872 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16873 (qhs_zextenddi_cstr): Likewise.
16874
16875 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16876
16877 PR middle-end/56420
16878 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16879 avoid signed wrapping.
16880 (expand_mult): Handle properly multiplication by
16881 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16882 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16883 in the compiler if coeff is HOST_WIDE_INT_MIN.
16884 (expand_divmod): Don't make ext_op1 static, change it's type to
16885 uhwi. Avoid undefined behavior in -INTVAL (op1).
16886
16887 PR rtl-optimization/50339
16888 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16889 field.
16890 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16891 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16892 into splitting_ashiftrt field.
16893 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16894 ASHIFTRT.
16895 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16896 choices.
16897
16898 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16899
16900 PR middle-end/56108
16901 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16902 are sure to go irrevocable.
16903
16904 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16905
16906 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16907 scalars are valid operands.
16908
16909 2013-02-21 Martin Jambor <mjambor@suse.cz>
16910
16911 PR tree-optimization/56310
16912 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16913 only matching indices and non-negative final offsets.
16914 (intersect_aggregates_with_edge): Pass src_idx to
16915 agg_replacements_to_vector. Pass src_idx insstead of index to
16916 intersect_with_agg_replacements.
16917
16918 2013-02-21 Martin Jambor <mjambor@suse.cz>
16919
16920 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16921 instead of hard-wired defaults.
16922
16923 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16924
16925 * doc/invoke.texi (MIPS Options): Update documentation of the
16926 floating-point multiply-accumulate instruction restrictions.
16927
16928 2013-02-21 Kostya Serebryany <kcc@google.com>
16929
16930 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16931 asan_shadow_offset on x86_64 linux.
16932
16933 2013-02-21 Richard Biener <rguenther@suse.de>
16934
16935 PR tree-optimization/56415
16936 Revert
16937 2013-02-11 Richard Biener <rguenther@suse.de>
16938
16939 PR tree-optimization/56273
16940 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16941 first VRP run.
16942
16943 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16944
16945 PR bootstrap/56258
16946 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16947 instead of @itemx.
16948
16949 PR inline-asm/56405
16950 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16951 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16952
16953 2013-02-20 Jan Hubicka <jh@suse.cz>
16954
16955 PR tree-optimization/56265
16956 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16957 when target is referenced for first time.
16958
16959 2013-02-20 Richard Biener <rguenther@suse.de>
16960
16961 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16962 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16963 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16964 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16965 not return anything.
16966 (rename_ssa_copies): Do not remove unused locals.
16967 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16968 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16969 * passes.c (execute_function_todo): Do not schedule unused locals
16970 removal if cleanup_tree_cfg did something.
16971 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16972 about the number of removed locals.
16973
16974 2013-02-20 Richard Biener <rguenther@suse.de>
16975
16976 PR tree-optimization/56398
16977 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16978
16979 2013-02-20 Martin Jambor <mjambor@suse.cz>
16980
16981 PR tree-optimization/55334
16982 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16983 restricted pointers to arrays.
16984
16985 2013-02-20 Richard Biener <rguenther@suse.de>
16986 Jakub Jelinek <jakub@redhat.com>
16987
16988 PR tree-optimization/56396
16989 * tree-ssa-ccp.c (n_const_val): New static variable.
16990 (get_value): Return NULL for SSA names we don't have a lattice
16991 entry for.
16992 (ccp_initialize): Initialize n_const_val.
16993 * tree-ssa-copy.c (n_copy_of): New static variable.
16994 (init_copy_prop): Initialize n_copy_of.
16995 (get_value): Return NULL_TREE for SSA names we don't have a
16996 lattice entry for.
16997
16998 2013-02-20 Martin Jambor <mjambor@suse.cz>
16999
17000 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
17001
17002 2013-02-20 Richard Biener <rguenther@suse.de>
17003
17004 * genpreds.c (write_lookup_constraint): Do not compare first
17005 letter of the constraint again.
17006
17007 2013-02-20 Richard Biener <rguenther@suse.de>
17008
17009 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
17010 and ceil_log2.
17011 (get_use_iv_cost): Terminate hashtable walk when coming across
17012 an empty entry.
17013
17014 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
17015
17016 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
17017 reassociation for avx2 targets.
17018
17019 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17020
17021 * config/microblaze/microblaze.c: microblaze_has_clz = 0
17022 Add version check for v8.10.a to enable microblaze_has_clz
17023 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
17024 version and TARGET_PATTERN_COMPARE check
17025 * config/microblaze/microblaze.md: New clzsi2 instruction
17026
17027 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17028
17029 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17030 function before branching.
17031
17032 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17033
17034 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17035 DUMP_INSN_RTX_UID.
17036 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17037
17038 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17039
17040 PR middle-end/55889
17041 * sel-sched.c: Include ira.h.
17042 (implicit_clobber_conflict_p): New function.
17043 (moveup_expr): Use it.
17044 * Makefile.in (sel-sched.o): Depend on ira.h.
17045
17046 2013-02-19 Richard Biener <rguenther@suse.de>
17047
17048 PR tree-optimization/56384
17049 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17050 (vn_hash_type): Split out from ...
17051 (vn_hash_constant_with_type): ... here.
17052 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17053 (vn_phi_eq): Compare types from vn_phi_s structure.
17054 (vn_phi_lookup): Populate vn_phi_s type.
17055 (vn_phi_insert): Likewise.
17056
17057 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17058
17059 PR tree-optimization/56350
17060 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17061 if haven't found reduction or nested cycle operand, rather than
17062 asserting we must find it.
17063
17064 PR tree-optimization/56381
17065 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17066 to fold_build3.
17067
17068 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17069 Jakub Jelinek <jakub@redhat.com>
17070
17071 PR target/52555
17072 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17073 (swap_optab_enable): Same.
17074 (init_all_optabs): Use argument instead of global.
17075 * tree.h (struct tree_optimization_option): New field target_optabs.
17076 * expr.h (init_all_optabs): Add argument to prototype.
17077 (TREE_OPTIMIZATION_OPTABS): New.
17078 (save_optabs_if_changed): Protoize.
17079 * optabs.h: Declare this_fn_optabs.
17080 * optabs.c (save_optabs_if_changed): New.
17081 Declare this_fn_optabs.
17082 (init_optabs): Add argument to init_all_optabs() call.
17083 * function.c (invoke_set_current_function_hook): Handle per
17084 function optabs.
17085 * function.h (struct function): New field optabs.
17086 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17087 optimization_current_node has changed.
17088 * target-globals.h (save_target_globals_default_opts): Protoize.
17089 * target-globals.c (save_target_globals_default_opts): New.
17090
17091 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17092
17093 PR target/56347
17094 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17095 registers %fr12 and %fr12R as call used.
17096
17097 PR target/56214
17098 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17099 and HImode, require all displacements to be an integer multiple of
17100 their mode size.
17101 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17102 only allow QImode and HImode when reload is in progress and strict is
17103 true. Likewise for symbolic addresses. Use base14_operand to check
17104 displacements in REG+BASE addresses.
17105
17106 2013-02-18 Richard Biener <rguenther@suse.de>
17107
17108 PR tree-optimization/56366
17109 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17110 handle sign-conversion of outer-loop initial induction value.
17111
17112 2013-02-18 Richard Biener <rguenther@suse.de>
17113
17114 PR middle-end/56349
17115 * cfghooks.c (merge_blocks): If we merge a latch into another
17116 block adjust references to it.
17117 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17118 (verify_loop_structure): Verify that a recorded latch is in fact
17119 a latch.
17120
17121 2013-02-18 Richard Biener <rguenther@suse.de>
17122
17123 PR tree-optimization/56321
17124 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17125 order SSA name release and virtual operand unlinking.
17126
17127 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17128
17129 * config/microblaze/microblaze.md (save_stack_block): Define.
17130 (restore_stack_block): Likewise.
17131
17132 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17133
17134 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17135 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17136 * config/microblaze/microblaze.c (microblaze_option_override):
17137 Bail out early for PIC modes when target does not support PIC.
17138
17139 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17140
17141 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17142 Replace with a microblaze version.
17143 (microblaze_trampoline_init): Adapt for microblaze.
17144 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17145 microblaze.
17146
17147 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17148 Dodji Seketeli <dodji@redhat.com>
17149
17150 PR asan/56330
17151 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17152 (instrument_mem_region_access): Do not forget to always put
17153 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17154 0) statement, even for cases where either 'base' or 'base + len'
17155 are not instrumented -- because they have been previously
17156 instrumented. Simplify the logic by putting all the statements
17157 instrument 'base + len' inside a sequence, and then insert that
17158 sequence right before the current insertion point. Then, to
17159 instrument 'base + len', just get an iterator on that statement.
17160 And do not forget to update the pointer to iterator the function
17161 received as argument.
17162
17163 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17164
17165 PR rtl-optimization/56348
17166 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17167
17168 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17169
17170 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17171 (clean_graph_dump_file): Pass base to start_graph_dump.
17172
17173 2013-02-14 Richard Henderson <rth@redhat.com>
17174
17175 PR target/55941
17176 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17177
17178 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17179
17180 * collect2-aix.h: Define F_LOADONLY.
17181
17182 2013-02-14 Richard Biener <rguenther@suse.de>
17183
17184 PR lto/50494
17185 * varasm.c (output_constant_def_1): Get the decl representing
17186 the constant as argument.
17187 (output_constant_def): Wrap output_constant_def_1.
17188 (make_decl_rtl): Use output_constant_def_1 with the decl
17189 representing the constant.
17190 (build_constant_desc): Optionally re-use a decl already
17191 representing the constant.
17192 (tree_output_constant_def): Adjust.
17193
17194 2013-02-14 Dodji Seketeli <dodji@redhat.com>
17195
17196 Fix an asan crash
17197 * asan.c (instrument_builtin_call): Really put the length of the
17198 second source argument into src1_len.
17199
17200 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17201
17202 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
17203 argument. If it is false, don't create edge from then_bb to
17204 fallthru_bb.
17205 (insert_if_then_before_iter): Pass true to it.
17206 (build_check_stmt): Pass false to it.
17207 (transform_statements): Flush hash table only on extended basic
17208 block boundaries, rather than at the beginning of every bb.
17209 Don't flush hash table on nonfreeing_call_p calls.
17210 * tree-flow.h (nonfreeing_call_p): New prototype.
17211 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
17212
17213 2013-02-13 David S. Miller <davem@davemloft.net>
17214
17215 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
17216
17217 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
17218
17219 PR target/56184
17220 * ira.c (max_regno_before_ira): Move from ...
17221 (ira): ... here.
17222 (fix_reg_equiv_init): Use max_regno_before_ira instead of
17223 vec_safe_length.
17224
17225 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17226
17227 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
17228
17229 2013-02-13 Richard Biener <rguenther@suse.de>
17230
17231 PR lto/56295
17232 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
17233 globals in MEM_REFs.
17234
17235 2013-02-13 Richard Biener <rguenther@suse.de>
17236
17237 * loop-init.c (loop_optimizer_init): Clear loop state when
17238 re-initializing preserved loops.
17239 * loop-unswitch.c (unswitch_single_loop): Return whether
17240 we unswitched the loop. Do not verify loop state here.
17241 (unswitch_loops): When we unswitched a loop discover new loops.
17242
17243 2013-02-13 Kostya Serebryany <kcc@google.com>
17244
17245 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
17246 on x86_64 linux.
17247 * sanitizer.def: Rename __asan_init to __asan_init_v1.
17248
17249 2013-02-12 Dodji Seketeli <dodji@redhat.com>
17250
17251 Avoid instrumenting duplicated memory access in the same basic block
17252 * Makefile.in (asan.o): Add new dependency on hash-table.h
17253 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
17254 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
17255 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
17256 (free_mem_ref_resources, has_mem_ref_been_instrumented)
17257 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
17258 (get_mem_ref_of_assignment): New functions.
17259 (get_mem_refs_of_builtin_call): Extract from
17260 instrument_builtin_call and tweak a little bit to make it fit with
17261 the new signature.
17262 (instrument_builtin_call): Use the new
17263 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
17264 of is_gimple_builtin_call.
17265 (instrument_derefs, instrument_mem_region_access): Insert the
17266 instrumented memory reference into the hash table.
17267 (maybe_instrument_assignment): Renamed instrument_assignment into
17268 this, and change it to advance the iterator when instrumentation
17269 actually happened and return true in that case. This makes it
17270 homogeneous with maybe_instrument_assignment, and thus give a
17271 chance to callers to be more 'regular'.
17272 (transform_statements): Clear the memory reference hash table
17273 whenever we enter a new BB, when we cross a function call, or when
17274 we are done transforming statements. Use
17275 maybe_instrument_assignment instead of instrumentation. No more
17276 need to special case maybe_instrument_assignment and advance the
17277 iterator after calling it; it's now handled just like
17278 maybe_instrument_call. Update comment.
17279
17280 2013-02-13 Richard Biener <rguenther@suse.de>
17281
17282 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
17283 Fix loop discovery code.
17284
17285 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
17286
17287 PR inline-asm/56148
17288 * lra-constraints.c (process_alt_operands): Match early clobber
17289 operand with itself. Check conflicts with earlyclobber only if
17290 the operand is not reloaded. Prefer to reload conflicting operand
17291 if earlyclobber and matching operands are the same.
17292
17293 2013-02-12 Richard Biener <rguenther@suse.de>
17294
17295 PR lto/56297
17296 * lto-streamer-out.c (write_symbol): Do not output symbols
17297 for hard register variables.
17298
17299 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
17300
17301 PR target/54222
17302 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
17303 (umulsidi3_insn, mulsidi3_insn): New insns.
17304
17305 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
17306
17307 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
17308 (struct tune_params): Add vec_costs field.
17309 * config/arm/arm.c (arm_builtin_vectorization_cost)
17310 (arm_add_stmt_cost): New functions.
17311 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
17312 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
17313 (arm_default_vec_cost): New struct of type cpu_vec_costs.
17314 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
17315 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
17316 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
17317 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
17318
17319 2013-02-12 Richard Biener <rguenther@suse.de>
17320
17321 PR lto/56295
17322 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
17323 decls again if possible.
17324
17325 2013-02-12 Richard Biener <rguenther@suse.de>
17326
17327 PR middle-end/56288
17328 * tree-ssa.c (verify_ssa_name): Fix check, move
17329 SSA_NAME_IN_FREE_LIST check up.
17330
17331 2013-02-12 Jakub Jelinek <jakub@redhat.com>
17332 Steven Bosscher <steven@gcc.gnu.org>
17333
17334 PR rtl-optimization/56151
17335 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
17336 equal to op0 or op1, and last_insn pattern is CODE operation
17337 with MEM dest and one of the operands matches that MEM.
17338
17339 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
17340
17341 * doc/extend.texi: Document Function Multiversioning and "default"
17342 parameter string to target attribute.
17343 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
17344 target attribute parameter is "default".
17345 (ix86_compare_version_priority): Remove checks for target attribute.
17346 (ix86_mangle_function_version_assembler_name): Change error to sorry.
17347 Remove check for target attribute equal to NULL. Add assert.
17348 (ix86_generate_version_dispatcher_body): Change error to sorry.
17349
17350 2013-02-11 Iain Sandoe <iain@codesourcery.com>
17351 Jack Howarth <howarth@bromo.med.uc.edu>
17352 Patrick Marlier <patrick.marlier@gmail.com>
17353
17354 PR libitm/55693
17355 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
17356 define ENDFILE_SPEC as TM_DESTRUCTOR.
17357 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
17358
17359 2013-02-11 Alexander Potapenko <glider@google.com>
17360 Jack Howarth <howarth@bromo.med.uc.edu>
17361 Jakub Jelinek <jakub@redhat.com>
17362
17363 PR sanitizer/55617
17364 * config/darwin.c (cdtor_record): Rename ctor_record.
17365 (sort_cdtor_records): Rename sort_ctor_records.
17366 (finalize_dtors): New routine to sort destructors by
17367 priority before use in assemble_integer.
17368 (machopic_asm_out_destructor): Use finalize_dtors if needed.
17369
17370 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
17371
17372 PR rtl-optimization/56275
17373 * simplify-rtx.c (avoid_constant_pool_reference): Check that
17374 offset is non-negative and less than cmode size before
17375 calling simplify_subreg.
17376
17377 2013-02-11 Richard Biener <rguenther@suse.de>
17378
17379 PR tree-optimization/56264
17380 * cfgloop.h (fix_loop_structure): Adjust prototype.
17381 * loop-init.c (fix_loop_structure): Return the number of
17382 newly discovered loops.
17383 * tree-cfgcleanup.c (repair_loop_structures): When new loops
17384 are discovered, do a full loop-closed SSA rewrite.
17385
17386 2013-02-11 Richard Biener <rguenther@suse.de>
17387
17388 PR tree-optimization/56273
17389 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17390 first VRP run.
17391 (check_array_ref): Fix missing newline in dumps.
17392 (search_for_addr_array): Likewise.
17393
17394 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
17395
17396 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
17397
17398 2013-02-09 Jakub Jelinek <jakub@redhat.com>
17399
17400 PR target/56256
17401 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
17402
17403 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
17404
17405 PR rtl-optimization/56246
17406 * lra-constraints.c (simplify_operand_subreg): Try to reuse
17407 reload pseudo.
17408 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
17409 constraints are satisfied.
17410
17411 2013-02-08 Jeff Law <law@redhat.com>
17412
17413 PR debug/53948
17414 * emit-rtl.c (reg_is_parm_p): New function.
17415 * regs.h (reg_is_parm_p): New prototype.
17416 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
17417 callee-clobbered registers.
17418
17419 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
17420
17421 PR target/56043
17422 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
17423 If there is no implicit builtin declaration, just return NULL.
17424
17425 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17426
17427 * config/i386/sse.md (FMAMODEM): New mode iterator.
17428 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
17429 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
17430
17431 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17432
17433 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
17434 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
17435 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
17436
17437 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17438
17439 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
17440 (microblaze*-*-elf): Likewise.
17441 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
17442 LINK_SPEC.
17443 * config/microblaze/microblaze-c.c: Add builtin defines for
17444 _LITTLE_ENDIAN and _BIG_ENDIAN.
17445 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
17446 add to TARGET_DEFAULT flags.
17447 Expand ASM_SPEC and LINK_SPEC.
17448 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
17449 * config/microblaze/microblaze.md: Update extendsidi2 and
17450 movdi_internal instructions to use low-order / high-order reg
17451 print_operands.
17452 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
17453 options and inversemask / mask of LITTLE_ENDIAN.
17454 * config/microblaze/t-microblaze: Expand multilib options to
17455 include mlittle-endian (le) and update exceptions patterns.
17456
17457 2013-02-08 Jakub Jelinek <jakub@redhat.com>
17458
17459 PR rtl-optimization/56195
17460 * lra-constraints.c (get_reload_reg): Don't reuse regs
17461 if they have smaller mode than requested, if they have
17462 wider mode than requested, try to return a SUBREG.
17463
17464 PR tree-optimization/56250
17465 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
17466 if type is unsigned and code isn't MULT_EXPR.
17467
17468 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17469
17470 PR tree-optimization/56064
17471 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
17472 bits according to mode.
17473 * fixed-value.h (fixed_from_double_int)
17474 (const_fixed_from_double_int): Adjust comments.
17475
17476 2013-02-08 Richard Biener <rguenther@suse.de>
17477
17478 PR lto/56231
17479 * lto-streamer.h (struct data_in): Remove current_file, current_line
17480 and current_col members.
17481 * lto-streamer-out.c (lto_output_location): Stream changed bits
17482 en-block for efficiency.
17483 * lto-streamer-in.c (clear_line_info): Remove.
17484 (lto_input_location): Cache current file, line and column
17485 globally via local statics. Read changed bits en-block.
17486 (input_function): Do not call clear_line_info.
17487 (lto_read_body): Likewise.
17488 (lto_input_toplevel_asms): Likewise.
17489
17490 2013-02-08 Michael Matz <matz@suse.de>
17491
17492 PR tree-optimization/52448
17493 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
17494 (nt_call_phase): New static.
17495 (add_or_mark_expr): Only mark accesses with newer phase than any
17496 call seen.
17497 (nonfreeing_call_p): New.
17498 (nt_init_block): Update nt_call_phase, mark blocks as visited.
17499 (nt_fini_block): Keep blocks marked as visited.
17500 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
17501
17502 2013-02-08 Richard Biener <rguenther@suse.de>
17503
17504 * ira.c (ira): Free broken dominator information.
17505
17506 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17507
17508 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
17509
17510 2013-02-08 Marek Polacek <polacek@redhat.com>
17511
17512 * cfgloop.c (verify_loop_structure): Add more checking of headers.
17513
17514 2013-02-08 Richard Biener <rguenther@suse.de>
17515
17516 PR middle-end/56181
17517 * cfgloop.h (flow_loops_find): Adjust.
17518 (bb_loop_header_p): Declare.
17519 * cfgloop.c (bb_loop_header_p): New function split out from ...
17520 (flow_loops_find): ... here. Adjust function signature,
17521 support incremental loop structure update.
17522 (verify_loop_structure): Cleanup. Verify a loop is a loop.
17523 * cfgloopmanip.c (fix_loop_structure): Move ...
17524 * loop-init.c (fix_loop_structure): ... here.
17525 (apply_loop_flags): Split out from ...
17526 (loop_optimizer_init): ... here.
17527 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
17528 in incremental mode, only remove dead loops here.
17529
17530 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17531
17532 PR target/54222
17533 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
17534 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
17535 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
17536 (*round<mode>3.libgcc): New insns for fixed-modes.
17537 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
17538 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
17539 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
17540 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
17541 implementations. Define to __builtin_avr_absFX,
17542 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
17543 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
17544 __builtin_avr_countlsFX, respectively.
17545 * config/avr/avr-c.c (target.h): Include it.
17546 (enum avr_builtin_id): New enum.
17547 (avr_resolve_overloaded_builtin): New static function.
17548 (avr_register_target_pragmas): Use it to set
17549 targetm.resolve_overloaded_builtin.
17550 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
17551 tree nodes used by DEF_BUILTIN.
17552 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
17553 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
17554 <AVR_BUILTIN_xxBITS>: Same.
17555
17556 2013-02-08 Richard Biener <rguenther@suse.de>
17557
17558 * cfgloop.c (verify_loop_structure): Properly handle
17559 a loop exiting to another loop header.
17560 * ira-int.h (ira_loops): Remove.
17561 * ira.c (ira_loops): Remove.
17562 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
17563 (do_reload): Use loop_optimizer_finalize.
17564 * ira-build.c (create_loop_tree_nodes): Use get_loops and
17565 number_of_loops to access the loop tree.
17566 (more_one_region_p): Likewise.
17567 (finish_loop_tree_nodes): Likewise.
17568 (rebuild_regno_allocno_maps): Likewise.
17569 (mark_loops_for_removal): Likewise.
17570 (mark_all_loops_for_removal): Likewise.
17571 (remove_unnecessary_regions): Likewise.
17572 (ira_build): Likewise.
17573 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17574
17575 2013-02-08 Richard Biener <rguenther@suse.de>
17576
17577 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
17578 * ipa-pure-const.c (analyze_function): Avoid calling
17579 mark_irreducible_loops twice.
17580 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
17581
17582 2013-02-07 David S. Miller <davem@davemloft.net>
17583
17584 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
17585 on 'reg'.
17586 * var-tracking.c (vt_add_function_parameter): Test the presence of
17587 HAVE_window_save properly and do not remap argument registers when
17588 we have a leaf function.
17589
17590 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
17591
17592 PR bootstrap/56227
17593 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
17594 instead of "ll".
17595 * config/i386/i386.c (ix86_print_operand): Ditto.
17596
17597 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17598
17599 * lra-constraints.c (process_alt_operands): Fix recently added comment.
17600
17601 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17602
17603 PR rtl-optimization/56225
17604 * lra-constraints.c (process_alt_operands): Check that reload hard
17605 reg can hold value for strict_low_part.
17606
17607 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17608
17609 PR debug/56154
17610 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
17611 dwarf2out_end_function.
17612 (in_first_function_p, maybe_at_text_label_p,
17613 first_loclabel_num_not_at_text_label): New variables.
17614 (dwarf2out_var_location): In the first function find out
17615 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
17616 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
17617 functions.
17618
17619 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
17620
17621 PR rtl-optimization/56178
17622 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
17623 SUBREG of a register. Tidy up related block of code.
17624 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
17625 note if the source is a register or a SUBREG of a register.
17626
17627 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17628
17629 PR target/56228
17630 * config/rs6000/rs6000.md (ptrm): New mode attr.
17631 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
17632 call_value_indirect_aix<pttrsize>,
17633 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
17634 m in constraints.
17635
17636 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17637
17638 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
17639 if -bnortl. Convert to strcmp and strncmp.
17640
17641 2013-02-07 Alan Modra <amodra@gmail.com>
17642
17643 PR target/54009
17644 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
17645 addresses won't wrap when offsetting.
17646 (rs6000_secondary_reload): Provide secondary reloads needed for
17647 wrapping LO_SUM addresses.
17648
17649 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
17650
17651 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
17652 MACH, just __MACH__.
17653
17654 2013-02-06 Richard Biener <rguenther@suse.de>
17655
17656 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
17657 instead of calling fix_loop_structure.
17658
17659 2013-02-06 Jakub Jelinek <jakub@redhat.com>
17660
17661 PR middle-end/56217
17662 * omp-low.c (use_pointer_for_field): Return false if
17663 lower_send_shared_vars doesn't generate any copy-out code.
17664
17665 2013-02-06 Tom de Vries <tom@codesourcery.com>
17666
17667 PR rtl-optimization/56131
17668 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
17669 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
17670 of the label is NULL. Add comment.
17671
17672 2013-02-05 Jakub Jelinek <jakub@redhat.com>
17673
17674 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
17675
17676 PR sanitizer/55374
17677 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
17678 (STATIC_LIBTSAN_LIBS): Likewise.
17679 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
17680 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
17681 is defined, don't add anything else beyond that.
17682 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
17683 (LINK_COMMAND_SPEC): Use them.
17684
17685 PR tree-optimization/56205
17686 * tree-stdarg.c (check_all_va_list_escapes): Return true if
17687 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
17688 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
17689
17690 2013-02-05 Richard Biener <rguenther@suse.de>
17691
17692 PR tree-optimization/53342
17693 PR tree-optimization/53185
17694 * tree-vectorizer.h (vect_check_strided_load): Remove.
17695 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
17696 not disallow peeling for vectorized strided loads.
17697 (vect_check_strided_load): Make static and simplify.
17698 (vect_analyze_data_refs): Adjust.
17699 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
17700 correctly when vectorizing strided loads.
17701
17702 2013-02-05 Richard Biener <rguenther@suse.de>
17703
17704 * doc/install.texi: Refer to ISL, not PPL.
17705
17706 2013-02-05 Jan Hubicka <jh@suse.cz>
17707
17708 PR tree-optimization/55789
17709 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
17710
17711 2013-02-05 Jan Hubicka <jh@suse.cz>
17712
17713 PR tree-optimization/55789
17714 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
17715 the dead call anyway.
17716
17717 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
17718
17719 PR sanitizer/55374
17720 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
17721
17722 2013-02-04 Alexander Potapenko <glider@google.com>
17723 Jack Howarth <howarth@bromo.med.uc.edu>
17724 Jakub Jelinek <jakub@redhat.com>
17725
17726 PR sanitizer/55617
17727 * config/darwin.c (sort_ctor_records): Stabilized qsort
17728 on constructor priority by using original position.
17729 (finalize_ctors): New routine to sort constructors by
17730 priority before use in assemble_integer.
17731 (machopic_asm_out_constructor): Use finalize_ctors if needed.
17732
17733 2013-02-04 Jakub Jelinek <jakub@redhat.com>
17734
17735 PR libstdc++/54314
17736 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
17737 about visibility on artificial decls.
17738 * config/sol2.c (solaris_assemble_visibility): Likewise.
17739
17740 2013-02-04 Kai Tietz <ktietz@redhat.com>
17741
17742 PR target/56186
17743 * config/i386/i386.c (function_value_ms_64): Add additional valtype
17744 argument and improve checking of return-argument types for 16-byte
17745 modes.
17746 (ix86_function_value_1): Add additional valtype argument on call
17747 of function_value_64.
17748 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
17749 handling infunction_value_64 function.
17750
17751 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
17752
17753 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
17754
17755 2013-02-04 Richard Biener <rguenther@suse.de>
17756
17757 PR tree-optimization/56188
17758 * tree-ssa-structalias.c (label_visit): Consider case with
17759 initially non-empty points-to set.
17760 (perform_var_substitution): Dump node mapping and clean up.
17761
17762 2013-02-04 Richard Guenther <rguenther@suse.de>
17763
17764 PR lto/56168
17765 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
17766 node prevail as last resort.
17767 (lto_symtab_merge_decls): Remove guard on LTRANS here.
17768 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
17769
17770 2013-02-04 Richard Biener <rguenther@suse.de>
17771
17772 PR tree-optimization/56113
17773 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
17774 Merge into ...
17775 (equiv_class_lookup_or_add): ... this.
17776 (label_visit): Adjust and fix error in previous patch.
17777 (perform_var_substitution): Adjust.
17778
17779 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
17780
17781 * config/sh/divtab.c: Fix formatting and comments throughout the file.
17782 * config/sh/sh4-300.md: Likewise.
17783 * config/sh/sh4a.md: Likewise.
17784 * config/sh/constraints.md: Likewise.
17785 * config/sh/sh.md: Likewise.
17786 * config/sh/netbsd-elf.h: Likewise.
17787 * config/sh/predicates.md: Likewise.
17788 * config/sh/sh-protos.h: Likewise.
17789 * config/sh/ushmedia.h: Likewise.
17790 * config/sh/linux.h: Likewise.
17791 * config/sh/sh.c: Likewise.
17792 * config/sh/superh.h: Likewise.
17793 * config/sh/elf.h: Likewise.
17794 * config/sh/sh4.md: Likewise.
17795 * config/sh/sh.h: Likewise.
17796
17797 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17798
17799 * config/pa/constraints.md: Adjust unused letters. Change "T"
17800 constraint to match_test floating_point_store_memory_operand().
17801 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17802 (base14_operand): New.
17803 (floating_point_store_memory_operand): New.
17804 (integer_store_memory_operand): Revise to use base14_operand and
17805 reg_plus_base_memory_operand.
17806 (move_dest_operand): Allow symbolic_memory_operands.
17807 (symbolic_memory_operand): Check for LO_SOM.
17808 (symbolic_operand): Change default case to break.
17809 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17810 CONST_DOUBLE values to be reloaded by putting them into memory when
17811 the destination is a floating point register.
17812 (movdf): Remove code to handle CONST_DOUBLE.
17813 (movsf): Likewise.
17814 (reload_indf_r1): New.
17815 (reload_insf_r1): New.
17816 Consistently use "Q" and "T" constraints with integer and floating
17817 point move instructions, respectively.
17818 (movdi): Remove FAIL.
17819 Change predicate for source operand unamed DImode move from
17820 general_operand to move_src_operand.
17821 (umulsidi3): Change predicate for destination operand to
17822 register_operand.
17823 Likewise for similar unamed patterns.
17824 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17825 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17826 (hppa_legitimize_address): Simplify mask calculation.
17827 (pa_emit_move_sequence): Revised handling of secondary reloads from
17828 REG+D addresses for floating point loads and stores. Directly handle
17829 loading CONST0_RTX (mode) to a floating point register.
17830 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17831 to floating point registers. Don't restrict secondary reloads to
17832 floating point registers to integer modes. Revise some comments and
17833 cleanup some code.
17834 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17835 (pa_legitimate_address_p): New.
17836 (pa_legitimize_reload_address): New.
17837 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17838 (STRICT_REG_OK_FOR_BASE_P): New.
17839 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17840 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17841
17842 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17843 Andrew Dixie <andrewd@gentrack.com>
17844
17845 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17846 flag set.
17847
17848 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17849
17850 * expmed.c (extract_bit_field_1): Pass the full width of the
17851 structure to get_best_reg_extraction_insn.
17852
17853 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17854
17855 PR target/54601
17856 * configure.ac (use_cxa_atexit): Add AIX.
17857 * configure: Regenerate.
17858
17859 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17860
17861 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17862
17863 PR debug/54793
17864 * final.c (need_profile_function): New variable.
17865 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17866 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17867 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17868 notes, targetm.asm_out.function_prologue doesn't emit anything,
17869 HAVE_prologue and profiler should be emitted before prologue,
17870 set need_profile_function instead of emitting it.
17871 (final_scan_insn): If need_profile_function, emit
17872 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17873 NOTE_INSN_FUNCTION_BEG note.
17874
17875 2013-02-01 Richard Henderson <rth@redhat.com>
17876
17877 * config/rs6000/rs6000.md (smulditi3): New.
17878 (umulditi3): New.
17879
17880 * config/alpha/alpha.md (umulditi3): New.
17881
17882 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17883
17884 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17885 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17886
17887 2013-02-01 Richard Biener <rguenther@suse.de>
17888
17889 PR tree-optimization/56113
17890 * tree-ssa-structalias.c (label_visit): Reduce work for
17891 single-predecessor nodes.
17892
17893 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17894
17895 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17896 range isn't testing for zero.
17897
17898 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17899
17900 PR middle-end/56113
17901 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17902
17903 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17904 Nick Clifton <nickc@redhat.com>
17905
17906 * config/v850/constraints.md (Q): Define as a memory constraint.
17907 * config/v850/predicates.md (label_ref_operand): New predicate.
17908 (e3v5_shift_operand): New predicate.
17909 (ior_operator): New predicate.
17910 * config/v850/t-v850: Add e3v5 multilib.
17911 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17912 (v850_gen_movdi): Prototype.
17913 * config/v850/v850.c: Add support for e3v5 architecture.
17914 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17915 TARGET_V850E_UP.
17916 (construct_save_jarl): Add e3v5 long JARL support.
17917 (v850_adjust_insn_length): New function. Adjust length of call
17918 insns when using e3v5 instructions.
17919 (v850_gen_movdi): New function: Generate instructions to move a
17920 DImode value.
17921 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17922 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17923 (TARGET_USE_FPU): Enable for e3v5.
17924 (CONST_OK_FOR_W): New macro.
17925 (ADJUST_INSN_LENGTH): Define.
17926 * config/v850/v850.md (UNSPEC_LOOP): Define.
17927 (attr cpu): Add v850e3v5.
17928 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17929 (movdi): New pattern.
17930 (movdi_internal): New pattern.
17931 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17932 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17933 (cstoresf4): Likewise.
17934 (cstoredf4): Likewise.
17935 (insv): New pattern.
17936 (rotlso3_a): New pattern.
17937 (rotlsi3_b): New pattern
17938 (rotlsi3_v850e3v5): New pattern.
17939 (doloop_begin): New pattern.
17940 (fix_loop_counter): New pattern.
17941 (doloop_end): New pattern.
17942 (branch_normal): Add e3v5 long branch support.
17943 (branch_invert): Likewise.
17944 (branch_z_normal): Likewise.
17945 (branch_z_invert): Likewise.
17946 (branch_nz_normal): Likewise.
17947 (branch_nz_invert): Likewise.
17948 (call_internal_short): Add e3v5 register-indirect JARL support.
17949 (call_internal_long): Likewise.
17950 (call_value_internal_short): Likewise.
17951 (call_value_internal_long): Likewise.
17952 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17953 (mloop): New option.
17954 * config.gcc: Add support for configuring v840e3v5 target.
17955 * doc/invoke.texi: Document new v850 specific command line options.
17956
17957 2013-01-31 Paul Koning <ni1d@arrl.net>
17958
17959 PR debug/55059
17960 PR debug/54508
17961 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17962 children if parent is a class.
17963 (prune_unused_types_prune): Don't add DW_AT_declaration.
17964
17965 2013-01-31 Richard Biener <rguenther@suse.de>
17966
17967 PR tree-optimization/56157
17968 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17969 match up operand with SLP child.
17970
17971 2013-01-31 Jason Merrill <jason@redhat.com>
17972
17973 PR debug/54410
17974 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17975 parameters the first time.
17976 (gen_scheduled_generic_parms_dies): Check completeness here.
17977
17978 2013-01-31 Richard Biener <rguenther@suse.de>
17979
17980 PR middle-end/53073
17981 * common.opt (faggressive-loop-optimizations): New flag,
17982 enabled by default.
17983 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17984 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17985 infer_loop_bounds_from_undefined by it.
17986
17987 2013-01-31 Richard Biener <rguenther@suse.de>
17988
17989 PR tree-optimization/56150
17990 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17991 visit virtual operands.
17992 (find_uses_to_rename_bb): Likewise.
17993
17994 2013-01-31 Richard Biener <rguenther@suse.de>
17995
17996 PR tree-optimization/56150
17997 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17998 mixed store non-store stmts.
17999
18000 2013-01-30 Jakub Jelinek <jakub@redhat.com>
18001
18002 PR sanitizer/55374
18003 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
18004 LIBASAN_EARLY_SPEC is defined.
18005 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
18006 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
18007 before %o.
18008 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
18009
18010 PR c++/55742
18011 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
18012 invalid args instead of ICEing on it.
18013 (ix86_valid_target_attribute_tree): Return error_mark_node if
18014 ix86_valid_target_attribute_inner_p failed.
18015 (ix86_valid_target_attribute_p): Return false only if
18016 ix86_valid_target_attribute_tree returned error_mark_node. Allow
18017 target("default") attribute.
18018 (sorted_attr_string): Change argument from const char * to tree,
18019 merge in all target attribute arguments rather than just one.
18020 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
18021 instead of free. Avoid using strcat.
18022 (ix86_mangle_function_version_assembler_name): Mangle
18023 target("default") as if no target attribute is present. Adjust
18024 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18025 instead of xmalloc and XDELETEVEC instead of free.
18026 (ix86_function_versions): Don't return true if one of the decls
18027 doesn't have target attribute. If they don't and one of the decls
18028 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18029 sorted_attr_string caller. Use XDELETEVEC instead of free.
18030 (ix86_supports_function_versions): Remove.
18031 (make_name): Fix up formatting.
18032 (make_dispatcher_decl): Remove resolver_name and its initialization.
18033 Avoid leaking memory.
18034 (is_function_default_version): Return true if there is
18035 target("default") attribute rather than no target attribute at all.
18036 (make_resolver_func): Avoid leaking memory.
18037 (ix86_generate_version_dispatcher_body): Likewise.
18038 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18039 * target.def (supports_function_versions): Remove.
18040 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18041 * doc/tm.texi: Regenerated.
18042
18043 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18044
18045 PR rtl-optimization/56144
18046 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18047 for values with side effects.
18048
18049 2013-01-30 Richard Biener <rguenther@suse.de>
18050
18051 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18052 (sparseset_pop): Likewise.
18053 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18054 to be able to use quick_push in the worker loop.
18055
18056 2013-01-30 Marek Polacek <polacek@redhat.com>
18057
18058 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18059
18060 2013-01-30 Richard Biener <rguenther@suse.de>
18061
18062 PR lto/56147
18063 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18064
18065 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18066
18067 PR tree-optimization/56064
18068 * fixed-value.c (fixed_from_double_int): New function.
18069 * fixed-value.h (fixed_from_double_int): New prototype.
18070 (const_fixed_from_double_int): New static inline function.
18071 * fold-const.c (native_interpret_fixed): New static function.
18072 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18073 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18074 (native_encode_fixed): New static function.
18075 (native_encode_expr) <FIXED_CST>: Use it.
18076 (native_interpret_int): Move double_int worker code to...
18077 * double-int.c (double_int::from_buffer): ...this new static method.
18078 * double-int.h (double_int::from_buffer): Prototype it.
18079
18080 2013-01-30 Richard Biener <rguenther@suse.de>
18081
18082 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18083 New pointer-map and obstack.
18084 (init_alias_vars): Allocate pointer-map and obstack.
18085 (delete_points_to_sets): Free them.
18086 (find_what_var_points_to): Cache result.
18087 (find_what_p_points_to): Adjust for changed interface of
18088 find_what_var_points_to.
18089 (compute_points_to_sets): Likewise.
18090 (ipa_pta_execute): Likewise.
18091
18092 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18093
18094 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18095 * configure: Regenerate.
18096 * config.in: Regenerate.
18097 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18098 #nobits/#progbits if supported.
18099
18100 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18101
18102 PR target/56121
18103 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18104 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18105 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18106
18107 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18108
18109 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18110 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18111
18112 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18113
18114 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18115 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18116
18117 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18118
18119 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18120 declaration.
18121 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18122 * config/arm/cortex-a7.md: New bypasses using
18123 arm_mac_accumulator_is_result.
18124
18125 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18126
18127 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18128 (cortex_a7_neon_mla): Likewise.
18129 (cortex_a7_fpfmad): New reservation.
18130 (cortex_a7_fpmacs): Use ffmas and update required units.
18131 (cortex_a7_fpmuld): Update required units and latency.
18132 (cortex_a7_fpmacd): Likewise.
18133 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18134 (cortex_a7_neon). Likewise.
18135 (bypass) Update participating units.
18136
18137 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18138
18139 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18140 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18141 from fmac to ffma.
18142 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18143 (vfp_fmul): Use ffmad.
18144 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18145 (cortex_r4_fmacd): Use ffmad.
18146 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18147 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18148 (cortex_a9_fmacd): Use ffmad.
18149 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18150 (cortex_a8_vfp_macd): Use ffmad.
18151 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18152 (cortex_a5_fpmacd): Use ffmad.
18153 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18154 (cortex_a15_vfp_macd): Use ffmad.
18155 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18156
18157 2013-01-29 Jason Merrill <jason@redhat.com>
18158
18159 PR libstdc++/54314
18160 * varasm.c (default_assemble_visibility): Don't warn about
18161 visibility on artificial decls.
18162
18163 2013-01-29 Richard Biener <rguenther@suse.de>
18164
18165 PR tree-optimization/56113
18166 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18167 the bitmap leader.
18168 (label_visit): Free duplicate bitmaps and record the leader instead.
18169 (perform_var_substitution): Adjust.
18170
18171 2013-01-29 Richard Biener <rguenther@suse.de>
18172
18173 PR tree-optimization/55270
18174 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18175 the CFG, schedule loops for fixup.
18176
18177 2013-01-29 Nick Clifton <nickc@redhat.com>
18178
18179 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18180 SP_REG.
18181
18182 2013-01-28 Leif Ekblad <leif@rdos.net>
18183
18184 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18185 * config/i386/i386.h (TARGET_RDOS): New macro.
18186 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18187 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18188 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
18189 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
18190 DEFAULT_LARGE_SECTION_THRESHOLD.
18191 * config/i386/i386.md (R14_REG, R15_REG): New constants.
18192 * config/i386/rdos.h: New file.
18193 * config/i386/rdos64.h: New file.
18194
18195 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
18196
18197 PR other/54814
18198 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
18199 TEST_HARD_REG_BIT.
18200
18201 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18202
18203 PR rtl-optimization/56117
18204 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
18205 call cselib_lookup_from_insn on the MEM before calling
18206 add_insn_mem_dependence.
18207
18208 2013-01-28 Richard Biener <rguenther@suse.de>
18209
18210 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
18211 to a stmt that didn't have one.
18212 (copy_phis_for_bb): Likewise for PHI arguments.
18213 (copy_debug_stmt): Likewise for debug stmts.
18214
18215 2013-01-28 Richard Biener <rguenther@suse.de>
18216
18217 PR tree-optimization/56034
18218 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
18219 (partition_builtin_p): Adjust.
18220 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
18221 it is the last partition.
18222 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
18223 up the vertex for the definition.
18224 (classify_partition): Classify whether a partition is a
18225 PKIND_REDUCTION, thus has uses outside of the loop.
18226 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
18227 Merge all PKIND_REDUCTION partitions into the last partition.
18228 (tree_loop_distribution): Seed partitions from reductions as well.
18229
18230 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18231
18232 PR tree-optimization/56125
18233 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
18234 pow(x,c) into sqrt(x) * powi(x, n/2) or
18235 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
18236 optimizing for size.
18237 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
18238 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
18239 integer.
18240
18241 PR tree-optimization/56094
18242 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
18243 to UNKNOWN_LOCATION while gimplifying expr.
18244
18245 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
18246
18247 PR target/56114
18248 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
18249 operand 0 in movabs insn template for -masm=intel asm alternative.
18250 (*movabs<mode>_2): Ditto for operand 1.
18251
18252 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
18253
18254 PR target/54663
18255 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
18256 of microblaze-c.o
18257
18258 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18259
18260 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
18261 tm_file.
18262
18263 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18264
18265 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
18266 Undef to avoid warning.
18267
18268 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18269
18270 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
18271 * configure: Regenerate.
18272
18273 2013-01-25 Jakub Jelinek <jakub@redhat.com>
18274
18275 PR tree-optimization/56098
18276 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
18277 for stmts with volatile ops.
18278 (cond_store_replacement): Don't optimize if assign has volatile ops.
18279 (cond_if_else_store_replacement_1): Don't optimize if either
18280 then_assign or else_assign have volatile ops.
18281 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
18282 volatile ops.
18283
18284 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18285
18286 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
18287
18288 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18289
18290 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
18291 missing ':' in asm example.
18292
18293 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
18294
18295 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
18296 entries into lane and laneq entries.
18297 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
18298 Remove AdvSIMD scalar modes.
18299 (aarch64_sq<r>dmulh_laneq<mode>): New.
18300 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
18301 modes.
18302 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
18303 builtin implementations to relfect changes in RTL in aarch64-simd.md.
18304 * config/aarch64/iterators.md (VCOND): New.
18305 (VCONQ): New.
18306
18307 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18308
18309 PR target/54222
18310 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
18311 Add NULL LIBNAME argument to existing definitions.
18312 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
18313 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
18314 * config/avr/avr.c (DEF_BUILTIN): Same.
18315 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
18316 (avr_expand_builtin): Expand to a vanilla call if a libgcc
18317 implementation is available (DECL_ASSEMBLER_NAME is set).
18318 (avr_fold_absfx): New static function.
18319 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
18320 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
18321 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
18322 AVR_BUILTIN_ABSLLK.
18323 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
18324 (abshk, absk, abslk, absllk): Provide as static inline functions.
18325
18326 2013-01-25 Marek Polacek <polacek@redhat.com>
18327
18328 PR tree-optimization/56035
18329 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
18330
18331 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
18332
18333 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
18334 (*movtf_internal_rex64): Add (!o,C) alternative
18335 (*movxf_internal_rex64): Ditto.
18336 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
18337
18338 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
18339
18340 * doc/invoke.texi: fix typo.
18341 * doc/objc.texi: fix typo.
18342
18343 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
18344
18345 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
18346 for the first two alternatives.
18347
18348 2013-01-24 Diego Novillo <dnovillo@google.com>
18349
18350 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
18351 (ggc-zone.o): Remove.
18352 * configure.ac: Remove option --with-gc.
18353 * configure: Re-generate.
18354 * doc/install.texi: Remove documentation for --with-gc.
18355 * gengtype.c (write_enum_defn): Remove. Update all users.
18356 (write_Types_process_field): Remove generation of gt_e_* argument.
18357 (output_type_enum): Remove. Update all users.
18358 (write_enum_defn): Remove. Update all users.
18359 (enum alloc_zone): Remove. Update all users.
18360 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
18361 * ggc-common.c (ggc_splay_alloc): Remove first argument.
18362 Update all callers.
18363 (struct ptr_data): Remove field TYPE. Update all users.
18364 (gt_pch_note_object): Remove argument TYPE. Update all users.
18365 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
18366 Update all users.
18367 * ggc-none.c (ggc_alloc_typed_stat): Remove.
18368 (struct alloc_zone): Remove.
18369 (ggc_internal_alloc_zone_stat): Remove.
18370 (ggc_internal_cleared_alloc_zone_stat): Remove.
18371 * ggc-page.c (ggc_alloc_typed_stat): Remove.
18372 (ggc_pch_count_object): Remove last argument. Update all users.
18373 (ggc_pch_alloc_object): Remove last argument. Update all users.
18374 (struct alloc_zone): Remove.
18375 * ggc-zone.c: Remove.
18376 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
18377 (struct alloc_zone): Remove.
18378 (ggc_alloc_typed_stat): Remove.
18379 (ggc_alloc_typed): Remove.
18380 (ggc_splay_alloc): Remove first argument.
18381 (rtl_zone): Remove. Update all users.
18382 (tree_zone): Remove. Update all users.
18383 (tree_id_zone): Remove. Update all users.
18384 (ggc_internal_zone_alloc_stat): Remove. Update all users.
18385 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
18386 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
18387 * tree-ssanames.c: Remove references to zone allocator in comments.
18388
18389 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
18390
18391 * config/avr/avr.c (avr_out_fract): Make register numbers that
18392 might be outside of source operand signed.
18393
18394 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
18395
18396 * config/i386/constraints.md (Yf): New constraint.
18397 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
18398 of f constraint to conditionaly disable x87 register preferences.
18399 (*movdf_internal): Ditto.
18400 (*movsf_internal): Ditto.
18401
18402 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
18403
18404 PR inline-asm/55934
18405 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
18406 that have operands with impossible constraints.
18407 Add a FIXME for a speed-up opportunity.
18408 * lra-constraints.c (process_alt_operands): Verify that a class
18409 selected from constraints on asms is valid for the operand mode.
18410 (curr_insn_transform): Remove incorrect comment.
18411
18412 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
18413
18414 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
18415 TOC operand is a valid symbol ref in the constant pool.
18416
18417 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18418
18419 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
18420
18421 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
18422
18423 PR target/54222
18424 * config/avr/stdfix.h: New file.
18425 * t-avr (stdfix-gcc.h): New rule to build it.
18426 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
18427
18428 2013-01-23 Kostya Serebryany <kcc@google.com>
18429
18430 * config/darwin.h: remove dependency on
18431 CoreFoundation (asan on Mac OS).
18432
18433 2013-01-23 Jakub Jelinek <jakub@redhat.com>
18434
18435 PR target/49069
18436 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
18437 instead of cmpdi_operand for first comparison operand.
18438 Don't assert that comparison operands aren't both constants.
18439
18440 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18441
18442 * doc/install.texi (Downloading the Source): Update references to
18443 downloading separate components.
18444
18445 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18446
18447 * doc/extend.texi (__int128): Improve grammar.
18448
18449 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
18450
18451 PR target/56028
18452 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
18453 alternative to (o,r).
18454 (*movdi_internal_rex64): Remove (!o,n) alternative.
18455 (DImode immediate->memory splitter): Remove.
18456 (DImode immediate->memory peephole2): Remove.
18457 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
18458 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
18459 alternative to (!o,*r).
18460 (*movtf_internal_sse): New pattern.
18461 (*movxf_internal_rex64): New pattern.
18462 (*movxf_internal): Disable for TARGET_64BIT.
18463 (*movdf_internal_rex64): Remove (!o,F) alternative.
18464
18465 2013-01-22 Jakub Jelinek <jakub@redhat.com>
18466
18467 PR middle-end/56074
18468 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
18469 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
18470 * tree-vect-loop-manip.c (find_loop_location): Also ignore
18471 stmt locations where LOCATION_LOCUS of the stmt location is
18472 UNKNOWN_LOCATION or BUILTINS_LOCATION.
18473
18474 PR target/55686
18475 * config/i386/i386.md (UNSPEC_STOS): New.
18476 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
18477 *strsetqi_1): Add UNSPEC_STOS.
18478
18479 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
18480
18481 PR c++/56067
18482 * doc/invoke.texi: Remove left over -Wsynth example.
18483
18484 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18485
18486 PR tree-optimization/56051
18487 * fold-const.c (fold_binary_loc): Don't fold
18488 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
18489 a narrowing conversion, or widening conversion from signed
18490 to unsigned.
18491
18492 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18493
18494 PR rtl-optimization/56023
18495 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
18496 dependent on debug instruction.
18497
18498 2013-01-21 Martin Jambor <mjambor@suse.cz>
18499
18500 PR middle-end/56022
18501 * function.c (allocate_struct_function): Call
18502 invoke_set_current_function_hook earlier.
18503
18504 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18505
18506 * reload1.c (init_reload): Only initialize reload_obstack
18507 during the first call.
18508
18509 2013-01-21 Marek Polacek <polacek@redhat.com>
18510
18511 * cfgloop.c (verify_loop_structure): Fix up grammar.
18512
18513 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
18514
18515 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
18516 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
18517
18518 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18519
18520 PR target/56058
18521 * config/arm/marvell-pj4.md: Update copyright year.
18522 Fix up use of alu to alu_reg and simple_alu_imm.
18523
18524 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18525
18526 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
18527
18528 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
18529
18530 PR target/55433
18531 * lra-constraints.c (curr_insn_transform): Don't reuse original
18532 insn for secondary memory move when memory mode should be different.
18533
18534 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18535
18536 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
18537 atomic_storedi_1): New patterns.
18538
18539 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18540
18541 btver2 pipeline descriptions.
18542 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
18543 descriptions.
18544 * config/i386/i386.md (btver2_decode): New type attributes.
18545 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
18546 type attributes.
18547 * config/i386/btver2.md: New file describing btver2 pipelines.
18548
18549 2013-01-19 Andrew Pinski <apinski@cavium.com>
18550
18551 PR tree-optimization/52631
18552 * tree-ssa-sccvn (visit_use): Before looking up the original
18553 statement, try looking up the simplified expression.
18554
18555 2013-01-19 Anthony Green <green@moxielogic.com>
18556
18557 * config/moxie/moxie.c (moxie_expand_prologue): Set
18558 current_function_static_stack_size.
18559
18560 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18561
18562 PR tree-optimization/56029
18563 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
18564 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
18565
18566 2013-01-18 Sharad Singhai <singhai@google.com>
18567
18568 PR tree-optimization/55995
18569 * dumpfile.c (dump_loc): Print location only if available.
18570 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
18571
18572 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
18573
18574 PR target/55433
18575 * lra-constraints.c (curr_insn_transform): Reuse original insn for
18576 secondary memory move.
18577 (inherit_reload_reg): Use rclass instead of cl for
18578 check_secondary_memory_needed_p.
18579
18580 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18581
18582 PR middle-end/56015
18583 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
18584 the case where writing real complex part of target modifies op1.
18585
18586 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
18587
18588 * config/aarch64/aarch64-simd.md
18589 (aarch64_vcond_internal<mode>): Handle unordered cases.
18590 * config/aarch64/iterators.md (v_cmp_result): New.
18591
18592 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
18593 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18594
18595 * config/arm/marvell-pj4.md: New file.
18596 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
18597 * config/arm/arm.md (generic_sched): Add marvell_pj4.
18598 (generic_vfp): Likewise.
18599 * config/arm/arm-cores.def: Add marvell-pj4.
18600 * config/arm/arm-tune.md: Regenerate.
18601 * config/arm/arm-tables.opt: Regenerate.
18602 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
18603 * doc/invoke.texi: Document marvell-pj4.
18604
18605 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
18606
18607 * config/aarch64/arm_neon.h: Map scalar types to standard types.
18608
18609 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18610
18611 PR debug/54114
18612 PR debug/54402
18613 PR debug/49888
18614 * var-tracking.c (negative_power_of_two_p): New.
18615 (global_get_addr_cache, local_get_addr_cache): New.
18616 (get_addr_from_global_cache, get_addr_from_local_cache): New.
18617 (vt_canonicalize_addr): Rewrite using the above. Adjust the
18618 heading comment.
18619 (vt_stack_offset_p): Remove.
18620 (vt_canon_true_dep): Always canonicalize loc's address.
18621 (clobber_overlapping_mems): Make sure we have a MEM.
18622 (local_get_addr_clear_given_value): New.
18623 (val_reset): Clear local cached entries.
18624 (compute_bb_dataflow): Create and release the local cache.
18625 Disable duplicate MEMs clobbering.
18626 (emit_notes_in_bb): Clobber MEMs likewise.
18627 (vt_emit_notes): Create and release the local cache.
18628 (vt_initialize, vt_finalize): Create and release the global
18629 cache, respectively.
18630 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
18631
18632 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18633
18634 PR libmudflap/53359
18635 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
18636 not found in the symtab.
18637
18638 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18639
18640 PR debug/56006
18641 PR rtl-optimization/55547
18642 PR rtl-optimization/53827
18643 PR debug/53671
18644 PR debug/49888
18645 * alias.c (offset_overlap_p): New, factored out of...
18646 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
18647 the conservative special case for symbolic constants. Don't
18648 adjust zero sizes on alignment.
18649
18650 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
18651
18652 PR rtl-optimization/52573
18653 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
18654 REG_UNUSED for the same register.
18655
18656 2013-01-17 Richard Biener <rguenther@suse.de>
18657 Marek Polacek <polacek@redhat.com>
18658
18659 PR rtl-optimization/55833
18660 * loop-unswitch.c (unswitch_loops): Move loop verification...
18661 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
18662 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
18663 Set it to true when we're removing a loop from hierarchy tree in
18664 an irreducible region.
18665 (fix_bb_placements): Adjust caller.
18666 (fix_loop_placements): Likewise.
18667
18668 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
18669
18670 * config/avr/builtins.def (DEF_BUILTIN): Factor out
18671 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
18672 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
18673 Remove ID. Adjust comments.
18674 * config/avr/avr-c.c (avr_builtin_name): Remove.
18675 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
18676 * config/avr/avr.c (avr_tolower): New static function.
18677 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
18678 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
18679 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
18680 default expansion.
18681
18682 2013-01-17 Jan Hubicka <jh@suse.cz>
18683
18684 PR tree-optimization/55273
18685 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
18686
18687 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
18688
18689 PR target/55981
18690 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
18691 store through atomic_store<mode>_1.
18692 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
18693
18694 2013-01-17 Martin Jambor <mjambor@suse.cz>
18695
18696 PR tree-optimizations/55264
18697 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
18698 for virtual methods.
18699 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
18700 virtual methods before inlining is over.
18701 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
18702 virtual functions.
18703 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
18704 non-virtual.
18705
18706 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
18707
18708 PR rtl-optimization/56005
18709 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
18710 pending reads for prefetch.
18711
18712 2013-01-16 Ian Bolton <ian.bolton@arm.com>
18713
18714 * config/aarch64/aarch64.md
18715 (*cstoresi_neg_uxtw): New pattern.
18716 (*cmovsi_insn_uxtw): New pattern.
18717 (*<optab>si3_uxtw): New pattern.
18718 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
18719 (*<optab>si3_insn_uxtw): New pattern.
18720 (*bswapsi2_uxtw): New pattern.
18721
18722 2013-01-16 Richard Biener <rguenther@suse.de>
18723
18724 * tree-inline.c (tree_function_versioning): Remove set but
18725 never used variable.
18726
18727 2013-01-16 Richard Biener <rguenther@suse.de>
18728
18729 PR tree-optimization/55964
18730 * tree-flow.h (rename_variables_in_loop): Remove.
18731 (rename_variables_in_bb): Likewise.
18732 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
18733 (copy_loop_before): Adjust and delete update-ssa status.
18734 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
18735 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
18736 (rename_variables_in_loop): Remove.
18737 (slpeel_update_phis_for_duplicate_loop): Likewise.
18738 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
18739 use available cfg machinery instead of duplicating it.
18740 Update PHI nodes and perform poor-mans SSA update here.
18741 (slpeel_tree_peel_loop_to_edge): Adjust.
18742
18743 2013-01-16 Richard Biener <rguenther@suse.de>
18744
18745 PR tree-optimization/54767
18746 PR tree-optimization/53465
18747 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
18748 (vrp_visit_phi_node): For PHI arguments coming via backedges
18749 drop all symbolical range information.
18750 (execute_vrp): Compute backedges.
18751
18752 2013-01-16 Richard Biener <rguenther@suse.de>
18753
18754 * doc/install.texi: Update CLooG and ISL requirements to
18755 0.18.0 and 0.11.1.
18756
18757 2013-01-16 Christian Bruel <christian.bruel@st.com>
18758
18759 PR target/55301
18760 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
18761 (broken_move): Handle UNSPECV_SP_SWITCH_B.
18762 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
18763
18764 2013-01-16 DJ Delorie <dj@redhat.com>
18765
18766 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
18767 (UNSPECV_SP_SWITCH_E): New.
18768 (sp_switch_1): Change to an unspec.
18769 (sp_switch_2): Change to an unspec. Don't use post-inc when we
18770 replace $r15.
18771
18772 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
18773
18774 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
18775 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
18776 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
18777 (expand_mem_thread_fence): Ditto.
18778 (expand_mem_signal_fence): Ditto.
18779 (expand_atomic_load): Ditto.
18780 (expand_atomic_store): Ditto.
18781
18782 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
18783
18784 PR rtl-optimization/55547
18785 PR rtl-optimization/53827
18786 PR debug/53671
18787 PR debug/49888
18788 * alias.c (memrefs_conflict_p): Set sizes to negative after
18789 AND adjustments.
18790
18791 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18792
18793 PR target/55940
18794 * function.c (thread_prologue_and_epilogue_insns): Always
18795 add crtl->drap_reg to set_up_by_prologue.set, even if
18796 stack_realign_drap is false.
18797
18798 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18799
18800 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18801 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18802 *call): Fix indention.
18803
18804 2013-01-15 Tom de Vries <tom@codesourcery.com>
18805
18806 PR target/55876
18807 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18808 Update comment.
18809
18810 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18811
18812 PR rtl-optimization/55153
18813 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18814
18815 2013-01-15 Martin Jambor <mjambor@suse.cz>
18816
18817 PR tree-optimization/55920
18818 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18819 accesses as grp_to_be_debug_replaced.
18820
18821 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18822
18823 PR tree-optimization/55920
18824 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18825 there is non-useless type conversion needed from debug rhs to lhs,
18826 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18827
18828 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18829 Mikael Pettersson <mikpe@it.uu.se>
18830
18831 PR target/43961
18832 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18833 Thumb.
18834 (ASM_OUTPUT_CASE_LABEL): Remove.
18835 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18836 * final.c (shorten_branches): Update alignment of labels before
18837 jump tables if CASE_VECTOR_SHORTEN_MODE.
18838
18839 2013-01-15 Richard Biener <rguenther@suse.de>
18840
18841 PR bootstrap/55961
18842 * system.h: Do not include gmp.h for building host tools.
18843
18844 2013-01-15 Richard Biener <rguenther@suse.de>
18845
18846 PR middle-end/55882
18847 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18848 account for bitpos when computing alignment.
18849
18850 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18851
18852 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18853 (ix86_target_macros_internal): Likewise.
18854
18855 * config/i386/i386.c (m_CORE2I7): Removed.
18856 (m_CORE_HASWELL): New macro.
18857 (m_CORE_ALL): Likewise.
18858 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18859 (initial_ix86_arch_features): Likewise.
18860 (processor_target_table): Initializations for Core avx2.
18861 (cpu_names): New names "core-avx2".
18862 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18863 PROCESSOR_CORE_HASWELL.
18864 (ix86_issue_rate): New case.
18865 (ia32_multipass_dfa_lookahead): Likewise.
18866 (ix86_sched_init_global): Likewise.
18867
18868 * config/i386/i386.h (TARGET_HASWELL): New macro.
18869 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18870 (processor_type): New PROCESSOR_HASWELL.
18871
18872 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18873
18874 PR tree-optimization/55955
18875 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18876 *SHIFT_EXPR and *ROTATE_EXPR codes.
18877
18878 PR tree-optimization/48766
18879 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18880 -ftrapv disable -fwrapv.
18881
18882 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18883
18884 PR target/55974
18885 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18886 etc. to 1 and not to __flash.
18887 Use LL suffix for __INT24_MAX__ with -mint8.
18888 Use ULL suffix for __UINT24_MAX__ with -mint8.
18889
18890 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18891
18892 * config/avr/avr-arch.h
18893 (struct base_arch_s): Use typedef avr_arch_t instead.
18894 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18895 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18896 * config/avr/avr.c: Same.
18897 * config/avr/avr-devices.c: Same.
18898 * config/avr/driver-avr.c: Same.
18899 * config/avr/gen-avr-mmcu-texi.c: Same.
18900 * config/avr/avr-mcus.def: Adjust comment.
18901
18902 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18903
18904 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18905 * config/aarch64/iterators.md (VALLDI): New.
18906
18907 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18908 Andi Kleen <ak@linux.intel.com>
18909
18910 PR target/55948
18911 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18912 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18913 memmodel flag.
18914
18915 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18916
18917 * config/avr/avr-stdint.h: Remove trailing blanks.
18918 * config/avr/avr-log.h: Same.
18919 * config/avr/avr-arch.h: Same.
18920 * config/avr/avr-devices.c: Same.
18921 * config/avr/avr-dimode.md: Same.
18922 * config/avr/predicates.md: Same.
18923 * config/avr/avr-c.c: Same. And fix typo.
18924
18925 * config/avr/avr-protos.h: Same. And:
18926 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18927 (init_cumulative_args): Rename to avr_init_cumulative_args.
18928 (expand_prologue): Rename to avr_expand_prologue.
18929 (expand_epilogue): Rename to avr_expand_epilogue.
18930 (adjust_insn_length): Rename to avr_adjust_insn_length.
18931 (notice_update_cc): Rename to avr_notice_update_cc.
18932 (final_prescan_insn): Rename to avr_final_prescan_insn.
18933 * config/avr/avr.c: Same.
18934 * config/avr/avr.h: Same.
18935 * config/avr/avr.md: Remove trailing blanks.
18936 (prologue): Use avr_expand_prologue.
18937 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18938
18939 2013-01-14 Richard Biener <rguenther@suse.de>
18940
18941 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18942 verify_location, collect_subblocks): New functions.
18943 (verify_gimple_in_cfg): Verify that locations only reference
18944 BLOCKs in the functions BLOCK tree.
18945
18946 2013-01-14 Richard Biener <rguenther@suse.de>
18947
18948 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18949 PHI argument.
18950 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18951 unshare reference.
18952 (insert_out_of_ssa_copy_on_edge): Likewise.
18953 (rewrite_close_phi_out_of_ssa): Likewise.
18954 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18955 debug expressions.
18956 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18957 propagated constants.
18958 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18959 can not be shared.
18960
18961 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18962
18963 * config/avr/avr-modes.def: Add GPL copyright notice.
18964
18965 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18966
18967 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18968 MEMMODEL_MASK to determine memory model.
18969 (atomic_store<mode>): Ditto from operands[2].
18970 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18971
18972 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18973
18974 PR fortran/55935
18975 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18976 (fold_gimple_assign): Don't call unshare_expr here.
18977 (fold_ctor_reference): Call unshare_expr.
18978
18979 2013-01-13 Terry Guo <terry.guo@arm.com>
18980
18981 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18982 * doc/fragments.texi: Document MULTILIB_REUSE.
18983 * gcc.c (multilib_reuse): New internal spec.
18984 (set_multilib_dir): Also search multilib from multilib_reuse.
18985 * genmultilib (tmpmultilib3): Refactor code.
18986 (tmpmultilib4): Ditto.
18987 (multilib_reuse): New multilib argument.
18988
18989 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18990
18991 * Makefile.in: Update copyright.
18992
18993 2013-01-12 Tom de Vries <tom@codesourcery.com>
18994
18995 PR middle-end/55890
18996 * calls.c (expand_call): Check if arg_nr is valid.
18997
18998 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18999
19000 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
19001 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
19002 documentation. Add missing '__' in front of
19003 __builtin_ia32_packssdw256.
19004
19005 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19006
19007 PR target/55719
19008 * config/s390/s390.c (s390_preferred_reload_class): Do not return
19009 NO_REGS for larl operands.
19010 (s390_reload_larl_operand): Use s390_load_address instead of
19011 emit_move_insn.
19012
19013 2013-01-11 Richard Biener <rguenther@suse.de>
19014
19015 * tree-cfg.c (verify_node_sharing_1): Split out from ...
19016 (verify_node_sharing): ... here.
19017 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
19018
19019 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
19020
19021 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
19022 Substitute TREECHECKING.
19023 * configure: Regenerate.
19024 * Makefile.in (TREECHECKING): New.
19025
19026 2013-01-11 Richard Guenther <rguenther@suse.de>
19027
19028 PR tree-optimization/44061
19029 * tree-vrp.c (extract_range_basic): Compute zero as
19030 value-range for __builtin_constant_p of function parameters.
19031
19032 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19033
19034 Update copyright years.
19035
19036 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19037
19038 PR rtl-optimization/55672
19039 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19040 const to be eliminable.
19041
19042 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19043
19044 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19045 * configure: Regenerate.
19046
19047 2013-01-10 Richard Biener <rguenther@suse.de>
19048
19049 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19050
19051 2013-01-10 Richard Biener <rguenther@suse.de>
19052
19053 PR bootstrap/55792
19054 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19055 locations for virtual PHI arguments.
19056 (rewrite_update_phi_arguments): Likewise.
19057
19058 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19059
19060 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19061 on to assembler.
19062
19063 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19064
19065 PR tree-optimization/55921
19066 * tree-complex.c (expand_complex_asm): New function.
19067 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19068
19069 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19070
19071 PR target/55718
19072 * config/s390/s390.c (s390_symref_operand_p)
19073 (s390_loadrelative_operand_p): Merge the two functions.
19074 (s390_check_qrst_address, print_operand_address): Add parameters
19075 to s390_loadrelative_operand_p invokation.
19076 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19077 (s390_reload_larl_operand, s390_secondary_reload): Use
19078 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19079 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19080
19081 2013-01-09 Mike Stump <mikestump@comcast.net>
19082
19083 * dse.c (record_store): Remove unnecessary assert.
19084
19085 2013-01-09 Jan Hubicka <jh@suse.cz>
19086
19087 PR tree-optimization/55569
19088 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19089 * cfgloop.h (scale_loop_profile): Likewise.
19090
19091 2013-01-09 Jan Hubicka <jh@suse.cz>
19092
19093 PR lto/45375
19094 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19095 functions.
19096 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19097
19098 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19099
19100 PR middle-end/55114
19101 * expr.h (maybe_emit_group_store): Declare.
19102 * expr.c (maybe_emit_group_store): New function.
19103 * builtins.c (expand_builtin_int_roundingfn): Call it.
19104 (expand_builtin_int_roundingfn_2): Likewise.
19105
19106 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19107
19108 PR rtl-optimization/55829
19109 * lra-constraints.c (match_reload): Add code for absent output.
19110 (curr_insn_transform): Add code for reloads of matched inputs
19111 without output.
19112
19113 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19114
19115 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19116 attribute of movddup insn to DF.
19117 (*vec_interleave_lowv2df): Ditto.
19118 (vec_dupv2df): Ditto.
19119
19120 2013-01-09 Jan Hubicka <jh@suse.cz>
19121
19122 PR tree-optimiation/55875
19123 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19124 EVERY_ITERATION parameter.
19125 (number_of_iterations_exit): Check if exit is executed every iteration.
19126 (idx_infer_loop_bounds): Similarly here.
19127 (n_of_executions_at_most): Simplify
19128 to only test for cases where statement is dominated by the
19129 particular bound; handle correctly the "postdominance" test.
19130 (scev_probably_wraps_p): Use max loop iterations info
19131 as a global bound first.
19132
19133 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19134 Nick Clifton <nickc@redhat.com>
19135
19136 * config/v850/v850.md (cbranchsf4): New pattern.
19137 (cstoresf4): New pattern.
19138 (cbranchdf4): New pattern.
19139 (cstoredf4): New pattern.
19140 (movsicc): Disallow floating point comparisons.
19141 (cmpsf_le_insn): Fix order of operators.
19142 (cmpsf_lt_insn): Likewise.
19143 (cmpsf_eq_insn): Likewise.
19144 (cmpdf_le_insn): Likewise.
19145 (cmpdf_lt_insn): Likewise.
19146 (cmpdf_eq_insn): Likewise.
19147 (cmpsf_ge_insn): Use LE comparison.
19148 (cmpdf_ge_insn): Likewise.
19149 (cmpsf_gt_insn): Use LT comparison.
19150 (cmpdf_gt_insn): Likewise.
19151 (cmpsf_ne_insn): Delete pattern.
19152 (cmpdf_ne_insn): Delete pattern.
19153 * config/v850/v850.c (v850_gen_float_compare): Use
19154 gen_cmpdf_eq_insn for NE comparison.
19155 (v850_float_z_comparison_operator)
19156 (v850_float_nz_comparison_operator): Move from here ...
19157 * config/v850/predicates.md: ... to here. Move GT and GE
19158 comparisons into v850_float_z_comparison_operator.
19159 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19160 Delete prototype.
19161 (v850_float_nz_comparison_operator): Likewise.
19162
19163 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19164
19165 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19166 with calls to gen_insvsi/gen_insvdi.
19167
19168 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19169
19170 * config/i386/i386.c (initial_ix86_tune_features): Set up
19171 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19172
19173 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19174 Jakub Jelinek <jakub@redhat.com>
19175
19176 PR tree-optimization/48189
19177 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19178 If nitercst is 0, don't predict the exit edge.
19179
19180 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19181
19182 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19183 in asm_fprintf with reg_names.
19184 (aarch64_print_operand_address): Likewise.
19185 (aarch64_return_addr): Likewise.
19186 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19187
19188 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19189
19190 * config/pa/pa.h (VAL_U6_BITS_P): Define.
19191 (INT_U6_BITS): Likewise.
19192 * config/pa/predicates.md (uint6_operand): New predicate.
19193 (shift5_operand, shift6_operand): Likewise.
19194 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
19195 arith32_operand.
19196 (lshrdi3): Use shift6_operand.
19197 (shrpsi4, shrpdi4): New insn patterns.
19198 (extzv): Delete expander.
19199 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
19200 predicates in unamed zero extract patterns. Tighten common constraint.
19201 (extv): Delete expander.
19202 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
19203 predicates in unamed sign extract patterns. Tighten common constraint.
19204 (insv): Delete expander.
19205 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
19206 predicates in unamed insert patterns. Tighten common constraint.
19207 Change uint32_operand predicate to uint6_operand predicate in unamed
19208 DImode pattern to insert constant values of type 1...1xxxx.
19209
19210 2013-01-04 Jan Hubicka <jh@suse.cz>
19211
19212 PR tree-optimization/55823
19213 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
19214 issue.
19215
19216 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19217 Uros Bizjak <ubizjak@gmail.com>
19218
19219 PR rtl-optimization/55845
19220 * df-problems.c (can_move_insns_across): Stop scanning at
19221 volatile_insn_p source instruction or give up if
19222 across_from .. across_to range contains any volatile_insn_p
19223 instructions.
19224
19225 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19226
19227 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
19228 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
19229 Declare.
19230 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
19231 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
19232
19233 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19234
19235 PR fortran/55341
19236 * asan.c (asan_clear_shadow): New function.
19237 (asan_emit_stack_protection): Use it.
19238
19239 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19240
19241 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
19242 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
19243 with tab instead of space.
19244
19245 2013-01-08 Nick Clifton <nickc@redhat.com>
19246
19247 * config/rl78/rl78.c (rl78_expand_prologue): Always select
19248 register bank 0 at the start of an interrupt handler.
19249 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
19250 MDBH registers.
19251
19252 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19253
19254 * config/aarch64/aarch64-simd.md
19255 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
19256 (aarch64_simd_bsl): Likewise.
19257 (aarch64_vcond_internal<mode>): Likewise.
19258 (vcond<mode><mode>): Likewise.
19259 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
19260 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
19261
19262 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19263
19264 * config/aarch64/aarch64-builtins.c
19265 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
19266
19267 2013-01-08 Martin Jambor <mjambor@suse.cz>
19268
19269 PR debug/55579
19270 * tree-sra.c (analyze_access_subtree): Return true also after
19271 potentially creating a debug-only replacement.
19272
19273 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19274
19275 PR middle-end/55890
19276 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
19277
19278 PR tree-optimization/54120
19279 * tree-vrp.c (range_fits_type_p): Don't allow
19280 src_precision < precision from signed vr to unsigned_p
19281 if vr->min or vr->max is negative.
19282 (simplify_float_conversion_using_ranges): Test can_float_p
19283 against CODE_FOR_nothing.
19284
19285 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19286 Richard Biener <rguenther@suse.de>
19287
19288 PR middle-end/55851
19289 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
19290 types instead of just INTEGER_TYPE types.
19291
19292 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
19293
19294 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
19295 TF_SIZE): Define.
19296
19297 2013-01-07 Steve Ellcey <sellcey@mips.com>
19298
19299 PR target/42661
19300 * config/mips/mips.opt: Change mad to mmad to match documentation.
19301
19302 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19303
19304 PR target/55897
19305 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
19306 .progmemx.data now.
19307
19308 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19309
19310 PR target/55897
19311 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
19312 (avr_addrspace_t): Add .section_name field.
19313 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
19314 array size.
19315 (avr_addrspace): Same. Initialize .section_name. Remove last
19316 NULL entry. Put __memx into .progmemx.data.
19317 (progmem_section_prefix): Remove.
19318 (avr_asm_init_sections): No need to initialize progmem_section.
19319 (avr_asm_named_section): Use avr_addrspace[].section_name to get
19320 section name prefix.
19321 (avr_asm_select_section): Ditto. And use get_unnamed_section to
19322 retrieve the progmem section.
19323 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
19324 boundary to run over avr_addrspace[].
19325 (avr_register_target_pragmas): Ditto.
19326
19327 2013-01-06 Jakub Jelinek <jakub@redhat.com>
19328
19329 * varasm.c (output_constant_def_contents): For asan_protect_global
19330 protected strings, adjust DECL_ALIGN if needed, before testing for
19331 anchored symbols.
19332 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
19333 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
19334 normal decls.
19335 (output_object_block): For asan protected decls, emit asan padding
19336 after their contents.
19337 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
19338 (asan_finish_file): Test it here instead.
19339
19340 2013-01-07 Nick Clifton <nickc@redhat.com>
19341 Matthias Klose <doko@debian.org>
19342 Doug Kwan <dougkwan@google.com>
19343 H.J. Lu <hongjiu.lu@intel.com>
19344
19345 PR driver/55470
19346 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
19347
19348 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
19349
19350 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
19351
19352 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
19353
19354 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
19355
19356 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19357
19358 PR target/54461
19359 * doc/install.texi (Cross-Compiler-Specific Options): Document
19360 --with-avrlibc.
19361
19362 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
19363
19364 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
19365 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
19366 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
19367 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
19368 vqmovun_high_s64): Fix source operand number and update copyright.
19369
19370 2013-01-07 Richard Biener <rguenther@suse.de>
19371
19372 PR middle-end/55890
19373 * gimple.h (gimple_call_builtin_p): New overload.
19374 * gimple.c (validate_call): New function.
19375 (gimple_call_builtin_p): Likewise.
19376 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19377 Use gimple_call_builtin_p.
19378 (find_func_clobbers): Likewise.
19379 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
19380 (strlen_optimize_stmt): Likewise.
19381
19382 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19383
19384 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
19385 (vld1q_dup_*): Likewise.
19386 (vld1_*): Likewise.
19387 (vld1q_*): Likewise.
19388 (vld1_lane_*): Likewise.
19389 (vld1q_lane_*): Likewise.
19390
19391 2013-01-07 Richard Biener <rguenther@suse.de>
19392
19393 * lto-streamer.h (LTO_minor_version): Bump to 2.
19394
19395 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19396
19397 * config/aarch64/aarch64-protos.h
19398 (aarch64_const_double_zero_rtx_p): Rename to...
19399 (aarch64_float_const_zero_rtx_p): ...this.
19400 (aarch64_float_const_representable_p): New.
19401 (aarch64_output_simd_mov_immediate): Likewise.
19402 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
19403 move immediate case.
19404 * config/aarch64/aarch64.c
19405 (aarch64_const_double_zero_rtx_p): Rename to...
19406 (aarch64_float_const_zero_rtx_p): ...this.
19407 (aarch64_print_operand): Allow printing of new constants.
19408 (aarch64_valid_floating_const): New.
19409 (aarch64_legitimate_constant_p): Check for valid floating-point
19410 constants.
19411 (aarch64_simd_valid_immediate): Likewise.
19412 (aarch64_vect_float_const_representable_p): New.
19413 (aarch64_float_const_representable_p): Likewise.
19414 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
19415 (aarch64_output_simd_mov_immediate): New.
19416 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
19417 (*movdf_aarch64): Likewise.
19418 * config/aarch64/constraints.md (Ufc): New.
19419 (Y): call aarch64_float_const_zero_rtx.
19420 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
19421
19422 2013-01-07 Richard Biener <rguenther@suse.de>
19423
19424 PR tree-optimization/55888
19425 PR tree-optimization/55862
19426 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
19427 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
19428 not if it is contained therein.
19429
19430 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19431
19432 * config/avr/t-avr: Typo.
19433
19434 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19435
19436 PR55243
19437 * config/avr/t-avr: Don't automatically rebuild
19438 $(srcdir)/config/avr/t-multilib
19439 $(srcdir)/config/avr/avr-tables.opt
19440 $(srcdir)/doc/avr-mmcu.texi
19441 (avr-mcus): New phony target to build them on request.
19442 (s-avr-mlib, s-avr-mmcu-texi): Remove.
19443 * avr/avr-mcus.def: Adjust comments.
19444
19445 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
19446
19447 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
19448
19449 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19450
19451 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
19452
19453 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19454
19455 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
19456
19457 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
19458
19459 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
19460 to generate profiling.
19461 * config/rs6000/aix64.h (LIB_SPEC): Same.
19462
19463 2013-01-04 Andrew Pinski <apinski@cavium.com>
19464
19465 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
19466 New function.
19467 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
19468
19469 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
19470
19471 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
19472 unconditionally.
19473 (ix86_expand_move): Ditto.
19474 (ix86_zero_extend_to_Pmode): Ditto.
19475 (ix86_expand_call): Ditto.
19476 (ix86_expand_special_args_builtin): Ditto.
19477 (ix86_expand_builtin): Ditto.
19478
19479 2013-01-04 Richard Biener <rguenther@suse.de>
19480
19481 PR tree-optimization/55862
19482 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
19483 translating them through PHI nodes.
19484
19485 2013-01-04 Martin Jambor <mjambor@suse.cz>
19486
19487 PR tree-optimization/55755
19488 * tree-sra.c (sra_modify_assign): Do not check that an access has no
19489 children when trying to avoid producing a VIEW_CONVERT_EXPR.
19490
19491 2013-01-04 Marek Polacek <polacek@redhat.com>
19492
19493 PR middle-end/55859
19494 * opts.c (default_options_optimization): Clarify error message.
19495
19496 2013-01-04 Richard Biener <rguenther@suse.de>
19497
19498 PR middle-end/55863
19499 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
19500 reassociation.
19501
19502 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19503
19504 PR target/53789
19505 * config/pa/pa.md (movsi): Revert previous change.
19506 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
19507 references.
19508
19509 2013-01-03 Richard Henderson <rth@redhat.com>
19510
19511 * config/i386/i386.c (ix86_expand_move): Always assign to op1
19512 after eliminating TLS symbols.
19513
19514 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
19515
19516 PR bootstrap/50167
19517 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
19518 * graphite-poly.c (debug_gmp_value): Likewise.
19519
19520 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
19521
19522 PR target/55712
19523 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
19524 selected code model, define __code_mode_small__, __code_model_medium__,
19525 __code_model_large__, __code_model_32__ or __code_model_kernel__.
19526 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
19527 xchg temporary register with %k. Declare temporary register as
19528 early clobbered.
19529 [__x86_64__]: For medium and large code models, preserve %rbx register.
19530
19531 2013-01-03 Richard Biener <rguenther@suse.de>
19532
19533 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
19534 (dump_subscript): Adjust.
19535 (finalize_ddr_dependent): Do not dump redundant info.
19536 (analyze_siv_subscript): Adjust.
19537 (subscript_dependence_tester): Likewise.
19538 (compute_affine_dependence): Likewise.
19539
19540 2013-01-03 Richard Biener <rguenther@suse.de>
19541
19542 Revert
19543 2013-01-03 Richard Biener <rguenther@suse.de>
19544
19545 PR tree-optimization/55857
19546 * tree-vect-stmts.c (vectorizable_load): Do not setup
19547 re-alignment for invariant loads.
19548
19549 2013-01-02 Richard Biener <rguenther@suse.de>
19550
19551 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19552 invariant load do not generate a vector load from the scalar location.
19553
19554 2013-01-03 Richard Biener <rguenther@suse.de>
19555
19556 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
19557 for not vectorizing.
19558 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
19559 not build INDIRECT_REFs, call get_name once only.
19560 (vect_create_data_ref_ptr): Likewise. Dump base object kind
19561 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
19562
19563 2013-01-03 Richard Biener <rguenther@suse.de>
19564
19565 PR tree-optimization/55857
19566 * tree-vect-stmts.c (vectorizable_load): Do not setup
19567 re-alignment for invariant loads.
19568
19569 2013-01-03 Richard Biener <rguenther@suse.de>
19570
19571 PR lto/55848
19572 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
19573 prefer a built-in decl.
19574
19575 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19576
19577 * gcc.c (process_command): Update copyright notice dates.
19578 * gcov.c (print_version): Likewise.
19579 * gcov-dump.c (print_version): Likewise.
19580
19581 PR rtl-optimization/55838
19582 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
19583 iv0.step, iv1.step and step.
19584
19585 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19586 Marc Glisse <marc.glisse@inria.fr>
19587
19588 PR tree-optimization/55832
19589 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
19590 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
19591 integer_{one,zero}_node.
19592
19593 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19594
19595 PR debug/54402
19596 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
19597 * var-tracking.c (reverse_op): Don't add reverse ops to
19598 VALUEs that have already
19599 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
19600
19601 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
19602
19603 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
19604
19605 2013-01-02 Teresa Johnson <tejohnson@google.com>
19606
19607 * dumpfile.c (dump_loc): Print filename with location.
19608 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
19609 new location_t parameter to emit complete unroll message with
19610 new dump framework.
19611 (canonicalize_loop_induction_variables): Compute loops location
19612 and pass to try_unroll_loop_completely.
19613 * loop-unroll.c (report_unroll_peel): New function.
19614 (peel_loops_completely): Use new dump format with location
19615 for main dumpfile message, and invoke report_unroll_peel on success.
19616 (decide_unrolling_and_peeling): Ditto.
19617 (decide_peel_once_rolling): Remove old dumpfile message subsumed
19618 by report_unroll_peel.
19619 (decide_peel_completely): Ditto.
19620 (decide_unroll_constant_iterations): Ditto.
19621 (decide_unroll_runtime_iterations): Ditto.
19622 (decide_peel_simple): Ditto.
19623 (decide_unroll_stupid): Ditto.
19624 * cfgloop.c (get_loop_location): New function.
19625 * cfgloop.h (get_loop_location): Declare.
19626
19627 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19628
19629 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
19630 NULL.
19631
19632 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19633
19634 PR middle-end/55198
19635 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
19636 BLKmode objects when EXPAND_MEMORY is specified.
19637
19638 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19639
19640 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
19641 in loop predicate.
19642 (fold_builtin_cpu): Do not share cpu model decls across statements.
19643
19644 2013-01-02 Jason Merrill <jason@redhat.com>
19645
19646 PR c++/55804
19647 * tree.c (build_array_type_1): Revert earlier change.
19648
19649 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
19650
19651 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
19652 "cortex-a57".
19653 * config/aarch64/aarch64-tune.md: Re-generate.
19654
19655 2013-01-02 Richard Biener <rguenther@suse.de>
19656
19657 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19658 invariant load do not generate a vector load from the scalar location.
19659
19660 2013-01-02 Richard Biener <rguenther@suse.de>
19661
19662 PR bootstrap/55784
19663 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
19664 * configure: Regenerate.
19665
19666 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19667
19668 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
19669 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
19670 (expand_builtin_int_roundingfn_2): Keep the original target around
19671 for the fallback case.
19672
19673 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19674
19675 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
19676 to be clear for sign changes.
19677
19678 2013-01-01 Jan Hubicka <jh@suse.cz>
19679
19680 * ipa-inline-analysis.c: Fix formatting.
19681
19682 2013-01-01 Jakub Jelinek <jakub@redhat.com>
19683
19684 PR tree-optimization/55831
19685 * tree-vect-loop.c (get_initial_def_for_induction): Use
19686 gsi_after_labels instead of gsi_start_bb.
19687 \f
19688 Copyright (C) 2013 Free Software Foundation, Inc.
19689
19690 Copying and distribution of this file, with or without modification,
19691 are permitted in any medium without royalty provided the copyright
19692 notice and this notice are preserved.