1 2013-09-03 Richard Biener <rguenther@suse.de>
4 * tree-streamer-out.c: Include tm.h.
5 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7 2013-09-03 Jan Hubicka <jh@suse.cz>
9 * tree-profile.c (tree_profiling): Cleanup CFG when done.
11 2013-09-03 Alan Modra <amodra@gmail.com>
13 * config.gcc (powerpc*-*-linux*): Add support for little-endian
14 multilibs to big-endian target and vice versa.
15 * config/rs6000/t-linux64: Use := assignment on all vars.
16 (MULTILIB_EXTRA_OPTS): Remove fPIC.
17 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
18 * config/rs6000/t-linux64le: New file.
19 * config/rs6000/t-linux64bele: New file.
20 * config/rs6000/t-linux64lebe: New file.
22 2013-09-02 Jan Hubicka <jh@suse.cz>
24 * ipa-inline-transform.c (inline_transform): Do not
25 optimize_inline_calls when not optimizing.
27 2013-09-02 Jan Hubicka <jh@suse.cz>
29 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
30 duplicated nodes for assembler names.
31 * symtab.c (symtab_unregister_node): Do not attempt to unlink
32 hard registers from assembler name hash.
34 2013-09-02 Jan Hubicka <jh@suse.cz>
36 * ipa-split.c (execute_split_functions): Split externally visible
37 functions called once.
39 2013-09-02 Martin Jambor <mjambor@suse.cz>
42 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
43 linked list. When finding the correct duplicate, also consider also
44 the caller in additon to its inlined_to node.
46 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
48 * config/aarch64/aarch64-simd-builtins.def
49 (dup_lane_scalar): Remove.
50 * config/aarch64/aarch64-simd.md
51 (aarch64_simd_dup): Add 'w->w' alternative.
52 (aarch64_dup_lane<mode>): Allow for VALL.
53 (aarch64_dup_lane_scalar<mode>): Remove.
54 (aarch64_dup_lane_<vswap_width_name><mode>): New.
55 (aarch64_get_lane_signed<mode>): Add w->w altenative.
56 (aarch64_get_lane_unsigned<mode>): Likewise.
57 (aarch64_get_lane<mode>): Likewise.
58 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
59 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
60 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
61 (VCON): Change container of V2SF.
62 (vswap_width_name): Likewise.
63 * config/aarch64/arm_neon.h
64 (__aarch64_vdup_lane_any): New.
65 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
66 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
67 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
69 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
72 * expr.c (emit_move_complex): Do not move complex FP values as parts if
73 the source or the destination is a single hard register.
75 2013-09-02 Richard Biener <rguenther@suse.de>
78 * tree-scalar-evolution.c (instantiate_scev_name): Allow
81 2013-09-02 Richard Biener <rguenther@suse.de>
83 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
84 arithmetic to sizetype.
86 2013-09-02 Bin Cheng <bin.cheng@arm.com>
88 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
89 Find auto-increment use both before and after candidate.
91 2013-09-02 Marek Polacek <polacek@redhat.com>
93 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
95 2013-09-01 Jan Hubicka <jh@suse.cz>
97 * Makefile.in: Add ipa-profile.o
98 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
99 * cgraph.c (struct cgraph_propagate_frequency_data,
100 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
101 ipa-profile.c; replace cgraph_ by ipa_ prefix.
102 * cgraph.h (cgraph_propagate_frequency): Remove.
103 * ipa-inline-analysis.c: Include ipa-utils.h; drop duplicated cfgloop.h.
104 (inline_update_callee_summaries): Update.
105 * ipa-profile.c: New file.
106 * ipa-utils.h (ipa_propagate_frequency): Declare.
107 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
108 data-streamer.h, value-prof.h
109 (symtab_remove_unreachable_nodes): Update profile.
110 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
111 account_time_size, cmp_counts, dump_histogram,
112 ipa_profile_generate_summary, ipa_profile_write_summary,
113 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
114 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
115 Move to ipa-profile.c
117 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
119 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
121 2013-09-01 Jan Hubicka <jh@suse.cz>
123 * common.opt (fdevirtualize-speculatively): New function.
124 * invoke.texi (fdevirtualize-speculatively): Document.
125 * ipa-devirt.c: Include ipa-inline.h
126 (likely_target_p): New function.
127 (ipa_devirt): New function.
128 (gate_ipa_devirt): New function.
129 (pass_data_ipa_devirt): New static var.
130 (pass_ipa_devirt): Likewise.
131 (make_pass_ipa_devirt): New function.
132 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
133 (common_handle_option): Disable devirtualization when
134 value range profiling is available.
135 * passes.def (pass_ipa_devirt): Add.
136 * timever.def (TV_IPA_DEVIRT): New timevar.
137 * tree-pass.h (make_pass_ipa_devirt):
139 2013-09-01 Iain Sandoe <iain@codesourcery.com>
141 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
142 include sanitize(undefined).
144 2013-08-31 Diego Novillo <dnovillo@google.com>
146 * Makefile.in (TREE_CORE_H): Define.
148 (GTFILES): Add tree-core.h.
149 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
151 * tree-core.h: New file.
152 Move all data structures, enum, typedefs, global
153 declarations and constants from ...
156 2013-08-31 Jan Hubicka <jh@suse.cz>
158 * bulitins.c (expand_builtin): Do not early exit for gcov
159 instrumented functions.
161 2013-08-31 Marek Polacek <polacek@redhat.com>
163 * ubsan.c: Include tm_p.h.
165 2013-08-31 Jan Hubicka <jh@suse.cz>
167 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
170 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
171 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
172 * tree-cfg.c (verify_gimple_label): ... here.
173 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
174 (ipa_merge_profiles): New function.
175 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
176 (lto_input_function_body): Likewise.
177 * ipa-utils.h (ipa_merge_profiles): Declare.
178 * lto-streamer.h (lto_input_function_body): Update prototype.
179 (emit_label_in_global_context_p): Remove.
180 * lto-symtab.c: Include ipa-utils.h
181 (lto_cgraph_replace_node): Use ipa_merge_profiles.
183 2013-08-31 Jan Hubicka <jh@suse.cz>
185 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
187 2013-08-31 Jan Hubicka <jh@suse.cz>
189 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
191 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
193 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
194 "cmp" RTX before signed_comparison_operator check to account
197 2013-08-30 Jan Hubicka <jh@suse.cz>
199 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
200 (detect_type_change_1): Rename to ...
201 (detect_type_change): ... this one; early return on non-polymorphic
203 (detect_type_change_ssa): Add comp_type parameter; update
204 use of detect_type_change.
205 (compute_complex_assign_jump_func): Add param_type parameter;
206 update use of detect_type_change_ssa.
207 (compute_complex_ancestor_jump_func): Likewise.
208 (ipa_get_callee_param_type): New function.
209 (ipa_compute_jump_functions_for_edge): Compute parameter type;
210 update calls to the jump function computation functions.
212 2013-08-30 Teresa Johnson <tejohnson@google.com>
213 Steven Bosscher <steven@gcc.gnu.org>
215 * cfgrtl.c (fixup_new_cold_bb): New routine.
216 (commit_edge_insertions): Invoke fixup_partitions.
217 (find_partition_fixes): New routine.
218 (fixup_partitions): Ditto.
219 (verify_hot_cold_block_grouping): Update comments.
220 (rtl_verify_edges): Invoke find_partition_fixes.
221 (rtl_verify_bb_pointers): Update comments.
222 (rtl_verify_bb_layout): Ditto.
223 * basic-block.h (probably_never_executed_edge_p): Declare.
224 (fixup_partitions): Ditto.
225 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
226 * bb-reorder.c (sanitize_hot_paths): New function.
227 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
229 * predict.c (probably_never_executed_edge_p): New routine.
230 * cfg.c (check_bb_profile): Add partition insanity warnings.
232 2013-08-30 Meador Inge <meadori@codesourcery.com>
234 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
236 2013-08-30 Marek Polacek <polacek@redhat.com>
238 * Makefile.in (ubsan.o): Add.
239 (c-family/c-ubsan.o): Add.
240 (builtins.o): Add ubsan.h dependency.
243 * common.opt: Add -fsanitize=undefined option.
244 (flag_sanitize): Add variable.
245 (fsanitize=): Add option. Add Driver.
246 (fsanitize=thread): Remove option.
247 (fsanitize=address): Likewise.
248 (static-libubsan): New option.
249 * doc/invoke.texi: Document the new flag and -static-libubsan.
250 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
251 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
252 * builtin-attrs.def (ATTR_COLD): Define.
253 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
254 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
255 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
256 * flag-types.h (sanitize_code): New enum.
257 * opts.c (common_handle_option): Parse command line arguments
258 of -fsanitize=. Add -fsanitize=unreachable option.
259 * varasm.c (get_variable_section): Adjust.
260 (assemble_noswitch_variable): Likewise.
261 (assemble_variable): Likewise.
262 (output_constant_def_contents): Likewise.
263 (categorize_decl_for_section): Likewise.
264 (place_block_symbol): Likewise.
265 (output_object_block): Likewise.
266 * builtins.def: Likewise.
267 * toplev.c (compile_file): Likewise.
268 (process_options): Likewise.
269 * cppbuiltin.c: Likewise.
270 * tsan.c (tsan_pass): Likewise.
271 (tsan_gate): Likewise.
272 (tsan_gate_O0): Likewise.
273 * cfgexpand.c (partition_stack_vars): Likewise.
274 (expand_stack_vars): Likewise.
275 (defer_stack_allocation): Likewise.
276 (expand_used_vars): Likewise.
277 * cfgcleanup.c (old_insns_match_p): Likewise.
278 * asan.c (asan_finish_file): Likewise.
279 (asan_instrument): Likewise.
280 (gate_asan): Likewise.
281 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
282 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
283 (asan_global_struct): Use pointer_sized_int_node instead
284 calling build_nonstandard_integer_type.
285 (initialize_sanitizer_builtins): Likewise.
286 (asan_finish_file): Likewise.
287 * gcc.c: Document %{%:function(args):X}.
288 (static_spec_functions): Add sanitize.
289 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
290 store funcval != NULL there.
291 (do_spec_1): Adjust handle_spec_function caller.
292 (handle_braces): Allow %:function(args) as condition.
293 (sanitize_spec_function): New function.
294 (ADD_STATIC_LIBUBSAN_LIBS): Define.
295 (LIBUBSAN_SPEC): Likewise.
296 (LIBUBSAN_EARLY_SPEC): Likewise.
297 (SANITIZER_SPEC): Handle libubsan.
298 (SANITIZER_EARLY_SPEC): Likewise.
299 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
300 instead of fsanitize=address.
301 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
302 instead of fsanitize=address*.
303 * builtins.c: Include ubsan.h.
304 (fold_builtin_0): Instrument __builtin_unreachable.
305 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
306 instead of flag_asan.
307 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
308 (pointer_sized_int_node): Define.
309 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
311 2013-08-30 Mike Stump <mikestump@comcast.net>
313 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
316 2013-08-29 Jan Hubicka <jh@suse.cz>
318 * cgraph.c (cgraph_function_body_availability): Handle weakref
320 * passes.def: Remove pass_fixup_cfg.
321 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
322 track when we need to remove functions.
323 (gate_ipa_inline): Execute inlining always; add comment why.
324 (pass_data_ipa_inline): Remove TODO_remove_functions.
325 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
326 do not produce summaries.
327 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
328 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
329 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
330 constant pool and vtable.
332 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
334 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
335 New arm_neon.h's internal macros to specify 64-bit constants.
336 Avoid using stdint.h's macros.
338 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
340 * recog.c (verify_changes): Verify that changes[i].old is non-zero
341 before applying REG_P.
343 2013-08-30 Jakub Jelinek <jakub@redhat.com>
345 PR tree-optimization/58277
346 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
347 after seeing too many stmts with vdef in between dombb and current
348 bb, invalidate everything.
350 2013-08-30 Richard Biener <rguenther@suse.de>
352 * fold-const.c (fold_single_bit_test): Fix overflow test.
354 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
356 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
357 and which can live in a register, always retrieve the value on entry.
358 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
359 passed by invisible reference specially.
360 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
361 (vt_add_function_parameter): Correctly deal with a parameter passed by
364 2013-08-30 Jan Hubicka <jh@suse.cz>
366 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
368 2013-08-30 Richard Biener <rguenther@suse.de>
370 PR tree-optimization/58228
371 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
372 allow invariant loads in nested loop vectorization.
374 2013-08-30 Richard Biener <rguenther@suse.de>
376 PR tree-optimization/58223
377 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
378 (has_anti_or_output_dependence): ... this and adjust to also
379 look for output dependences.
380 (mark_nodes_having_upstream_mem_writes): Adjust.
381 (rdg_flag_uses): Likewise.
383 2013-08-30 Richard Biener <rguenther@suse.de>
385 PR tree-optimization/58010
386 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
387 assert that we have a loop-closed PHI.
389 2013-08-29 Jan Hubicka <jh@suse.cz>
391 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
392 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
393 * lto-section-in.c (lto_free_function_in_decl_state): New function.
394 (lto_free_function_in_decl_state_for_node): New function.
396 2013-08-29 Xinliang David Li <davidxl@google.com>
398 * loop-unroll.c (report_unroll_peel): Minor message change.
399 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
400 Emit alignment peeling message with default -fopt-info.
401 (vect_loop_versioning): Emit loop version info message.
402 * tree-vectorizer.c (vectorize_loops): Minor message change.
403 (execute_vect_slp): Ditto.
405 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
407 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
408 of the clone from the DECL_NAME of the original function.
410 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
412 * passes.c (register_pass): Add overload.
413 * tree-pass.h (register_pass): Forward declare it. Add comment.
415 2013-08-29 Jan Hubicka <jh@suse.cz>
417 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
418 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
419 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
420 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
421 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
422 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
423 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
424 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
426 2013-08-29 Teresa Johnson <tejohnson@google.com>
428 * dumpfile.c (dump_loc): Output column number.
429 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
430 * doc/invoke.texi: Document optall -fopt-info flag.
431 * profile.c (read_profile_edge_counts): Use new dump framework.
432 (compute_branch_probabilities): Ditto.
433 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
434 when pass not in any opt group.
435 * pass_manager.h (pass_manager::get_pass_profile): New method.
436 * value-prof.c (check_counter): Use new dump framework.
437 (check_ic_target): Ditto.
438 * coverage.c (get_coverage_counts): Ditto.
439 (coverage_init): Setup new dump framework.
441 2013-08-29 Richard Biener <rguenther@suse.de>
443 PR tree-optimization/58246
444 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
445 handle the dominance check inside a basic-block.
447 2013-08-29 Richard Biener <rguenther@suse.de>
450 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
451 of default defs that appear in abnormal PHI nodes.
453 2013-08-29 Richard Biener <rguenther@suse.de>
455 PR tree-optimization/57685
456 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
457 single-use operands to avoid exponential complexity.
459 2013-08-28 Dehao Chen <dehao@google.com>
461 * ipa-inline.c (edge_badness): Fix integer underflow.
463 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
465 * gtm-builtins.def (_ITM_free): Declare leaf.
467 2013-08-28 Jakub Jelinek <jakub@redhat.com>
470 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
471 (*tls_local_dynamic_base_64_largepic): Likewise.
472 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
473 Remove predicate from call operand.
474 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
475 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
477 2013-08-28 Jeff Law <law@redhat.com>
479 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
480 checks for the number of predecessors and successors allowed.
481 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
482 which require copying a joiner block if there is a request which
483 is a subpath that requires no joiner block copying.
485 2013-08-28 Jan Hubicka <jh@suse.cz>
487 * lto-streamer-out.c (DFS_write_tree_body): Drop
488 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
489 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
490 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
491 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
492 Do not read DECL_ERROR_ISSUED.
493 (unpack_ts_decl_with_vis_value_fields): Do not read
495 (lto_input_ts_binfo_tree_pointers): Do not read
496 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
497 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
498 write DECL_ERROR_ISSUED..
499 (pack_ts_decl_with_vis_value_fields): Do not write
501 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
502 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
503 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
504 * tree.h (tree_decl_common): Update comment.
505 (DECL_ERROR_ISSUED): Remove.
507 2013-08-28 Jakub Jelinek <jakub@redhat.com>
510 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
512 2013-08-28 Jan Hubicka <jh@suse.cz>
514 * builtins.def (free): Declare leaf.
516 2013-08-27 David Malcolm <dmalcolm@redhat.com>
519 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
520 * configure: Regenerate.
522 2013-08-27 Martin Jambor <mjambor@suse.cz>
524 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
525 (ipa_ancestor_jf_data): Likewise.
526 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
527 (ipa_get_jf_pass_through_type_preserved): New function.
528 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
529 (ipa_get_jf_ancestor_type_preserved): New function.
530 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
531 (ipa_get_jf_ancestor_result): Likewise.
532 (propagate_vals_accross_pass_through): Use
533 ipa_get_jf_pass_through_result to do all the value mappings.
534 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
536 (ipa_set_jf_cst_copy): New function.
537 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
538 (ipa_set_jf_arith_pass_through): Likewise.
539 (ipa_set_ancestor_jf): Likewise.
540 (compute_complex_assign_jump_func): Set type_preserved instead of
542 (ipa_compute_jump_functions_for_edge): Likewise.
543 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
544 (update_jump_functions_after_inlining): Update type_preserved.
545 Explicitely create jump functions when combining one with pass_through.
546 (ipa_write_jump_function): Stream the type_preserved flags.
547 (ipa_read_jump_function): Likewise.
549 2013-08-27 Jakub Jelinek <jakub@redhat.com>
550 Aldy Hernandez <aldyh@redhat.com>
552 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
553 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
554 * function.h (struct function): Add has_force_vect_loops and
556 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
557 * gimple.c (gimple_build_omp_critical): Add KIND argument and
559 * gimple.def: Update CLAUSES comments.
560 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
561 (gimple_build_omp_for): Add argument to prototype.
562 (gimple_omp_for_kind): New.
563 (gimple_omp_for_set_kind): New.
564 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
565 GOVD_DATA_SHARE_CLASS.
566 (enum omp_region_type): Add ORT_SIMD.
567 (gimple_add_tmp_var): Handle ORT_SIMD.
568 (gimplify_var_or_parm_decl): Same.
569 (is_gimple_stmt): Same.
570 (omp_firstprivatize_variable): Same.
571 (omp_add_variable): Only use splay_tree_insert if lookup failed.
572 (omp_notice_variable): Handle ORT_SIMD.
573 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
574 (omp_check_private): Handle ORT_SIMD.
575 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
577 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
578 Handle OMP_CLAUSE_LASTPRIVATE.
579 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
581 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
582 (gimplify_expr): Handle OMP_SIMD.
583 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
584 (expand_GOMP_SIMD_VF): New.
585 (expand_GOMP_SIMD_LAST_LANE): New.
586 * internal-fn.def (GOMP_SIMD_LANE): New.
588 (GOMP_SIMD_LAST_LANE): New.
589 * omp-low.c: Include target.h.
590 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
592 (check_omp_nesting_restrictions): Same.
594 (lower_rec_simd_input_clauses): New.
595 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
597 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
598 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
599 (expand_omp_build_assign): New.
600 (expand_omp_for_init_counts): New.
601 (expand_omp_for_init_vars): New.
602 (extract_omp_for_update_vars): New.
603 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
604 and rewrite accordingly.
605 (expand_omp_simd): New.
606 (expand_omp_for): Use expand_omp_simd.
607 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
608 (lower_omp_for): Do not lower the body.
609 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
611 * tree-flow.h (find_omp_clause): Remove prototype.
612 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
613 forcing vectorization of the loop, or if flag_tree_vectorize.
614 (gate_tree_if_conversion): Similarly.
615 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
616 gimple_build_omp_for.
617 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
618 * tree-parloops (create_parallel_loop): Pass kind argument to
619 gimple_build_omp_for.
620 * tree-pretty-print.c (dump_omp_clause): Add cases for
621 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
622 OMP_CLAUSE__SIMDUID_.
623 (dump_generic_node): Handle OMP_SIMD.
624 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
625 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
626 unroll OMP_SIMD loops here.
627 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
628 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
630 (vect_analyze_data_refs): Handle simd loops.
631 * tree-vect-loop.c (vectorizable_live_operation): Handle
633 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
634 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
635 (vectorizable_load): Same.
636 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
637 (struct simduid_to_vf): New.
638 (simduid_to_vf::hash): New.
639 (simduid_to-vf::equal): New.
640 (struct simd_array_to_simduid): New.
641 (simd_array_to_simduid::hash): New.
642 (simd_array_to_simduid::equal): New.
643 (adjust_simduid_builtins): New.
644 (struct note_simd_array_uses_struct): New.
645 (note_simd_array_uses_cb): New.
646 (note_simd_array_uses): New.
647 (vectorize_loops): Handle simd hints and adjust simd builtins
649 * tree-vectorizer.h (struct _stmt_vec_info): Add
650 simd_lane_access_p field.
651 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
652 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
653 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
654 (omp_clause_code_name): Same.
655 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
656 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
657 * tree.def (OMP_SIMD): New entry.
658 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
659 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
660 (OMP_CLAUSE_DECL): Adjust range for new clauses.
661 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
662 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
663 (OMP_CLAUSE_LINEAR_STEP): New.
664 (OMP_CLAUSE_SAFELEN_EXPR): New.
665 (OMP_CLAUSE__SIMDUID__DECL): New.
666 (find_omp_clause): New prototype.
668 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
670 * config/i386/driver-i386.c (host_detect_local_cpu): Update
671 Haswell processor detection.
673 2013-08-27 Christian Widmer <shadow@umbrox.de>
676 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
677 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
678 AVX2 capable processors.
680 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
682 * config/aarch64/arm_neon.h: Replace all inline asm implementations
683 of vget_low_* with implementations in terms of other intrinsics.
685 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
688 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
689 values to -1, 0 and 1.
691 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
693 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
694 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
695 (<optab><mode>3_insn): Remove.
696 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
697 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
698 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
699 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
700 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
701 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
702 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
703 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
704 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
705 (ror<mode>3_insn): Likewise.
706 * config/aarch64/predicates.md (aarch64_simd_register): New.
708 2013-08-27 Richard Biener <rguenther@suse.de>
710 PR tree-optimization/57521
711 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
712 one edge is non-critical.
713 (find_phi_replacement_condition): Make sure to use a non-critical
714 edge. Cleanup and remove old bug workarounds.
715 (bb_postdominates_preds): Remove.
716 (if_convertible_loop_p_1): Do not compute post-dominators.
717 (combine_blocks): Do not free post-dominators.
718 (main_tree_if_conversion): Likewise.
719 (pass_data_if_conversion): Add TODO_verify_ssa.
721 2013-08-27 DJ Delorie <dj@redhat.com>
723 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
725 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
727 * function.c (assign_parm_find_data_types): Set passed_mode and
728 nominal_mode to the TYPE_MODE of nominal_type for the built
729 pointer type in case of the struct-pass-by-reference.
731 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
733 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
734 (UINT16_TYPE): Change default to "unsigned int".
736 * config/avr/avr.opt (mfract-convert-truncate): New option.
737 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
738 is set, round negative fractional integers according to n1169
739 when converting to integer types.
741 2013-08-26 Jan Hubicka <jh@suse.cz>
743 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
744 methods can not be called indirectly when their address is not taken.
746 2013-08-26 Jan Hubicka <jh@suse.cz>
748 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
751 2013-08-26 Jan Hubicka <jh@suse.cz>
753 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
756 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
758 * reload.c (find_valid_class): Allow classes that do not include
759 FIRST_PSEUDO_REGISTER - 1.
761 2013-08-26 Jan Hubicka <jh@suse.cz>
763 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
764 fix edge count/frequency when speculation failed; fix type check
767 2013-08-26 Jan Hubicka <jh@suse.cz>
769 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
771 2013-08-26 Jan Hubicka <jh@suse.cz>
773 * ipa-inline-transform.c (inline_transform): Be ready for basic block
774 to be changed by edge redirection.
776 2013-08-26 Jan Hubicka <jh@suse.cz>
778 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
779 formating; add sanity check.
780 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
781 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
782 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
783 (resolve_noninline_speculation): Update callee keys, too.
785 2013-08-26 Jan Hubicka <jh@suse.cz>
787 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
788 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
790 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
792 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
795 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
797 * config/i386/i386.c (ix86_debug_options): Remove prototype.
798 (x86_64_elf_select_section): Ditto.
799 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
801 (ix86_pass_by_reference): Ditto.
802 (output_set_got): Ditto.
803 (ix86_unary_operator_ok): Ditto.
804 (ix86_expand_builtin): Ditto.
806 2013-08-23 Jan Hubicka <jh@suse.cz>
808 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
810 2013-08-23 Jan Hubicka <jh@suse.cz>
812 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
813 (tree_decl_with_vis): Add FINAL field.
815 2013-08-23 Jeff Law <law@redhat.com>
817 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
818 the debugging dump when the expression is fully redundant.
820 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
822 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
823 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
824 * pretty-print.c (pp_formatted_text_data): Likewise.
825 (pp_write_text_to_stream): Likewise.
826 (pp_write_text_as_dot_label_to_stream): Likewise.
827 (pp_append_r): Likewise.
828 (pp_format): Likewise.
829 (pp_flush): Likewise.
830 (pp_clear_output_area): Likewise.
831 (pp_append_text): Likewise.
832 (pp_formatted_text): Likewise.
833 (pp_remaining_character_count_for_line): Likewise.
834 (pp_newline): Likewise.
835 (pp_character): Likewise.
836 (output_buffer::~output_buffer): Define.
837 (pretty_printer::~pretty_printer): Destruct output buffer.
838 * pretty-print.h (output_buffer::~output_buffer): Declare.
839 (pretty_printer::~pretty_printer): Declare virtual.
841 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
844 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
845 HOST_WIDE_INT_M1U): New macros.
846 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
847 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
848 unsigned -1 for lshift.
849 * cse.c (cse_insn): Likewise.
850 * double-int.c (rshift_double, lshift_double): Likewise.
851 * builtins.c (fold_builtin_bitop): Likewise.
852 * combine.c (force_to_mode): Likewise.
853 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
854 * simplify-rtx.c (simplify_const_unary_operation,
855 simplify_const_binary_operation): Likewise.
856 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
857 check_va_list_escapes): Likewise.
858 * rtlanal.c (nonzero_bits1): Likewise.
859 * expmed.c (expand_smod_pow2): Likewise.
860 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
862 2013-08-23 Jan Hubicka <jh@suse.cz>
864 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
865 as having address taken.
867 2013-08-23 Jan Hubicka <jh@suse.cz>
869 * ipa-utils.h (method_class_type): Declare.
870 * ipa-devirt.c (method_class_type): Export.
872 * cgraphunit.c (analyze_functions): Do basic devirtualization;
873 do not walk base classes of anonymous types.
875 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
877 PR rtl-optimization/58220
879 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
880 handle SEQUENCE insns properly.
882 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
884 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
886 (pp_newline_and_indent): Likewise.
887 (pp_separate_with): Likewise.
888 * pretty-print.c (pp_newline_and_flush): Define.
889 (pp_newline_and_indent): Likewise.
890 (pp_separate_with): Likewise.
892 2013-08-23 Jakub Jelinek <jakub@redhat.com>
895 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
896 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
898 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
900 * gcc/config/i386/predicates.md (ext_sse_reg_operand): New.
901 * gcc/config/i386/i386.md (*movti_internal): Use
902 predicate to determine if EVEX is needed.
903 (*movsi_internal): Ditto.
904 (*movdf_internal): Ditto.
905 (*movsf_internal): Ditto.
906 * gcc/config/i386/mmx.md (*mov<mode>_internal): Ditto.
908 2013-08-23 Jakub Jelinek <jakub@redhat.com>
910 PR tree-optimization/58209
911 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
912 (find_tail_calls): Give up for pointer result types if m is non-NULL.
913 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
914 emit POINTER_PLUS_EXPR.
915 (create_tailcall_accumulator): For pointer result type accumulate in
918 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
920 * configure.ac: Add backslashes missing from the last change.
921 * configure: Regenerate.
923 2013-08-22 Jan Hubicka <jh@susue.cz>
925 * ipa.c (function_and_variable_visibility): First remember function
926 was global and then make it local.
928 2013-08-22 Julian Brown <julian@codesourcery.com>
930 * configure.ac: Add aarch64 to list of arches which use "nop" in
932 * configure: Regenerate.
934 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
936 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
937 gnu_libc_has_function.
938 * config/s390/tpf.h: Likewise.
940 2013-08-22 Jan Hubicka <jh@susue.cz>
942 * timevar.c (validate_phases): Add cast.
944 2013-08-22 Jan Hubicka <jh@susue.cz>
946 * timevar.c (validate_phases): Use size_t for memory.
947 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
949 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
951 * pretty-print.h (output_buffer::output_buffer): Declare.
952 (pretty_printer::pretty_printer): Likewise.
953 (pp_construct): Remove.
954 * pretty-print.c (output_buffer::output_buffer): Define.
955 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
956 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
957 (print_gimple_expr): Likewise.
958 (print_gimple_seq): Likewise.
959 (gimple_dump_bb): Likewise.
960 * sched-vis.c (dump_value_slim): Likewise.
961 (dump_insn_slim): Likewise.
962 (dump_rtl_slim): Likewise.
963 (str_pattern_slim): Likewise.
964 * tree-mudflap.c (mf_varname_tree): Likewise.
965 * graph.c (print_graph_cfg): Likewise.
966 (start_graph_dump): Likewise.
967 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
969 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
970 pretty printer initialization.
971 * coretypes.h (diagnostic_context): Remove superflous type alias
973 (pretty_printer): Likewise. Declare directly as a class.
974 (pretty_print_info): Remove declaration as class.
975 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
976 and pp_clear_output_area.
977 (asan_add_global): Likewise.
979 2013-08-22 Jan Hubicka <jh@suse.cz>
981 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
982 * ipa-utils.h (update_type_inheritance_graph): Declare.
983 (possible_polymorphic_call_target_p): Declare.
984 (possible_polymorphic_call_target_p): New.
985 * ipa-devirt.c: Update toplevel comments.
986 (cached_polymorphic_call_targets): Move up.
987 (odr_type_d): Move ID down.
988 (polymorphic_type_binfo_p): Update comment.
989 (odr_hasher::remove): Likewise;
990 (get_odr_type): Set anonymous_namespace.
991 (dump_odr_type): Dump it.
992 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
993 (maybe_record_node): Record node in cached_polymorphic_call_targets.
994 (record_binfo): Add comment.
995 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
997 (devirt_node_removal_hook): Do not iCE when cache is freed.
998 (possible_polymorphic_call_target_p): New predicate.
999 (update_type_inheritance_graph): New function.
1001 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
1002 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1003 Sergey Lega <sergey.s.lega@intel.com>
1004 Anna Tikhonova <anna.tikhonova@intel.com>
1005 Ilya Tocar <ilya.tocar@intel.com>
1006 Andrey Turetskiy <andrey.turetskiy@intel.com>
1007 Ilya Verbin <ilya.verbin@intel.com>
1008 Kirill Yukhin <kirill.yukhin@intel.com>
1009 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1011 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
1012 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
1013 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
1014 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
1015 (OPTION_MASK_ISA_AVX2_UNSET): Update.
1016 (OPTION_MASK_ISA_AVX512F_UNSET): New.
1017 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
1018 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
1019 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
1020 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
1021 OPT_mavx512pf, OPT_mavx512er cases.
1022 * config/i386/constraints.md (v): New constraint.
1023 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
1024 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
1025 (bit_AVX512CD): New.
1026 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
1027 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
1028 * config/i386/i386-c.c (ix86_target_macros_internal):
1029 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
1031 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
1032 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
1033 * config/i386/i386.c (regclass_map, dbx_register_map)
1034 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
1035 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
1036 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
1038 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
1039 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
1040 -mavx512pf options. Fix formatting.
1041 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
1042 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
1043 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
1044 -mavx512cd, -mavx512pf options.
1045 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
1046 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
1047 (ix86_preferred_output_reload_class): Replace SSE_REGS with
1049 (ix86_hard_regno_mode_ok): Support 512-bit registers.
1050 (ix86_set_reg_reg_cost): Ditto.
1051 (x86_order_regs_for_local_alloc): Ditto.
1052 (MAX_VECT_LEN): Extend to 64-byte.
1053 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
1054 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
1055 (TARGET_AVX512ER, TARGET_AVX512CD): New.
1056 (BIGGEST_ALIGNMENT): Extend to 512-bits.
1057 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
1058 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
1059 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
1060 (SSE_REG_MODE_P): Support new modes.
1061 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
1062 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
1063 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
1064 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
1065 (REG_CLASS_CONTENTS): Add new registers.
1066 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
1067 (EXT_REX_SSE_REGNO_P): New.
1068 (HI_REGISTER_NAMES): Add new registers.
1069 * config/i386/i386.md: Define constants for new registers.
1070 (mode): Add new 512-bit modes.
1071 (prefix): Support evex prefix.
1072 (isa): Support avx512f, noavx512f, fma_avx512f.
1073 (ssemodesuffix): Add new 512-bit modes.
1075 (*movxi_internal_avx512f): Ditto.
1076 (*movdi_internal): Replace constraint "x" with the new constraint "v".
1078 (*movsi_internal): Likewise.
1079 (*movdf_internal): Likewise.
1080 (*movsf_internal): Likewise.
1081 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
1082 (<code><mode>3): Likewise.
1083 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
1085 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
1086 with the new constraint "v".
1087 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
1089 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
1090 with the new constraint "v".
1091 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
1092 (<sse2>_storedqu<avxsizesuffix>): Likewise.
1093 (*<plusminus_insn><mode>3): Likewise.
1094 (<sse>_vm<plusminus_insn><mode>3): Likewise.
1095 (*mul<mode>3): Likewise.
1096 (<sse>_vmmul<mode>3): Likewise.
1097 (<sse>_div<mode>3): Likewise.
1098 (<sse>_vmdiv<mode>3): Likewise.
1099 (<sse>_sqrt<mode>2): Likewise.
1100 (<sse>_vmsqrt<mode>2): Likewise.
1101 (*<code><mode>3_finite): Likewise.
1102 (*<code><mode>3) <smaxmin>: Likewise.
1103 (<sse>_vm<code><mode>3): Likewise.
1104 (*<code><mode>3) <any_logic>: Likewise.
1105 (*fma_fmadd_<mode>): Likewise.
1106 (*fma_fmsub_<mode>): Likewise.
1107 (*fma_fnmadd_<mode>): Likewise.
1108 (*fma_fnmsub_<mode>): Likewise.
1109 (*fma_fmaddsub_<mode>): Likewise.
1110 (*fma_fmsubadd_<mode>): Likewise.
1111 (*fmai_fmadd_<mode>): Likewise.
1112 (*fmai_fmsub_<mode>): Likewise.
1113 (*fmai_fnmadd_<mode>): Likewise.
1114 (*fmai_fnmsub_<mode>): Likewise.
1115 (sse_cvtsi2ss): Likewise.
1116 (sse_cvtsi2ssq): Likewise.
1117 (sse_cvtss2si): Likewise.
1118 (sse_cvtss2si_2): Likewise.
1119 (sse_cvtss2siq): Likewise.
1120 (sse_cvtss2siq_2): Likewise.
1121 (sse_cvttss2si): Likewise.
1122 (sse_cvtss2siq_2): Likewise.
1123 (float<sseintvecmodelower><mode>2): Likewise.
1124 (sse2_cvtsd2si_2): Likewise.
1125 (sse2_cvtsd2siq_2): Likewise.
1126 (*<plusminus_insn><mode>3): Likewise.
1127 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
1128 (*<sse4_1_avx2>_mul<mode>3): Likewise.
1129 (ashr<mode>3): Likewise.
1130 (<shift_insn><mode>3): Likewise.
1131 (avx2_<code><mode>3): Likewise.
1132 (*avx2_<code><mode>3): Likewise.
1133 (*andnot<mode>3): Likewise.
1134 (*<code><mode>3) <any_logic>: Likewise.
1135 (abs<mode>2): Likewise.
1136 (avx2_permvar<mode>): Likewise.
1137 (avx2_perm<mode>_1): Likewise.
1138 (*avx_vpermilp<mode>): Likewise.
1139 (avx_vpermilvar<mode>3): Likewise.
1140 (avx2_ashrv<mode>): Likewise.
1141 (avx2_<shift_insn>v<mode>): Likewise.
1142 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
1144 * doc/rtl.texi: Document XImode.
1146 2013-08-21 Jeff Law <law@redhat.com>
1148 * tree-flow.h (register_jump_thread): Pass vector of edges
1149 instead of each important edge.
1150 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
1151 thread path into a vector and pass that to register_jump_thread.
1152 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
1153 passed in edge vector to the current 3-edge form.
1156 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1158 * dce.c (fini_dce): Call df_analyze again just in case
1159 delete_unmarked_insns removed anything.
1161 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
1163 * reload.h (struct reg_equivs): Rename to ..
1164 (struct reg_equivs_s): .. this.
1166 2013-08-20 Martin Liska <marxin.liska@gmail.com>
1168 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
1170 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1172 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
1174 2013-08-21 Jeff Law <law@redhat.com>
1176 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
1177 simplify assignments too. If the RHS collapses to a singleton
1178 range, then return the value for the range.
1180 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
1182 * config/i386/sse.md (V16): Rename to...
1184 (mov<mode>): Update iterator name.
1185 (*mov<mode>_internal): Ditto.
1186 (push<mode>1): Ditto.
1187 (movmisalign<mode>): Ditto.
1189 2013-08-20 Jan Hubicka <jh@suse.cz>
1192 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
1193 entry for direct edges.
1194 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
1196 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1198 Revert my last two changes, r201865 and r201864:
1201 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1203 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1204 instances can own GC refs.
1206 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1207 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1208 (gcc::context::gt_pch_nx): Likewise.
1209 (gcc::context::gt_pch_nx): Likewise.
1210 * ggc.h (gt_ggc_mx <T>): New.
1211 (gt_pch_nx_with_op <T>): New.
1212 (gt_pch_nx <T>): New.
1213 * passes.c (opt_pass::gt_ggc_mx): New.
1214 (opt_pass::gt_pch_nx): New.
1215 (opt_pass::gt_pch_nx_with_op): New.
1216 (pass_manager::gt_ggc_mx): New.
1217 (pass_manager::gt_pch_nx): New.
1218 (pass_manager::gt_pch_nx_with_op): New.
1219 (pass_manager::operator new): Use
1220 ggc_internal_cleared_alloc_stat rather than xcalloc.
1221 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1222 (pass_manager::gt_ggc_mx): New.
1223 (pass_manager::gt_pch_nx): New.
1224 (pass_manager::gt_pch_nx_with_op): New.
1225 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1226 (opt_pass::operator new): New.
1227 (opt_pass::gt_ggc_mx): New.
1228 (opt_pass::gt_pch_nx): New.
1229 (opt_pass::gt_pch_nx_with_op): New.
1232 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1234 * Makefile.in (GTFILES): Add context.h.
1235 * context.c (gcc::context::operator new): New.
1236 (gcc::context::gt_ggc_mx): New.
1237 (gcc::context::gt_pch_nx): New.
1238 (gcc::context::gt_pch_nx): New.
1239 * context.h (gcc::context): Add GTY((user)) marking.
1240 (gcc::context::operator new): New.
1241 (gcc::context::gt_ggc_mx): New.
1242 (gcc::context::gt_pch_nx): New.
1243 (gcc::context::gt_pch_nx): New.
1244 (g): Add GTY marking.
1245 (gt_ggc_mx (gcc::context *)): New.
1246 (gt_pch_nx (gcc::context *)): New.
1247 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1248 void *cookie)): New.
1249 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1251 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
1253 * dce.c (fini_dce): Call df_analyze again just in case
1254 delete_unmarked_insns removed anything.
1256 2013-08-20 Teresa Johnson <tejohnson@google.com>
1258 PR rtl-optimizations/57451
1259 * final.c (reemit_insn_block_notes): Prevent lexical blocks
1260 from crossing split section boundaries.
1262 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1264 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
1265 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
1266 with MULTLIB_DEFAULTS.
1268 2013-08-20 Nick Clifton <nickc@redhat.com>
1270 * target.def (narrow_volatile_bitfield): Note that the default
1271 value is false, not !TARGET_STRICT_ALIGN.
1272 * doc/tm.texi: Regenerate.
1274 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
1276 Fix LIB_SPEC for systems without libpthread.
1278 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
1279 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
1281 * config/i386/linux-common.h: Likewise.
1282 * config/mips/linux-common.h: Likewise.
1284 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1286 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
1289 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1291 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
1292 instances can own GC refs.
1294 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
1295 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
1296 (gcc::context::gt_pch_nx): Likewise.
1297 (gcc::context::gt_pch_nx): Likewise.
1298 * ggc.h (gt_ggc_mx <T>): New.
1299 (gt_pch_nx_with_op <T>): New.
1300 (gt_pch_nx <T>): New.
1301 * passes.c (opt_pass::gt_ggc_mx): New.
1302 (opt_pass::gt_pch_nx): New.
1303 (opt_pass::gt_pch_nx_with_op): New.
1304 (pass_manager::gt_ggc_mx): New.
1305 (pass_manager::gt_pch_nx): New.
1306 (pass_manager::gt_pch_nx_with_op): New.
1307 (pass_manager::operator new): Use
1308 ggc_internal_cleared_alloc_stat rather than xcalloc.
1309 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
1310 (pass_manager::gt_ggc_mx): New.
1311 (pass_manager::gt_pch_nx): New.
1312 (pass_manager::gt_pch_nx_with_op): New.
1313 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
1314 (opt_pass::operator new): New.
1315 (opt_pass::gt_ggc_mx): New.
1316 (opt_pass::gt_pch_nx): New.
1317 (opt_pass::gt_pch_nx_with_op): New.
1319 2013-08-20 David Malcolm <dmalcolm@redhat.com>
1321 * Makefile.in (GTFILES): Add context.h.
1322 * context.c (gcc::context::operator new): New.
1323 (gcc::context::gt_ggc_mx): New.
1324 (gcc::context::gt_pch_nx): New.
1325 (gcc::context::gt_pch_nx): New.
1326 * context.h (gcc::context): Add GTY((user)) marking.
1327 (gcc::context::operator new): New.
1328 (gcc::context::gt_ggc_mx): New.
1329 (gcc::context::gt_pch_nx): New.
1330 (gcc::context::gt_pch_nx): New.
1331 (g): Add GTY marking.
1332 (gt_ggc_mx (gcc::context *)): New.
1333 (gt_pch_nx (gcc::context *)): New.
1334 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
1335 void *cookie)): New.
1336 * gengtype.c (open_base_files) <ifiles>: Add context.h.
1338 2013-08-20 Alan Modra <amodra@gmail.com>
1341 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
1342 (rs6000_emit_epilogue): Likewise.
1344 2013-08-19 Dehao Chen <dehao@google.com>
1346 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
1348 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
1349 Jakub Jelinek <jakub@redhat.com>
1351 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
1352 (BUILT_IN_FABSD64): Likewise.
1353 (BUILT_IN_FABSD128): Likewise.
1354 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
1355 (fold_builtin_1): Likewise.
1356 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
1357 destination and source operands.
1358 (*abstd2_fpr): Likewise.
1359 (*nabstd2_fpr): Likewise.
1361 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
1363 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
1366 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
1368 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
1371 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
1373 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
1374 * builtins.c (default_libc_has_function): New.
1375 (gnu_libc_has_function): Ditto.
1376 (no_c99_libc_has_function): Ditto.
1377 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
1378 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
1379 (fold_builtin_sincos): Likewise.
1380 (fold_builtin_cexp): Likewise.
1381 * builtins.def (DEF_C94_BUILTIN): Likewise.
1382 (DEF_C99_BUILTIN): Likewise.
1383 (DEF_C99_C90RES_BUILTIN): Likewise.
1384 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
1385 definitions to using this define.
1386 * config/darwin-protos.h (darwin_libc_has_function): New.
1387 * config/darwin.c (darwin_libc_has_function): Ditto.
1388 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
1389 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
1390 * config/darwin.h: Ditto.
1391 * config/elfos.h: Ditto.
1392 * config/freebsd.h: Ditto.
1393 * config/i386/cygming.h: Ditto.
1394 * config/i386/djgpp.h: Ditto.
1395 * config/i386/i386-interix.h: Ditto.
1396 * config/microblaze/microblaze.h: Ditto.
1397 * config/mmix/mmix.h: Ditto.
1398 * config/gnu-user.h: Ditto.
1399 * config/ia64/hpux.h: Ditto.
1400 * config/pa/pa-hpux.h: Ditto.
1401 * config/pdp11/pdp11.h: Ditto.
1402 * config/picochip/picochip.h: Ditto.
1403 * config/linux.h: Ditto.
1404 * config/netbsd.h: Ditto.
1405 * config/openbsd.h: Ditto.
1406 * config/rs6000/aix43.h: Ditto.
1407 * config/rs6000/aix51.h: Ditto.
1408 * config/rs6000/aix52.h: Ditto.
1409 * config/rs6000/aix53.h: Ditto.
1410 * config/rs6000/aix61.h: Ditto.
1411 * config/rs6000/darwin.h: Ditto.
1412 * config/rs6000/linux.h: Ditto.
1413 * config/rs6000/linux64.h: Ditto.
1414 * config/s390/tpf.h: Ditto.
1415 * config/sol2-10.h: Ditto.
1416 * config/sol2.h: Ditto.
1417 * config/vms/vms.h: Ditto.
1418 * config/vxworks.h: Ditto.
1419 * config/linux-android.c (linux_android_libc_has_function):
1420 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
1421 * config/linux-protos.h (linux_android_libc_has_function):
1423 * config/i386/i386.c (ix86_libc_has_function): New.
1424 * config/i386/i386-protos.h
1425 (ix86_libc_has_function): New declaration.
1426 * config/i386/i386.md
1427 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
1428 ("isinf<mode>2): Likewise.
1429 * convert.c (convert_to_integer): Using new target hook
1430 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
1431 TARGET_C99_FUNCTIONS.
1432 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
1433 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
1434 * coretypes.h (function_class): New enum for different
1435 classes of functions.
1436 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
1437 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
1438 (TARGET_HAS_SINCOS): Likewise.
1439 (TARGET_LIBC_HAS_FUNCTION): New.
1440 * doc/tm.texi: Regenerated.
1441 * targhooks.h (default_libc_has_function): New declaration.
1442 (no_c99_libc_has_function): Ditto.
1443 (gnu_libc_has_function): Ditto.
1444 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
1445 and TARGET_HAS_SINCOS.
1447 2013-08-18 Jan Hubicka <jh@suse.cz>
1449 * Makeifle-in (ipa-devirt.o): New.
1450 (GTFILES): Add ipa-utils.h and ipa-devirt.c
1451 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
1452 (analyze_functions): Look into possible targets of polymorphic call.
1453 * dumpfile.c (dump_files): Add type-inheritance dump.
1454 * dumpfile.h (TDI_inheritance): New.
1455 * ipa-devirt.c: New file.
1456 * ipa-utils.h (odr_type_d): Forward declare.
1457 (odr_type): New type.
1458 (build_type_inheritance_graph): Declare.
1459 (possible_polymorphic_call_targets): Declare and introduce inline
1460 variant when only edge is pased.
1461 (dump_possible_polymorphic_call_targets): Likewise.
1462 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
1463 * tree.c (type_in_anonymous_namespace_p): Break out from ...
1464 (types_same_for_odr): ... here.
1465 * tree.h (type_in_anonymous_namespace_p): Declare.
1467 2013-08-18 Jakub Jelinek <jakub@redhat.com>
1469 PR tree-optimization/58006
1470 * tree-parloops.c (take_address_of): Don't ICE if get_name
1472 (eliminate_local_variables_stmt): Remove clobber stmts.
1474 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
1476 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
1477 error message is issued for an alias to undefined symbol.
1479 2013-08-18 Jan Hubicka <jh@suse.cz>
1481 * cgraph.c (cgraph_create_indirect_edge): Discover
1482 polymorphic calls and record basic info into indirect_info.
1483 * gimple-fold.c (gimple_fold_call): When doing BINFO based
1484 devirtualization, ignore objc function calls.
1485 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
1486 call with no parm index info.
1487 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
1488 * tree.c (virtual_method_call_p): New function.
1489 * tree.h (virtual_method_call_p): Declare.
1491 2013-08-16 Jan Hubicka <jh@suse.cz>
1494 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
1496 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
1498 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
1501 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1503 * gengtype.c (type_for_name): Add special-case support for
1504 locating types within the "gcc::" namespace.
1505 (open_base_files): Emit a "using namespace gcc" directive.
1507 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1510 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
1511 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
1513 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
1514 array instead of each individual operand as a separate argument.
1515 (emit_fusion_gpr_load): Likewise.
1516 (expand_fusion_gpr_load): Add new function declaration.
1518 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
1519 signature to have the operands passed as an array, instead of as
1520 separate arguments. Allow ZERO_EXTEND to be in the memory
1521 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
1522 depend on the register live/dead flags when peepholes are run.
1523 (expand_fusion_gpr_load): New function to be called from the
1524 peephole2 pass, to change the register that addis sets to be the
1526 (emit_fusion_gpr_load): Change the calling signature to have the
1527 operands passed as an array, instead of as separate arguments.
1528 Allow ZERO_EXTEND to be in the memory address, and also
1529 SIGN_EXTEND if -mpower8-fusion-sign.
1531 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
1533 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
1534 adjust the register addis loads up in the peephole2 pass. Do not
1535 depend on the register live/dead state when the peephole pass is done.
1537 2013-08-16 David Malcolm <dmalcolm@redhat.com>
1539 * gengtype.c (create_user_defined_type): Ensure that the kind
1540 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
1541 declaration is seen before the GTY((user)) marking.
1543 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
1546 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
1548 2013-08-16 Jan Hubicka <jh@suse.cz>
1550 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
1551 arugment expected_type.
1552 (gimple_fold_call): Use it.
1553 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
1554 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
1555 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
1556 (try_make_edge_direct_virtual_call): Likewise.
1557 * tree.c (obj_type_ref_class): New.
1558 * tree.h (obj_type_ref_class): Use it.
1560 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
1562 * sched-vis.c (rtl_slim_pp_initialized): Remove.
1563 (rtl_slim_pp): Likewise.
1564 (init_rtl_slim_pretty_print): Likewise.
1565 (dump_value_slim): Don't call it. Use local pretty printer.
1566 (dump_insn_slim): Likewise.
1567 (dump_rtl_slim): Likewise.
1568 (str_pattern_slim): Likewise.
1569 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
1572 2013-08-16 Jakub Jelinek <jakub@redhat.com>
1574 PR tree-optimization/58164
1575 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
1576 walk gimple_goto_dest of GIMPLE_GOTO.
1578 PR tree-optimization/58165
1579 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
1580 bi_call must be the last stmt in a bb, don't split_block, instead
1581 use fallthru edge from it and give up if there is none.
1582 Release conds vector when returning early.
1584 2013-08-14 Xinliang David Li <davidxl@google.com>
1586 * config/i386/i386.c (ix86_option_override_internal):
1587 Remove unused variable and field.
1589 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1592 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
1593 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
1594 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
1595 and Linux, correct BLKmode alignment when 128-bit alignment is
1596 required and compatibility flag is not set.
1597 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
1598 for zero-size arguments when compatibility flag is not set.
1600 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1602 PR tree-optimization/58145
1603 * tree-sra.c (build_ref_for_offset): If prev_base has
1604 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
1606 2013-08-14 Xinliang David Li <davidxl@google.com>
1608 * config/i386/i386.c (ix86_option_override_internal):
1609 Fix uninitialized variable error.
1611 2013-08-14 Xinliang David Li <davidxl@google.com>
1613 * config/i386/i386.opt: Define two new options.
1614 * config/i386/x86-tune.def: Add arch selector field in macros.
1615 * config/i386/i386.h: Adjust macro definition.
1616 * config/i386/i386.c (ix86_option_override_internal):
1618 (parse_mtune_ctrl_str): New function.
1619 (set_ix86_tune_features): New function.
1620 (ix86_function_specific_restore): Call the new helper function.
1622 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
1624 PR rtl-optimization/57662
1625 * sel-sched.c (code_motion_process_successors): When the current insn
1626 is removed after the recursive traversal, break from the loop.
1627 Add comments and debug printouts.
1629 2013-08-14 Jakub Jelinek <jakub@redhat.com>
1630 Alexandre Oliva <aoliva@redhat.com>
1633 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
1634 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
1635 there also UNSPEC_PLTOFF.
1637 2013-08-14 Marek Polacek <polacek@redhat.com>
1639 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
1640 NUM_CONDITIONS bit positions.
1642 2013-08-13 Cary Coutant <ccoutant@google.com>
1644 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
1645 (attr_checksum): Hash vector contents instead of pointer.
1646 (attr_checksum_ordered): Likewise.
1648 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
1650 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
1651 when Pmode != word_mode. Add length_address attribute.
1652 (sse3_monitor_<mode>): Merge from sse3_monitor and
1653 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
1654 Pmode != word_mode. Update insn length attribute.
1655 * config/i386/i386.c (ix86_option_override_internal): Update
1656 ix86_gen_monitor selection for merged sse3_monitor insn.
1658 2013-08-13 Julian Brown <julian@codesourcery.com>
1660 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
1661 perform invalid legitimization on greater-than-word-size modes for
1664 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
1666 * ira.c (setup_class_translate_array): Use aclass instead of cl
1667 for classes not fully covered by allocno classes.
1669 2013-08-13 Jakub Jelinek <jakub@redhat.com>
1671 PR tree-optimization/57661
1672 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
1673 * tree-inline.c (tree_function_versioning): Initialize it.
1674 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
1675 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
1676 that is not being copied.
1679 * asan.c (instrument_strlen_call): Fix typo in comment.
1680 Use char * type even for the lhs of POINTER_PLUS_EXPR.
1682 2013-08-13 Steve Ellcey <sellcey@mips.com>
1684 * config/mips/mips.md (prefetch): Use lw instead of ld on
1685 loongson in 32bit mode.
1687 2013-08-13 Nick Clifton <nickc@redhat.com>
1689 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
1691 2013-08-13 Jan Hubicka <jh@suse.cz>
1693 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
1694 introduced edge; fix typo in sanity check.
1695 (cgraph_resolve_speculation): Export; improve diagnostic.
1696 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
1697 speculation at type mismatch.
1698 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
1699 (cgraph_resolve_speculation): Declare.
1700 (symtab_can_be_discarded): New function.
1701 * value-prof.c (gimple_ic_transform): Remove actual transform code.
1702 * ipa-inline-transform.c (speculation_removed): New global var.
1703 (clone_inlined_nodes): See if speculation can be removed.
1704 (inline_call): If speculations was removed, we growths may not match.
1705 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
1706 (speculation_useful_p): New function.
1707 (resolve_noninline_speculation): New function.
1708 (inline_small_functions): Resolve useless speculations.
1709 * ipa-inline.h (speculation_useful_p): Declare
1710 * ipa.c (can_replace_by_local_alias): Simplify.
1711 (ipa_profile): Produce speculative calls in non-lto, too;
1712 add simple cost model; produce local aliases.
1714 2013-08-13 David Malcolm <dmalcolm@redhat.com>
1716 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
1719 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
1721 * config/i386/i386.c (ix86_function_versions): Use error + inform.
1723 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
1725 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
1726 iterator instead of X87MODEF.
1728 2013-08-12 Perez Read <netfirewall@gmail.com>
1731 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
1732 operand 0 for intel asm alternative.
1733 (*movabs<mode>_2): Ditto for operand 1.
1735 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
1737 * config/aarch64/arm_none.h
1738 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
1740 2013-08-12 Nick Clifton <nickc@redhat.com>
1742 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
1744 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
1746 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
1747 expand for QI/HImode operand to produce more effictive code for
1748 unsigned char(short) --> float(double) conversion.
1750 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
1752 * doc/invoke.texi: Mention that -ftls-model does not force the final
1755 2013-08-12 Marek Polacek <polacek@redhat.com>
1756 Marc Glisse <marc.glisse@inria.fr>
1758 PR tree-optimization/57980
1759 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
1760 when creating -1 constant.
1762 2013-08-10 Jan Hubicka <jh@suse.cz>
1764 Workaround binutils PR14342.
1765 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
1766 (gimple_init_edge_profiler): Likewise.
1767 (gimple_gen_ic_func_profiler): Likewise.
1769 2013-08-09 Jan Hubicka <jh@suse.cz>
1771 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
1773 2013-08-09 Xinliang David Li <davidxl@google.com>
1775 * config/i386/stringop.def: New file.
1776 * config/i386/stringop.opt: New file.
1777 * config/i386/i386-opts.h: Include stringopt.def.
1778 * config/i386/i386.opt: Include stringopt.opt.
1779 * config/i386/i386.c (ix86_option_override_internal):
1780 Override default size based stringop inline strategies with options.
1781 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
1784 2013-08-09 Jan Hubicka <jh@suse.cz>
1786 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
1788 2013-08-09 Jan Hubicka <jh@suse.cz>
1790 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
1792 (dump_cgraph_node): Dump profile-id.
1793 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
1794 and common_target_probability.
1795 * lto-cgraph.c (lto_output_edge): Stream common targets.
1796 (lto_output_node): Stream profile ids.
1797 (input_node): Stream profile ids.
1798 (input_edge): Stream common targets.
1799 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
1800 * ipa.c: Include value-prof.h
1801 (ipa_profile_generate_summary): Turn indirect call statement histograms
1802 into common targets.
1803 (ipa_profile): Turn common targets into speculative edges.
1805 2013-08-09 Jan Hubicka <jh@suse.cz>
1807 * cgraph.h (cgraph_node): Add profile_id.
1808 * value-prof.c (cgraph_node_map): Turn into pointer_map.
1809 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
1810 (del_node_map): Update.
1811 (find_func_by_funcdef_no): Replace by ...
1812 (find_func_by_profile_id): ... this one.
1813 (gimple_ic_transform): Do not remove useful histograms when
1814 speculation is not done; dump info when indirect call removal
1816 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
1817 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
1818 (__gcov_indirect_call_profiler_v2): .. this one.
1819 * profile.h (init_node_map): Update.
1820 * coverage.c (coverage_compute_profile_id): New function.
1821 * coverage.h (coverage_compute_profile_id): Declare.
1822 * tree-profile.c (init_ic_make_global_vars): Make
1823 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
1824 (gimple_init_edge_profiler): Update prototype of
1825 __gcov_indirect_call_profiler.
1826 (gimple_gen_ic_func_profiler): Simplify.
1827 (tree_profiling): Use init_node_map
1829 2013-08-09 Jan Hubicka <jh@suse.cz>
1831 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
1832 non-speculative refs.
1833 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
1834 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
1835 (cgraph_set_call_stmt): Likewise.
1836 (cgraph_create_edge_1): Fix release checking compilatoin;
1838 (cgraph_free_edge): Free indirect info.
1839 (cgraph_turn_edge_to_speculative): New function.
1840 (cgraph_speculative_call_info): New function.
1841 (cgraph_make_edge_direct): Return direct edge; handle speculation.
1842 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
1843 (dump_cgraph_node): Dump speculation.
1844 (verify_edge_count_and_frequency): Accept speculative edges.
1845 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
1846 (verify_cgraph_node): Handle speculation.
1847 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
1848 (cgraph_set_call_stmt): Update prototype.
1849 (cgraph_make_edge_direct): Update prototype.
1850 (cgraph_speculative_call_info): Declare.
1851 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
1852 to change; update call of ipa_find_references.
1853 * ipa-ref.c (ipa_record_reference): Fix return value; clear
1854 lto_stmt_uid and speculative flags.
1855 (ipa_dump_references): Dump speculation.
1856 (ipa_clone_references): Clone speculative flag.
1857 (ipa_clone_referring): Likewise.
1858 (ipa_clone_ref): New function.
1859 (ipa_find_reference): Look into lto_stmt_uids
1860 (ipa_clear_stmts_in_references): Do not clear speculative calls.
1861 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
1862 (ipa_find_reference): Update declaration.
1863 (ipa_clone_ref): Declare.
1864 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
1865 stream speculative flag.
1866 (lto_output_ref): Stream statements uids and speculation.
1867 (input_ref): Likewise.
1868 (input_edge): Stream speuclation.
1869 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
1870 (cgraph_set_call_stmt_including_clones): Handle speculation.
1871 * ipa-inline.c (heap_edge_removal_hook): New function.
1872 (inline_small_functions): Register it.
1873 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
1874 also initialize refs.
1875 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
1877 (try_make_edge_direct_simple_call): Likewise.
1878 (try_make_edge_direct_simple_call): Likewise.
1879 (update_indirect_edges_after_inlining): Likewise.
1880 (remove_described_reference): Look proper lto_stmt_uid.
1881 (propagate_controlled_uses): Likewise.
1882 (propagate_controlled_uses): Liekwise.
1883 * tree-inline.c (copy_bb): Copy speculative edges.
1884 (redirect_all_calls): New function.
1885 (copy_cfg_body): Do redirection after loop info is updated.
1886 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
1888 2013-08-09 Jan Hubicka <jh@suse.cz>
1890 * lto-streamer-out.c (output_function): Renumber PHIs.
1891 * lto-streamer-in.c (input_function): Likewise.
1893 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
1895 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
1896 (get_lane_unsigned): Likewise.
1897 (dup_lane_scalar): Likewise.
1898 (get_lane): enable for VALL.
1899 * config/aarch64/aarch64-simd.md
1900 (aarch64_dup_lane_scalar<mode>): Remove.
1901 (aarch64_get_lane_signed<mode>): Likewise.
1902 (aarch64_get_lane_unsigned<mode>): Likewise.
1903 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
1904 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
1905 (aarch64_get_lane<mode>): Enable for all vector modes.
1906 (aarch64_get_lanedi): Remove misleading constraints.
1907 * config/aarch64/arm_neon.h
1908 (__aarch64_vget_lane_any): Define.
1909 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
1910 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
1911 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
1912 * config/aarch64/iterators.md (VDQQH): New.
1916 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
1918 * configure.ac: Add GAS check for LEON instructions on SPARC.
1919 * configure: Regenerate.
1920 * config.in: Likewise.
1921 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
1923 * config/sparc/sparc.opt (LEON, LEON3): New masks.
1924 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
1926 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
1927 (AS_LEON_FLAG): New macro.
1928 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
1929 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
1930 Deal with LEON and LEON3 for the memory model.
1931 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1932 (atomic_compare_and_swap<mode>_1): Likewise.
1933 (*atomic_compare_and_swap<mode>_1): Likewise.
1935 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1937 * config/arm/neon.md (vcond): Fix floating-point vector
1938 comparisons against 0.
1940 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
1942 * lra-constraints.c (emit_spill_move): Remove assert.
1943 (process_alt_operands): Add more debugging
1944 output. Increase reject for spilling into memory. Decrease
1945 reject for reloading scratch.
1946 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
1948 2013-08-08 Steve Ellcey <sellcey@mips.com>
1950 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
1951 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
1952 micromips incompatible. Add nan2008.
1953 (MULTILIB_DIRNAMES): Add nan2008.
1954 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1955 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
1956 and micromips incompatible. Add nan2008.
1957 (MULTILIB_DIRNAMES): Add nan2008.
1958 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
1960 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
1962 PR rtl-optimization/58079
1963 * combine.c (combine_simplify_rtx): Avoid using SUBST if
1964 simplify_comparison has widened a comparison with an integer.
1966 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1968 * config/arm/neon.md (movmisalign<mode>): Disable when we
1969 don't allow unaligned accesses.
1970 (*movmisalign<mode>_neon_store): Likewise.
1971 (*movmisalign<mode>_neon_load): Likewise.
1972 (*movmisalign<mode>_neon_store): Likewise.
1973 (*movmisalign<mode>_neon_load): Likewise.
1975 2013-08-08 Jan Hubicka <jh@suse.cz>
1977 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
1978 (make_pass_rebuild_cgraph_edges): Also clear references.
1979 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
1980 * ipa-inline-transform.c (inline_transform): Remove all references
1982 * cgraphunit.c (expand_function): Remove all references after
1984 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
1985 (ipa_find_reference): Rewrite to iterator.
1986 (remove_stmt_references): Likewise.
1987 (ipa_clear_stmts_in_references): New function.
1988 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
1989 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
1991 * ipa-split.c (split_function): Remove references in split function.
1993 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
1996 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
1997 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
1999 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
2002 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
2003 suggested mode for the assignment isn't compatible with the
2006 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
2009 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
2011 2013-08-07 Xinliang David Li <davidxl@google.com>
2013 * config/i386/i386.opt: New option -mtune-ctrl=.
2014 * config/i386/x86-tune.def: New file.
2015 * config/i386/i386.h: include x86-tune.def.
2016 * config/i386/i386.c (ix86_option_override_internal):
2017 Parsing -mtune-ctrl= option and set tune features.
2019 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
2022 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
2023 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
2026 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2028 * rtl.h (update_alignments): Declare.
2029 * final.c (grow_label_align): New function extracted from...
2030 (shorten_branches): ...here. Call it.
2031 (update_alignments): New function.
2032 * reorg.c (sibling_labels): New variable.
2033 (get_label_before): Add SIBLING parameter. If it is non-zero, push
2034 the new label along with it onto the sibling_labels vector.
2035 (fill_simple_delay_slots): Adjust call to get_label_before.
2036 (fill_slots_from_thread): Likewise.
2037 (relax_delay_slots): Likewise.
2038 (make_return_insns): Likewise.
2039 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
2041 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
2043 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
2044 document its semantics.
2045 (diagnostic_report_diagnostic): Adjust accordingly.
2047 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2049 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
2050 (sparc_option_override): ...and port to new C++ pass API.
2051 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
2053 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
2055 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
2057 2013-08-06 Caroline Tice <cmtice@google.com>
2059 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
2060 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
2061 * tree-pass.h: Add pass_vtable_verify.
2062 * varasm.c (assemble_variable): Add code to properly set the comdat
2063 section and name for the .vtable_map_vars section.
2064 (assemble_vtyv_preinit_initializer): New function.
2065 (default_sectin_type_flags): Make sure .vtable_map_vars section has
2067 * output.h: Add function decl for assemble_vtv_preinit_initializer.
2068 * vtable-verify.c: New file.
2069 * vtable-verify.h: New file.
2070 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
2071 initialiation levels.
2072 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
2073 * passes.def: Insert pass_vtable_verify.
2074 * aclocal.m4: Reorder includes.
2075 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
2076 -fvtv-counts options.
2077 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
2078 as appropriate, if -fvtable-verify=... is used.
2079 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
2080 -fvtable-verify=... is used.
2081 * Makefile.in (OBJS): Add vtable-verify.o to list.
2082 (vtable-verify.o): Add new build rule.
2083 (GTFILES): Add vtable-verify.c to list.
2084 * common.opt (fvtable-verify=): New flag.
2085 (vtv_priority): Values for fvtable-verify= flag.
2086 (fvtv-counts): New flag.
2087 (fvtv-debug): New flag.
2088 * tree.h (save_vtable_map_decl): New extern function decl.
2090 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2092 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
2093 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
2094 (pass_data_rl78_devirt): ...new pass_data instance and...
2095 (make_pass_rl78_devirt): ...new function.
2096 (rl78_asm_file_start): Port pass registration to new C++ API.
2098 2013-08-07 David Malcolm <dmalcolm@redhat.com>
2100 * coretypes.h (rtl_opt_pass): Add.
2101 (gcc::context): Add.
2102 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
2103 (epiphany_init): Port to new C++ pass API.
2104 (epiphany_optimize_mode_switching): Likewise.
2105 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
2106 (pass_manager::get_pass_mode_switching): New.
2107 (pass_manager::get_pass_peephole2): New.
2108 * mode-switching.c (pass_mode_switching): Add clone method.
2109 * recog.c (pass_peephole2): Add clone method.
2110 (pass_split_all_insns): Add clone method.
2112 2013-08-06 David Malcolm <dmalcolm@redhat.com>
2114 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
2115 (mips_option_override): ...here, porting to new C++ API for passes.
2117 2013-08-06 Jan Hubicka <jh@suse.cz>
2119 * cgraph.c (cgraph_get_body): New function based on lto.c
2121 * cgraph.h (cgraph_get_body): Declare.
2122 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
2124 * cgraphunit.c (expand_function): Get body prior expanding.
2125 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
2126 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
2128 * passes.c (do_per_function_toporder): Get body.
2129 * tree-inline.c (expand_call_inline): Get body prior inlining it.
2130 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
2132 2013-08-06 Martin Jambor <mjambor@suse.cz>
2135 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
2136 is not re-finalized. Rename second parameter to no_collect.
2138 2013-08-06 Martin Jambor <mjambor@suse.cz>
2141 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
2142 MEM_REF has proper alignment information.
2144 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
2147 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
2149 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
2150 rtx (*) (rtx, ...) with insn_gen_fn.
2151 * genoutput.c (output_insn_data): Cast gen_? function pointers to
2152 insn_gen_fn::stored_funcptr. Add initializer braces.
2154 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2156 Rewrite how instances of passes are cloned to remove assumptions
2157 about their sizes (thus allowing pass subclasses to have
2158 additional data fields, albeit non-GC-managed ones at this point).
2160 * passes.c (make_pass_instance): Now that passes have clone
2161 methods, rewrite this function to eliminate XNEW and memcpy
2162 calls that used hardcoded sizes. Since this function no longer
2163 creates pass instances, rename it to...
2164 (add_pass_instance): ...this. Document the old way that passes were
2165 numbered and flagged, and rework this function to continue using it.
2166 (next_pass_1): Add an initial_pass argument for use by
2168 (position_pass): When adding multiple instances of a pass, use
2169 the pass's clone method, rather than relying on the XNEW/memcpy
2170 within the former make_pass_instance (now add_pass_instance).
2171 (pass_manager::pass_manager): When invoking next_pass_1, also supply
2172 the initial instance of the current pass within the pass manager.
2174 2013-08-05 David Malcolm <dmalcolm@redhat.com>
2176 This is the automated part of the conversion of passes from C
2177 structs to C++ classes.
2179 Patch autogenerated by refactor_passes.py from
2180 https://github.com/davidmalcolm/gcc-refactoring-scripts
2181 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
2183 * asan.c (pass_asan): Convert from a global struct to a subclass of
2184 gimple_opt_pass along with...
2185 (pass_data_asan): ...new pass_data instance and...
2186 (make_pass_asan): ...new function.
2187 (pass_asan_O0): Convert from a global struct to a subclass of
2188 gimple_opt_pass along with...
2189 (pass_data_asan_O0): ...new pass_data instance and...
2190 (make_pass_asan_O0): ...new function.
2191 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
2192 subclass of rtl_opt_pass along with...
2193 (pass_data_inc_dec): ...new pass_data instance and...
2194 (make_pass_inc_dec): ...new function.
2195 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
2196 a subclass of rtl_opt_pass along with...
2197 (pass_data_reorder_blocks): ...new pass_data instance and...
2198 (make_pass_reorder_blocks): ...new function.
2199 (pass_duplicate_computed_gotos): Convert from a global struct to a
2200 subclass of rtl_opt_pass along with...
2201 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
2202 (make_pass_duplicate_computed_gotos): ...new function.
2203 (pass_partition_blocks): Convert from a global struct to a subclass of
2204 rtl_opt_pass along with...
2205 (pass_data_partition_blocks): ...new pass_data instance and...
2206 (make_pass_partition_blocks): ...new function.
2207 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
2208 struct to a subclass of rtl_opt_pass along with...
2209 (pass_data_branch_target_load_optimize1): ...new pass_data instance
2211 (make_pass_branch_target_load_optimize1): ...new function.
2212 (pass_branch_target_load_optimize2): Convert from a global struct to a
2213 subclass of rtl_opt_pass along with...
2214 (pass_data_branch_target_load_optimize2): ...new pass_data instance
2216 (make_pass_branch_target_load_optimize2): ...new function.
2217 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
2218 of rtl_opt_pass along with...
2219 (pass_data_jump): ...new pass_data instance and...
2220 (make_pass_jump): ...new function.
2221 (pass_jump2): Convert from a global struct to a subclass of
2222 rtl_opt_pass along with...
2223 (pass_data_jump2): ...new pass_data instance and...
2224 (make_pass_jump2): ...new function.
2225 * cfgexpand.c (pass_expand): Convert from a global struct to a
2226 subclass of rtl_opt_pass along with...
2227 (pass_data_expand): ...new pass_data instance and...
2228 (make_pass_expand): ...new function.
2229 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
2230 of rtl_opt_pass along with...
2231 (pass_data_free_cfg): ...new pass_data instance and...
2232 (make_pass_free_cfg): ...new function.
2233 (pass_into_cfg_layout_mode): Convert from a global struct to a
2234 subclass of rtl_opt_pass along with...
2235 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
2236 (make_pass_into_cfg_layout_mode): ...new function.
2237 (pass_outof_cfg_layout_mode): Convert from a global struct to a
2238 subclass of rtl_opt_pass along with...
2239 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
2240 (make_pass_outof_cfg_layout_mode): ...new function.
2241 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
2242 struct to a subclass of gimple_opt_pass along with...
2243 (pass_data_build_cgraph_edges): ...new pass_data instance and...
2244 (make_pass_build_cgraph_edges): ...new function.
2245 (pass_rebuild_cgraph_edges): Convert from a global struct to a
2246 subclass of gimple_opt_pass along with...
2247 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
2248 (make_pass_rebuild_cgraph_edges): ...new function.
2249 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
2250 subclass of gimple_opt_pass along with...
2251 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
2253 (make_pass_remove_cgraph_callee_edges): ...new function.
2254 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
2255 struct to a subclass of rtl_opt_pass along with...
2256 (pass_data_stack_adjustments): ...new pass_data instance and...
2257 (make_pass_stack_adjustments): ...new function.
2258 * combine.c (pass_combine): Convert from a global struct to a subclass
2259 of rtl_opt_pass along with...
2260 (pass_data_combine): ...new pass_data instance and...
2261 (make_pass_combine): ...new function.
2262 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
2263 global struct to a subclass of rtl_opt_pass along with...
2264 (pass_data_compare_elim_after_reload): ...new pass_data instance
2266 (make_pass_compare_elim_after_reload): ...new function.
2267 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
2268 of rtl_opt_pass along with...
2269 (pass_data_rtl_cprop): ...new pass_data instance and...
2270 (make_pass_rtl_cprop): ...new function.
2271 * cse.c (pass_cse): Convert from a global struct to a subclass of
2272 rtl_opt_pass along with...
2273 (pass_data_cse): ...new pass_data instance and...
2274 (make_pass_cse): ...new function.
2275 (pass_cse2): Convert from a global struct to a subclass of
2276 rtl_opt_pass along with...
2277 (pass_data_cse2): ...new pass_data instance and...
2278 (make_pass_cse2): ...new function.
2279 (pass_cse_after_global_opts): Convert from a global struct to a
2280 subclass of rtl_opt_pass along with...
2281 (pass_data_cse_after_global_opts): ...new pass_data instance and...
2282 (make_pass_cse_after_global_opts): ...new function.
2283 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
2284 of rtl_opt_pass along with...
2285 (pass_data_ud_rtl_dce): ...new pass_data instance and...
2286 (make_pass_ud_rtl_dce): ...new function.
2287 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
2288 rtl_opt_pass along with...
2289 (pass_data_fast_rtl_dce): ...new pass_data instance and...
2290 (make_pass_fast_rtl_dce): ...new function.
2291 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
2292 a subclass of rtl_opt_pass along with...
2293 (pass_data_df_initialize_opt): ...new pass_data instance and...
2294 (make_pass_df_initialize_opt): ...new function.
2295 (pass_df_initialize_no_opt): Convert from a global struct to a
2296 subclass of rtl_opt_pass along with...
2297 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
2298 (make_pass_df_initialize_no_opt): ...new function.
2299 (pass_df_finish): Convert from a global struct to a subclass of
2300 rtl_opt_pass along with...
2301 (pass_data_df_finish): ...new pass_data instance and...
2302 (make_pass_df_finish): ...new function.
2303 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
2304 rtl_opt_pass along with...
2305 (pass_data_rtl_dse1): ...new pass_data instance and...
2306 (make_pass_rtl_dse1): ...new function.
2307 (pass_rtl_dse2): Convert from a global struct to a subclass of
2308 rtl_opt_pass along with...
2309 (pass_data_rtl_dse2): ...new pass_data instance and...
2310 (make_pass_rtl_dse2): ...new function.
2311 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
2312 subclass of rtl_opt_pass along with...
2313 (pass_data_dwarf2_frame): ...new pass_data instance and...
2314 (make_pass_dwarf2_frame): ...new function.
2315 * except.c (pass_set_nothrow_function_flags): Convert from a global
2316 struct to a subclass of rtl_opt_pass along with...
2317 (pass_data_set_nothrow_function_flags): ...new pass_data instance
2319 (make_pass_set_nothrow_function_flags): ...new function.
2320 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
2321 subclass of rtl_opt_pass along with...
2322 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
2324 (make_pass_convert_to_eh_region_ranges): ...new function.
2325 * final.c (pass_compute_alignments): Convert from a global struct to a
2326 subclass of rtl_opt_pass along with...
2327 (pass_data_compute_alignments): ...new pass_data instance and...
2328 (make_pass_compute_alignments): ...new function.
2329 (pass_final): Convert from a global struct to a subclass of
2330 rtl_opt_pass along with...
2331 (pass_data_final): ...new pass_data instance and...
2332 (make_pass_final): ...new function.
2333 (pass_shorten_branches): Convert from a global struct to a subclass of
2334 rtl_opt_pass along with...
2335 (pass_data_shorten_branches): ...new pass_data instance and...
2336 (make_pass_shorten_branches): ...new function.
2337 (pass_clean_state): Convert from a global struct to a subclass of
2338 rtl_opt_pass along with...
2339 (pass_data_clean_state): ...new pass_data instance and...
2340 (make_pass_clean_state): ...new function.
2341 * function.c (pass_instantiate_virtual_regs): Convert from a global
2342 struct to a subclass of rtl_opt_pass along with...
2343 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
2344 (make_pass_instantiate_virtual_regs): ...new function.
2345 (pass_leaf_regs): Convert from a global struct to a subclass of
2346 rtl_opt_pass along with...
2347 (pass_data_leaf_regs): ...new pass_data instance and...
2348 (make_pass_leaf_regs): ...new function.
2349 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
2350 subclass of rtl_opt_pass along with...
2351 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
2353 (make_pass_thread_prologue_and_epilogue): ...new function.
2354 (pass_match_asm_constraints): Convert from a global struct to a
2355 subclass of rtl_opt_pass along with...
2356 (pass_data_match_asm_constraints): ...new pass_data instance and...
2357 (make_pass_match_asm_constraints): ...new function.
2358 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
2359 subclass of rtl_opt_pass along with...
2360 (pass_data_rtl_fwprop): ...new pass_data instance and...
2361 (make_pass_rtl_fwprop): ...new function.
2362 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
2363 rtl_opt_pass along with...
2364 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
2365 (make_pass_rtl_fwprop_addr): ...new function.
2366 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
2367 rtl_opt_pass along with...
2368 (pass_data_rtl_pre): ...new pass_data instance and...
2369 (make_pass_rtl_pre): ...new function.
2370 (pass_rtl_hoist): Convert from a global struct to a subclass of
2371 rtl_opt_pass along with...
2372 (pass_data_rtl_hoist): ...new pass_data instance and...
2373 (make_pass_rtl_hoist): ...new function.
2374 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
2375 subclass of gimple_opt_pass along with...
2376 (pass_data_lower_cf): ...new pass_data instance and...
2377 (make_pass_lower_cf): ...new function.
2378 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
2379 from a global struct to a subclass of gimple_opt_pass along with...
2380 (pass_data_strength_reduction): ...new pass_data instance and...
2381 (make_pass_strength_reduction): ...new function.
2382 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
2383 of rtl_opt_pass along with...
2384 (pass_data_rtl_ifcvt): ...new pass_data instance and...
2385 (make_pass_rtl_ifcvt): ...new function.
2386 (pass_if_after_combine): Convert from a global struct to a subclass of
2387 rtl_opt_pass along with...
2388 (pass_data_if_after_combine): ...new pass_data instance and...
2389 (make_pass_if_after_combine): ...new function.
2390 (pass_if_after_reload): Convert from a global struct to a subclass of
2391 rtl_opt_pass along with...
2392 (pass_data_if_after_reload): ...new pass_data instance and...
2393 (make_pass_if_after_reload): ...new function.
2394 * init-regs.c (pass_initialize_regs): Convert from a global struct to
2395 a subclass of rtl_opt_pass along with...
2396 (pass_data_initialize_regs): ...new pass_data instance and...
2397 (make_pass_initialize_regs): ...new function.
2398 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
2399 of ipa_opt_pass_d along with...
2400 (pass_data_ipa_cp): ...new pass_data instance and...
2401 (make_pass_ipa_cp): ...new function.
2402 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
2403 global struct to a subclass of gimple_opt_pass along with...
2404 (pass_data_inline_parameters): ...new pass_data instance and...
2405 (make_pass_inline_parameters): ...new function.
2406 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
2407 subclass of gimple_opt_pass along with...
2408 (pass_data_early_inline): ...new pass_data instance and...
2409 (make_pass_early_inline): ...new function.
2410 (pass_ipa_inline): Convert from a global struct to a subclass of
2411 ipa_opt_pass_d along with...
2412 (pass_data_ipa_inline): ...new pass_data instance and...
2413 (make_pass_ipa_inline): ...new function.
2414 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
2415 struct to a subclass of gimple_opt_pass along with...
2416 (pass_data_local_pure_const): ...new pass_data instance and...
2417 (make_pass_local_pure_const): ...new function.
2418 (pass_ipa_pure_const): Convert from a global struct to a subclass of
2419 ipa_opt_pass_d along with...
2420 (pass_data_ipa_pure_const): ...new pass_data instance and...
2421 (make_pass_ipa_pure_const): ...new function.
2422 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
2423 to a subclass of ipa_opt_pass_d along with...
2424 (pass_data_ipa_reference): ...new pass_data instance and...
2425 (make_pass_ipa_reference): ...new function.
2426 * ipa-split.c (pass_split_functions): Convert from a global struct to
2427 a subclass of gimple_opt_pass along with...
2428 (pass_data_split_functions): ...new pass_data instance and...
2429 (make_pass_split_functions): ...new function.
2430 (pass_feedback_split_functions): Convert from a global struct to a
2431 subclass of gimple_opt_pass along with...
2432 (pass_data_feedback_split_functions): ...new pass_data instance and...
2433 (make_pass_feedback_split_functions): ...new function.
2434 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
2435 global struct to a subclass of simple_ipa_opt_pass along with...
2436 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
2438 (make_pass_ipa_function_and_variable_visibility): ...new function.
2439 (pass_ipa_free_inline_summary): Convert from a global struct to a
2440 subclass of simple_ipa_opt_pass along with...
2441 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
2442 (make_pass_ipa_free_inline_summary): ...new function.
2443 (pass_ipa_whole_program_visibility): Convert from a global struct to a
2444 subclass of ipa_opt_pass_d along with...
2445 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
2447 (make_pass_ipa_whole_program_visibility): ...new function.
2448 (pass_ipa_profile): Convert from a global struct to a subclass of
2449 ipa_opt_pass_d along with...
2450 (pass_data_ipa_profile): ...new pass_data instance and...
2451 (make_pass_ipa_profile): ...new function.
2452 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
2453 ipa_opt_pass_d along with...
2454 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
2455 (make_pass_ipa_cdtor_merge): ...new function.
2456 * ira.c (pass_ira): Convert from a global struct to a subclass of
2457 rtl_opt_pass along with...
2458 (pass_data_ira): ...new pass_data instance and...
2459 (make_pass_ira): ...new function.
2460 (pass_reload): Convert from a global struct to a subclass of
2461 rtl_opt_pass along with...
2462 (pass_data_reload): ...new pass_data instance and...
2463 (make_pass_reload): ...new function.
2464 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
2465 subclass of rtl_opt_pass along with...
2466 (pass_data_cleanup_barriers): ...new pass_data instance and...
2467 (make_pass_cleanup_barriers): ...new function.
2468 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
2469 of rtl_opt_pass along with...
2470 (pass_data_loop2): ...new pass_data instance and...
2471 (make_pass_loop2): ...new function.
2472 (pass_rtl_loop_init): Convert from a global struct to a subclass of
2473 rtl_opt_pass along with...
2474 (pass_data_rtl_loop_init): ...new pass_data instance and...
2475 (make_pass_rtl_loop_init): ...new function.
2476 (pass_rtl_loop_done): Convert from a global struct to a subclass of
2477 rtl_opt_pass along with...
2478 (pass_data_rtl_loop_done): ...new pass_data instance and...
2479 (make_pass_rtl_loop_done): ...new function.
2480 (pass_rtl_move_loop_invariants): Convert from a global struct to a
2481 subclass of rtl_opt_pass along with...
2482 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
2483 (make_pass_rtl_move_loop_invariants): ...new function.
2484 (pass_rtl_unswitch): Convert from a global struct to a subclass of
2485 rtl_opt_pass along with...
2486 (pass_data_rtl_unswitch): ...new pass_data instance and...
2487 (make_pass_rtl_unswitch): ...new function.
2488 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
2489 subclass of rtl_opt_pass along with...
2490 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
2492 (make_pass_rtl_unroll_and_peel_loops): ...new function.
2493 (pass_rtl_doloop): Convert from a global struct to a subclass of
2494 rtl_opt_pass along with...
2495 (pass_data_rtl_doloop): ...new pass_data instance and...
2496 (make_pass_rtl_doloop): ...new function.
2497 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
2498 a subclass of rtl_opt_pass along with...
2499 (pass_data_lower_subreg): ...new pass_data instance and...
2500 (make_pass_lower_subreg): ...new function.
2501 (pass_lower_subreg2): Convert from a global struct to a subclass of
2502 rtl_opt_pass along with...
2503 (pass_data_lower_subreg2): ...new pass_data instance and...
2504 (make_pass_lower_subreg2): ...new function.
2505 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
2506 struct to a subclass of ipa_opt_pass_d along with...
2507 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
2508 (make_pass_ipa_lto_gimple_out): ...new function.
2509 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
2510 of ipa_opt_pass_d along with...
2511 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
2512 (make_pass_ipa_lto_finish_out): ...new function.
2513 * mode-switching.c (pass_mode_switching): Convert from a global struct
2514 to a subclass of rtl_opt_pass along with...
2515 (pass_data_mode_switching): ...new pass_data instance and...
2516 (make_pass_mode_switching): ...new function.
2517 * modulo-sched.c (pass_sms): Convert from a global struct to a
2518 subclass of rtl_opt_pass along with...
2519 (pass_data_sms): ...new pass_data instance and...
2520 (make_pass_sms): ...new function.
2521 * omp-low.c (pass_expand_omp): Convert from a global struct to a
2522 subclass of gimple_opt_pass along with...
2523 (pass_data_expand_omp): ...new pass_data instance and...
2524 (make_pass_expand_omp): ...new function.
2525 (pass_lower_omp): Convert from a global struct to a subclass of
2526 gimple_opt_pass along with...
2527 (pass_data_lower_omp): ...new pass_data instance and...
2528 (make_pass_lower_omp): ...new function.
2529 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
2530 of gimple_opt_pass along with...
2531 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
2532 (make_pass_diagnose_omp_blocks): ...new function.
2533 * passes.c (pass_early_local_passes): Convert from a global struct to
2534 a subclass of simple_ipa_opt_pass along with...
2535 (pass_data_early_local_passes): ...new pass_data instance and...
2536 (make_pass_early_local_passes): ...new function.
2537 (pass_all_early_optimizations): Convert from a global struct to a
2538 subclass of gimple_opt_pass along with...
2539 (pass_data_all_early_optimizations): ...new pass_data instance and...
2540 (make_pass_all_early_optimizations): ...new function.
2541 (pass_all_optimizations): Convert from a global struct to a subclass
2542 of gimple_opt_pass along with...
2543 (pass_data_all_optimizations): ...new pass_data instance and...
2544 (make_pass_all_optimizations): ...new function.
2545 (pass_all_optimizations_g): Convert from a global struct to a subclass
2546 of gimple_opt_pass along with...
2547 (pass_data_all_optimizations_g): ...new pass_data instance and...
2548 (make_pass_all_optimizations_g): ...new function.
2549 (pass_rest_of_compilation): Convert from a global struct to a subclass
2550 of rtl_opt_pass along with...
2551 (pass_data_rest_of_compilation): ...new pass_data instance and...
2552 (make_pass_rest_of_compilation): ...new function.
2553 (pass_postreload): Convert from a global struct to a subclass of
2554 rtl_opt_pass along with...
2555 (pass_data_postreload): ...new pass_data instance and...
2556 (make_pass_postreload): ...new function.
2557 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
2558 subclass of rtl_opt_pass along with...
2559 (pass_data_gcse2): ...new pass_data instance and...
2560 (make_pass_gcse2): ...new function.
2561 * postreload.c (pass_postreload_cse): Convert from a global struct to
2562 a subclass of rtl_opt_pass along with...
2563 (pass_data_postreload_cse): ...new pass_data instance and...
2564 (make_pass_postreload_cse): ...new function.
2565 * predict.c (pass_profile): Convert from a global struct to a subclass
2566 of gimple_opt_pass along with...
2567 (pass_data_profile): ...new pass_data instance and...
2568 (make_pass_profile): ...new function.
2569 (pass_strip_predict_hints): Convert from a global struct to a subclass
2570 of gimple_opt_pass along with...
2571 (pass_data_strip_predict_hints): ...new pass_data instance and...
2572 (make_pass_strip_predict_hints): ...new function.
2573 * recog.c (pass_peephole2): Convert from a global struct to a subclass
2574 of rtl_opt_pass along with...
2575 (pass_data_peephole2): ...new pass_data instance and...
2576 (make_pass_peephole2): ...new function.
2577 (pass_split_all_insns): Convert from a global struct to a subclass of
2578 rtl_opt_pass along with...
2579 (pass_data_split_all_insns): ...new pass_data instance and...
2580 (make_pass_split_all_insns): ...new function.
2581 (pass_split_after_reload): Convert from a global struct to a subclass
2582 of rtl_opt_pass along with...
2583 (pass_data_split_after_reload): ...new pass_data instance and...
2584 (make_pass_split_after_reload): ...new function.
2585 (pass_split_before_regstack): Convert from a global struct to a
2586 subclass of rtl_opt_pass along with...
2587 (pass_data_split_before_regstack): ...new pass_data instance and...
2588 (make_pass_split_before_regstack): ...new function.
2589 (pass_split_before_sched2): Convert from a global struct to a subclass
2590 of rtl_opt_pass along with...
2591 (pass_data_split_before_sched2): ...new pass_data instance and...
2592 (make_pass_split_before_sched2): ...new function.
2593 (pass_split_for_shorten_branches): Convert from a global struct to a
2594 subclass of rtl_opt_pass along with...
2595 (pass_data_split_for_shorten_branches): ...new pass_data instance
2597 (make_pass_split_for_shorten_branches): ...new function.
2598 * ree.c (pass_ree): Convert from a global struct to a subclass of
2599 rtl_opt_pass along with...
2600 (pass_data_ree): ...new pass_data instance and...
2601 (make_pass_ree): ...new function.
2602 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
2603 subclass of rtl_opt_pass along with...
2604 (pass_data_stack_regs): ...new pass_data instance and...
2605 (make_pass_stack_regs): ...new function.
2606 (pass_stack_regs_run): Convert from a global struct to a subclass of
2607 rtl_opt_pass along with...
2608 (pass_data_stack_regs_run): ...new pass_data instance and...
2609 (make_pass_stack_regs_run): ...new function.
2610 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
2611 subclass of rtl_opt_pass along with...
2612 (pass_data_cprop_hardreg): ...new pass_data instance and...
2613 (make_pass_cprop_hardreg): ...new function.
2614 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
2615 subclass of rtl_opt_pass along with...
2616 (pass_data_reginfo_init): ...new pass_data instance and...
2617 (make_pass_reginfo_init): ...new function.
2618 * regmove.c (pass_regmove): Convert from a global struct to a subclass
2619 of rtl_opt_pass along with...
2620 (pass_data_regmove): ...new pass_data instance and...
2621 (make_pass_regmove): ...new function.
2622 * regrename.c (pass_regrename): Convert from a global struct to a
2623 subclass of rtl_opt_pass along with...
2624 (pass_data_regrename): ...new pass_data instance and...
2625 (make_pass_regrename): ...new function.
2626 * reorg.c (pass_delay_slots): Convert from a global struct to a
2627 subclass of rtl_opt_pass along with...
2628 (pass_data_delay_slots): ...new pass_data instance and...
2629 (make_pass_delay_slots): ...new function.
2630 (pass_machine_reorg): Convert from a global struct to a subclass of
2631 rtl_opt_pass along with...
2632 (pass_data_machine_reorg): ...new pass_data instance and...
2633 (make_pass_machine_reorg): ...new function.
2634 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
2635 of rtl_opt_pass along with...
2636 (pass_data_sched): ...new pass_data instance and...
2637 (make_pass_sched): ...new function.
2638 (pass_sched2): Convert from a global struct to a subclass of
2639 rtl_opt_pass along with...
2640 (pass_data_sched2): ...new pass_data instance and...
2641 (make_pass_sched2): ...new function.
2642 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
2643 to a subclass of rtl_opt_pass along with...
2644 (pass_data_stack_ptr_mod): ...new pass_data instance and...
2645 (make_pass_stack_ptr_mod): ...new function.
2646 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
2647 to a subclass of rtl_opt_pass along with...
2648 (pass_data_rtl_store_motion): ...new pass_data instance and...
2649 (make_pass_rtl_store_motion): ...new function.
2650 * tracer.c (pass_tracer): Convert from a global struct to a subclass
2651 of gimple_opt_pass along with...
2652 (pass_data_tracer): ...new pass_data instance and...
2653 (make_pass_tracer): ...new function.
2654 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
2655 to a subclass of gimple_opt_pass along with...
2656 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
2657 (make_pass_diagnose_tm_blocks): ...new function.
2658 (pass_lower_tm): Convert from a global struct to a subclass of
2659 gimple_opt_pass along with...
2660 (pass_data_lower_tm): ...new pass_data instance and...
2661 (make_pass_lower_tm): ...new function.
2662 (pass_tm_init): Convert from a global struct to a subclass of
2663 gimple_opt_pass along with...
2664 (pass_data_tm_init): ...new pass_data instance and...
2665 (make_pass_tm_init): ...new function.
2666 (pass_tm_mark): Convert from a global struct to a subclass of
2667 gimple_opt_pass along with...
2668 (pass_data_tm_mark): ...new pass_data instance and...
2669 (make_pass_tm_mark): ...new function.
2670 (pass_tm_edges): Convert from a global struct to a subclass of
2671 gimple_opt_pass along with...
2672 (pass_data_tm_edges): ...new pass_data instance and...
2673 (make_pass_tm_edges): ...new function.
2674 (pass_tm_memopt): Convert from a global struct to a subclass of
2675 gimple_opt_pass along with...
2676 (pass_data_tm_memopt): ...new pass_data instance and...
2677 (make_pass_tm_memopt): ...new function.
2678 (pass_ipa_tm): Convert from a global struct to a subclass of
2679 simple_ipa_opt_pass along with...
2680 (pass_data_ipa_tm): ...new pass_data instance and...
2681 (make_pass_ipa_tm): ...new function.
2682 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
2683 subclass of gimple_opt_pass along with...
2684 (pass_data_call_cdce): ...new pass_data instance and...
2685 (make_pass_call_cdce): ...new function.
2686 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
2687 subclass of gimple_opt_pass along with...
2688 (pass_data_build_cfg): ...new pass_data instance and...
2689 (make_pass_build_cfg): ...new function.
2690 (pass_split_crit_edges): Convert from a global struct to a subclass of
2691 gimple_opt_pass along with...
2692 (pass_data_split_crit_edges): ...new pass_data instance and...
2693 (make_pass_split_crit_edges): ...new function.
2694 (pass_warn_function_return): Convert from a global struct to a
2695 subclass of gimple_opt_pass along with...
2696 (pass_data_warn_function_return): ...new pass_data instance and...
2697 (make_pass_warn_function_return): ...new function.
2698 (pass_warn_function_noreturn): Convert from a global struct to a
2699 subclass of gimple_opt_pass along with...
2700 (pass_data_warn_function_noreturn): ...new pass_data instance and...
2701 (make_pass_warn_function_noreturn): ...new function.
2702 (pass_warn_unused_result): Convert from a global struct to a subclass
2703 of gimple_opt_pass along with...
2704 (pass_data_warn_unused_result): ...new pass_data instance and...
2705 (make_pass_warn_unused_result): ...new function.
2706 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
2707 a subclass of gimple_opt_pass along with...
2708 (pass_data_merge_phi): ...new pass_data instance and...
2709 (make_pass_merge_phi): ...new function.
2710 * tree-complex.c (pass_lower_complex): Convert from a global struct to
2711 a subclass of gimple_opt_pass along with...
2712 (pass_data_lower_complex): ...new pass_data instance and...
2713 (make_pass_lower_complex): ...new function.
2714 (pass_lower_complex_O0): Convert from a global struct to a subclass of
2715 gimple_opt_pass along with...
2716 (pass_data_lower_complex_O0): ...new pass_data instance and...
2717 (make_pass_lower_complex_O0): ...new function.
2718 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
2719 subclass of gimple_opt_pass along with...
2720 (pass_data_lower_eh): ...new pass_data instance and...
2721 (make_pass_lower_eh): ...new function.
2722 (pass_refactor_eh): Convert from a global struct to a subclass of
2723 gimple_opt_pass along with...
2724 (pass_data_refactor_eh): ...new pass_data instance and...
2725 (make_pass_refactor_eh): ...new function.
2726 (pass_lower_resx): Convert from a global struct to a subclass of
2727 gimple_opt_pass along with...
2728 (pass_data_lower_resx): ...new pass_data instance and...
2729 (make_pass_lower_resx): ...new function.
2730 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
2731 of gimple_opt_pass along with...
2732 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
2733 (make_pass_lower_eh_dispatch): ...new function.
2734 (pass_cleanup_eh): Convert from a global struct to a subclass of
2735 gimple_opt_pass along with...
2736 (pass_data_cleanup_eh): ...new pass_data instance and...
2737 (make_pass_cleanup_eh): ...new function.
2738 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
2739 to a subclass of simple_ipa_opt_pass along with...
2740 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
2741 (make_pass_ipa_lower_emutls): ...new function.
2742 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
2743 a subclass of gimple_opt_pass along with...
2744 (pass_data_if_conversion): ...new pass_data instance and...
2745 (make_pass_if_conversion): ...new function.
2746 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
2747 subclass of gimple_opt_pass along with...
2748 (pass_data_build_ssa): ...new pass_data instance and...
2749 (make_pass_build_ssa): ...new function.
2750 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
2751 global struct to a subclass of gimple_opt_pass along with...
2752 (pass_data_loop_distribution): ...new pass_data instance and...
2753 (make_pass_loop_distribution): ...new function.
2754 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
2755 subclass of gimple_opt_pass along with...
2756 (pass_data_mudflap_1): ...new pass_data instance and...
2757 (make_pass_mudflap_1): ...new function.
2758 (pass_mudflap_2): Convert from a global struct to a subclass of
2759 gimple_opt_pass along with...
2760 (pass_data_mudflap_2): ...new pass_data instance and...
2761 (make_pass_mudflap_2): ...new function.
2762 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
2763 subclass of gimple_opt_pass along with...
2764 (pass_data_mudflap_1): ...new pass_data instance and...
2765 (make_pass_mudflap_1): ...new function.
2766 (pass_mudflap_2): Convert from a global struct to a subclass of
2767 gimple_opt_pass along with...
2768 (pass_data_mudflap_2): ...new pass_data instance and...
2769 (make_pass_mudflap_2): ...new function.
2770 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
2771 gimple_opt_pass along with...
2772 (pass_data_nrv): ...new pass_data instance and...
2773 (make_pass_nrv): ...new function.
2774 (pass_return_slot): Convert from a global struct to a subclass of
2775 gimple_opt_pass along with...
2776 (pass_data_return_slot): ...new pass_data instance and...
2777 (make_pass_return_slot): ...new function.
2778 * tree-object-size.c (pass_object_sizes): Convert from a global struct
2779 to a subclass of gimple_opt_pass along with...
2780 (pass_data_object_sizes): ...new pass_data instance and...
2781 (make_pass_object_sizes): ...new function.
2782 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
2783 global struct to a subclass of gimple_opt_pass along with...
2784 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
2786 (make_pass_cleanup_cfg_post_optimizing): ...new function.
2787 (pass_fixup_cfg): Convert from a global struct to a subclass of
2788 gimple_opt_pass along with...
2789 (pass_data_fixup_cfg): ...new pass_data instance and...
2790 (make_pass_fixup_cfg): ...new function.
2791 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
2792 (make_pass_mudflap_1): ...new function.
2793 (pass_mudflap_2): Replace declaration with that of...
2794 (make_pass_mudflap_2): ...new function.
2795 (pass_asan): Replace declaration with that of...
2796 (make_pass_asan): ...new function.
2797 (pass_asan_O0): Replace declaration with that of...
2798 (make_pass_asan_O0): ...new function.
2799 (pass_tsan): Replace declaration with that of...
2800 (make_pass_tsan): ...new function.
2801 (pass_tsan_O0): Replace declaration with that of...
2802 (make_pass_tsan_O0): ...new function.
2803 (pass_lower_cf): Replace declaration with that of...
2804 (make_pass_lower_cf): ...new function.
2805 (pass_refactor_eh): Replace declaration with that of...
2806 (make_pass_refactor_eh): ...new function.
2807 (pass_lower_eh): Replace declaration with that of...
2808 (make_pass_lower_eh): ...new function.
2809 (pass_lower_eh_dispatch): Replace declaration with that of...
2810 (make_pass_lower_eh_dispatch): ...new function.
2811 (pass_lower_resx): Replace declaration with that of...
2812 (make_pass_lower_resx): ...new function.
2813 (pass_build_cfg): Replace declaration with that of...
2814 (make_pass_build_cfg): ...new function.
2815 (pass_early_tree_profile): Replace declaration with that of...
2816 (make_pass_early_tree_profile): ...new function.
2817 (pass_cleanup_eh): Replace declaration with that of...
2818 (make_pass_cleanup_eh): ...new function.
2819 (pass_sra): Replace declaration with that of...
2820 (make_pass_sra): ...new function.
2821 (pass_sra_early): Replace declaration with that of...
2822 (make_pass_sra_early): ...new function.
2823 (pass_early_ipa_sra): Replace declaration with that of...
2824 (make_pass_early_ipa_sra): ...new function.
2825 (pass_tail_recursion): Replace declaration with that of...
2826 (make_pass_tail_recursion): ...new function.
2827 (pass_tail_calls): Replace declaration with that of...
2828 (make_pass_tail_calls): ...new function.
2829 (pass_tree_loop): Replace declaration with that of...
2830 (make_pass_tree_loop): ...new function.
2831 (pass_tree_loop_init): Replace declaration with that of...
2832 (make_pass_tree_loop_init): ...new function.
2833 (pass_lim): Replace declaration with that of...
2834 (make_pass_lim): ...new function.
2835 (pass_tree_unswitch): Replace declaration with that of...
2836 (make_pass_tree_unswitch): ...new function.
2837 (pass_predcom): Replace declaration with that of...
2838 (make_pass_predcom): ...new function.
2839 (pass_iv_canon): Replace declaration with that of...
2840 (make_pass_iv_canon): ...new function.
2841 (pass_scev_cprop): Replace declaration with that of...
2842 (make_pass_scev_cprop): ...new function.
2843 (pass_empty_loop): Replace declaration with that of...
2844 (make_pass_empty_loop): ...new function.
2845 (pass_record_bounds): Replace declaration with that of...
2846 (make_pass_record_bounds): ...new function.
2847 (pass_graphite): Replace declaration with that of...
2848 (make_pass_graphite): ...new function.
2849 (pass_graphite_transforms): Replace declaration with that of...
2850 (make_pass_graphite_transforms): ...new function.
2851 (pass_if_conversion): Replace declaration with that of...
2852 (make_pass_if_conversion): ...new function.
2853 (pass_loop_distribution): Replace declaration with that of...
2854 (make_pass_loop_distribution): ...new function.
2855 (pass_vectorize): Replace declaration with that of...
2856 (make_pass_vectorize): ...new function.
2857 (pass_slp_vectorize): Replace declaration with that of...
2858 (make_pass_slp_vectorize): ...new function.
2859 (pass_complete_unroll): Replace declaration with that of...
2860 (make_pass_complete_unroll): ...new function.
2861 (pass_complete_unrolli): Replace declaration with that of...
2862 (make_pass_complete_unrolli): ...new function.
2863 (pass_parallelize_loops): Replace declaration with that of...
2864 (make_pass_parallelize_loops): ...new function.
2865 (pass_loop_prefetch): Replace declaration with that of...
2866 (make_pass_loop_prefetch): ...new function.
2867 (pass_iv_optimize): Replace declaration with that of...
2868 (make_pass_iv_optimize): ...new function.
2869 (pass_tree_loop_done): Replace declaration with that of...
2870 (make_pass_tree_loop_done): ...new function.
2871 (pass_ch): Replace declaration with that of...
2872 (make_pass_ch): ...new function.
2873 (pass_ccp): Replace declaration with that of...
2874 (make_pass_ccp): ...new function.
2875 (pass_phi_only_cprop): Replace declaration with that of...
2876 (make_pass_phi_only_cprop): ...new function.
2877 (pass_build_ssa): Replace declaration with that of...
2878 (make_pass_build_ssa): ...new function.
2879 (pass_build_alias): Replace declaration with that of...
2880 (make_pass_build_alias): ...new function.
2881 (pass_build_ealias): Replace declaration with that of...
2882 (make_pass_build_ealias): ...new function.
2883 (pass_dominator): Replace declaration with that of...
2884 (make_pass_dominator): ...new function.
2885 (pass_dce): Replace declaration with that of...
2886 (make_pass_dce): ...new function.
2887 (pass_dce_loop): Replace declaration with that of...
2888 (make_pass_dce_loop): ...new function.
2889 (pass_cd_dce): Replace declaration with that of...
2890 (make_pass_cd_dce): ...new function.
2891 (pass_call_cdce): Replace declaration with that of...
2892 (make_pass_call_cdce): ...new function.
2893 (pass_merge_phi): Replace declaration with that of...
2894 (make_pass_merge_phi): ...new function.
2895 (pass_split_crit_edges): Replace declaration with that of...
2896 (make_pass_split_crit_edges): ...new function.
2897 (pass_pre): Replace declaration with that of...
2898 (make_pass_pre): ...new function.
2899 (pass_profile): Replace declaration with that of...
2900 (make_pass_profile): ...new function.
2901 (pass_strip_predict_hints): Replace declaration with that of...
2902 (make_pass_strip_predict_hints): ...new function.
2903 (pass_lower_complex_O0): Replace declaration with that of...
2904 (make_pass_lower_complex_O0): ...new function.
2905 (pass_lower_complex): Replace declaration with that of...
2906 (make_pass_lower_complex): ...new function.
2907 (pass_lower_vector): Replace declaration with that of...
2908 (make_pass_lower_vector): ...new function.
2909 (pass_lower_vector_ssa): Replace declaration with that of...
2910 (make_pass_lower_vector_ssa): ...new function.
2911 (pass_lower_omp): Replace declaration with that of...
2912 (make_pass_lower_omp): ...new function.
2913 (pass_diagnose_omp_blocks): Replace declaration with that of...
2914 (make_pass_diagnose_omp_blocks): ...new function.
2915 (pass_expand_omp): Replace declaration with that of...
2916 (make_pass_expand_omp): ...new function.
2917 (pass_expand_omp_ssa): Replace declaration with that of...
2918 (make_pass_expand_omp_ssa): ...new function.
2919 (pass_object_sizes): Replace declaration with that of...
2920 (make_pass_object_sizes): ...new function.
2921 (pass_strlen): Replace declaration with that of...
2922 (make_pass_strlen): ...new function.
2923 (pass_fold_builtins): Replace declaration with that of...
2924 (make_pass_fold_builtins): ...new function.
2925 (pass_stdarg): Replace declaration with that of...
2926 (make_pass_stdarg): ...new function.
2927 (pass_early_warn_uninitialized): Replace declaration with that of...
2928 (make_pass_early_warn_uninitialized): ...new function.
2929 (pass_late_warn_uninitialized): Replace declaration with that of...
2930 (make_pass_late_warn_uninitialized): ...new function.
2931 (pass_cse_reciprocals): Replace declaration with that of...
2932 (make_pass_cse_reciprocals): ...new function.
2933 (pass_cse_sincos): Replace declaration with that of...
2934 (make_pass_cse_sincos): ...new function.
2935 (pass_optimize_bswap): Replace declaration with that of...
2936 (make_pass_optimize_bswap): ...new function.
2937 (pass_optimize_widening_mul): Replace declaration with that of...
2938 (make_pass_optimize_widening_mul): ...new function.
2939 (pass_warn_function_return): Replace declaration with that of...
2940 (make_pass_warn_function_return): ...new function.
2941 (pass_warn_function_noreturn): Replace declaration with that of...
2942 (make_pass_warn_function_noreturn): ...new function.
2943 (pass_cselim): Replace declaration with that of...
2944 (make_pass_cselim): ...new function.
2945 (pass_phiopt): Replace declaration with that of...
2946 (make_pass_phiopt): ...new function.
2947 (pass_forwprop): Replace declaration with that of...
2948 (make_pass_forwprop): ...new function.
2949 (pass_phiprop): Replace declaration with that of...
2950 (make_pass_phiprop): ...new function.
2951 (pass_tree_ifcombine): Replace declaration with that of...
2952 (make_pass_tree_ifcombine): ...new function.
2953 (pass_dse): Replace declaration with that of...
2954 (make_pass_dse): ...new function.
2955 (pass_nrv): Replace declaration with that of...
2956 (make_pass_nrv): ...new function.
2957 (pass_rename_ssa_copies): Replace declaration with that of...
2958 (make_pass_rename_ssa_copies): ...new function.
2959 (pass_sink_code): Replace declaration with that of...
2960 (make_pass_sink_code): ...new function.
2961 (pass_fre): Replace declaration with that of...
2962 (make_pass_fre): ...new function.
2963 (pass_check_data_deps): Replace declaration with that of...
2964 (make_pass_check_data_deps): ...new function.
2965 (pass_copy_prop): Replace declaration with that of...
2966 (make_pass_copy_prop): ...new function.
2967 (pass_vrp): Replace declaration with that of...
2968 (make_pass_vrp): ...new function.
2969 (pass_uncprop): Replace declaration with that of...
2970 (make_pass_uncprop): ...new function.
2971 (pass_return_slot): Replace declaration with that of...
2972 (make_pass_return_slot): ...new function.
2973 (pass_reassoc): Replace declaration with that of...
2974 (make_pass_reassoc): ...new function.
2975 (pass_rebuild_cgraph_edges): Replace declaration with that of...
2976 (make_pass_rebuild_cgraph_edges): ...new function.
2977 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
2978 (make_pass_remove_cgraph_callee_edges): ...new function.
2979 (pass_build_cgraph_edges): Replace declaration with that of...
2980 (make_pass_build_cgraph_edges): ...new function.
2981 (pass_local_pure_const): Replace declaration with that of...
2982 (make_pass_local_pure_const): ...new function.
2983 (pass_tracer): Replace declaration with that of...
2984 (make_pass_tracer): ...new function.
2985 (pass_warn_unused_result): Replace declaration with that of...
2986 (make_pass_warn_unused_result): ...new function.
2987 (pass_diagnose_tm_blocks): Replace declaration with that of...
2988 (make_pass_diagnose_tm_blocks): ...new function.
2989 (pass_lower_tm): Replace declaration with that of...
2990 (make_pass_lower_tm): ...new function.
2991 (pass_tm_init): Replace declaration with that of...
2992 (make_pass_tm_init): ...new function.
2993 (pass_tm_mark): Replace declaration with that of...
2994 (make_pass_tm_mark): ...new function.
2995 (pass_tm_memopt): Replace declaration with that of...
2996 (make_pass_tm_memopt): ...new function.
2997 (pass_tm_edges): Replace declaration with that of...
2998 (make_pass_tm_edges): ...new function.
2999 (pass_split_functions): Replace declaration with that of...
3000 (make_pass_split_functions): ...new function.
3001 (pass_feedback_split_functions): Replace declaration with that of...
3002 (make_pass_feedback_split_functions): ...new function.
3003 (pass_strength_reduction): Replace declaration with that of...
3004 (make_pass_strength_reduction): ...new function.
3005 (pass_ipa_lower_emutls): Replace declaration with that of...
3006 (make_pass_ipa_lower_emutls): ...new function.
3007 (pass_ipa_function_and_variable_visibility): Replace declaration with
3009 (make_pass_ipa_function_and_variable_visibility): ...new function.
3010 (pass_ipa_tree_profile): Replace declaration with that of...
3011 (make_pass_ipa_tree_profile): ...new function.
3012 (pass_early_local_passes): Replace declaration with that of...
3013 (make_pass_early_local_passes): ...new function.
3014 (pass_ipa_whole_program_visibility): Replace declaration with that
3016 (make_pass_ipa_whole_program_visibility): ...new function.
3017 (pass_ipa_lto_gimple_out): Replace declaration with that of...
3018 (make_pass_ipa_lto_gimple_out): ...new function.
3019 (pass_ipa_increase_alignment): Replace declaration with that of...
3020 (make_pass_ipa_increase_alignment): ...new function.
3021 (pass_ipa_inline): Replace declaration with that of...
3022 (make_pass_ipa_inline): ...new function.
3023 (pass_ipa_free_lang_data): Replace declaration with that of...
3024 (make_pass_ipa_free_lang_data): ...new function.
3025 (pass_ipa_free_inline_summary): Replace declaration with that of...
3026 (make_pass_ipa_free_inline_summary): ...new function.
3027 (pass_ipa_cp): Replace declaration with that of...
3028 (make_pass_ipa_cp): ...new function.
3029 (pass_ipa_reference): Replace declaration with that of...
3030 (make_pass_ipa_reference): ...new function.
3031 (pass_ipa_pure_const): Replace declaration with that of...
3032 (make_pass_ipa_pure_const): ...new function.
3033 (pass_ipa_pta): Replace declaration with that of...
3034 (make_pass_ipa_pta): ...new function.
3035 (pass_ipa_lto_finish_out): Replace declaration with that of...
3036 (make_pass_ipa_lto_finish_out): ...new function.
3037 (pass_ipa_tm): Replace declaration with that of...
3038 (make_pass_ipa_tm): ...new function.
3039 (pass_ipa_profile): Replace declaration with that of...
3040 (make_pass_ipa_profile): ...new function.
3041 (pass_ipa_cdtor_merge): Replace declaration with that of...
3042 (make_pass_ipa_cdtor_merge): ...new function.
3043 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
3045 (make_pass_cleanup_cfg_post_optimizing): ...new function.
3046 (pass_init_datastructures): Replace declaration with that of...
3047 (make_pass_init_datastructures): ...new function.
3048 (pass_fixup_cfg): Replace declaration with that of...
3049 (make_pass_fixup_cfg): ...new function.
3050 (pass_expand): Replace declaration with that of...
3051 (make_pass_expand): ...new function.
3052 (pass_instantiate_virtual_regs): Replace declaration with that of...
3053 (make_pass_instantiate_virtual_regs): ...new function.
3054 (pass_rtl_fwprop): Replace declaration with that of...
3055 (make_pass_rtl_fwprop): ...new function.
3056 (pass_rtl_fwprop_addr): Replace declaration with that of...
3057 (make_pass_rtl_fwprop_addr): ...new function.
3058 (pass_jump): Replace declaration with that of...
3059 (make_pass_jump): ...new function.
3060 (pass_jump2): Replace declaration with that of...
3061 (make_pass_jump2): ...new function.
3062 (pass_lower_subreg): Replace declaration with that of...
3063 (make_pass_lower_subreg): ...new function.
3064 (pass_cse): Replace declaration with that of...
3065 (make_pass_cse): ...new function.
3066 (pass_fast_rtl_dce): Replace declaration with that of...
3067 (make_pass_fast_rtl_dce): ...new function.
3068 (pass_ud_rtl_dce): Replace declaration with that of...
3069 (make_pass_ud_rtl_dce): ...new function.
3070 (pass_rtl_dce): Replace declaration with that of...
3071 (make_pass_rtl_dce): ...new function.
3072 (pass_rtl_dse1): Replace declaration with that of...
3073 (make_pass_rtl_dse1): ...new function.
3074 (pass_rtl_dse2): Replace declaration with that of...
3075 (make_pass_rtl_dse2): ...new function.
3076 (pass_rtl_dse3): Replace declaration with that of...
3077 (make_pass_rtl_dse3): ...new function.
3078 (pass_rtl_cprop): Replace declaration with that of...
3079 (make_pass_rtl_cprop): ...new function.
3080 (pass_rtl_pre): Replace declaration with that of...
3081 (make_pass_rtl_pre): ...new function.
3082 (pass_rtl_hoist): Replace declaration with that of...
3083 (make_pass_rtl_hoist): ...new function.
3084 (pass_rtl_store_motion): Replace declaration with that of...
3085 (make_pass_rtl_store_motion): ...new function.
3086 (pass_cse_after_global_opts): Replace declaration with that of...
3087 (make_pass_cse_after_global_opts): ...new function.
3088 (pass_rtl_ifcvt): Replace declaration with that of...
3089 (make_pass_rtl_ifcvt): ...new function.
3090 (pass_into_cfg_layout_mode): Replace declaration with that of...
3091 (make_pass_into_cfg_layout_mode): ...new function.
3092 (pass_outof_cfg_layout_mode): Replace declaration with that of...
3093 (make_pass_outof_cfg_layout_mode): ...new function.
3094 (pass_loop2): Replace declaration with that of...
3095 (make_pass_loop2): ...new function.
3096 (pass_rtl_loop_init): Replace declaration with that of...
3097 (make_pass_rtl_loop_init): ...new function.
3098 (pass_rtl_move_loop_invariants): Replace declaration with that of...
3099 (make_pass_rtl_move_loop_invariants): ...new function.
3100 (pass_rtl_unswitch): Replace declaration with that of...
3101 (make_pass_rtl_unswitch): ...new function.
3102 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
3103 (make_pass_rtl_unroll_and_peel_loops): ...new function.
3104 (pass_rtl_doloop): Replace declaration with that of...
3105 (make_pass_rtl_doloop): ...new function.
3106 (pass_rtl_loop_done): Replace declaration with that of...
3107 (make_pass_rtl_loop_done): ...new function.
3108 (pass_web): Replace declaration with that of...
3109 (make_pass_web): ...new function.
3110 (pass_cse2): Replace declaration with that of...
3111 (make_pass_cse2): ...new function.
3112 (pass_df_initialize_opt): Replace declaration with that of...
3113 (make_pass_df_initialize_opt): ...new function.
3114 (pass_df_initialize_no_opt): Replace declaration with that of...
3115 (make_pass_df_initialize_no_opt): ...new function.
3116 (pass_reginfo_init): Replace declaration with that of...
3117 (make_pass_reginfo_init): ...new function.
3118 (pass_inc_dec): Replace declaration with that of...
3119 (make_pass_inc_dec): ...new function.
3120 (pass_stack_ptr_mod): Replace declaration with that of...
3121 (make_pass_stack_ptr_mod): ...new function.
3122 (pass_initialize_regs): Replace declaration with that of...
3123 (make_pass_initialize_regs): ...new function.
3124 (pass_combine): Replace declaration with that of...
3125 (make_pass_combine): ...new function.
3126 (pass_if_after_combine): Replace declaration with that of...
3127 (make_pass_if_after_combine): ...new function.
3128 (pass_ree): Replace declaration with that of...
3129 (make_pass_ree): ...new function.
3130 (pass_partition_blocks): Replace declaration with that of...
3131 (make_pass_partition_blocks): ...new function.
3132 (pass_match_asm_constraints): Replace declaration with that of...
3133 (make_pass_match_asm_constraints): ...new function.
3134 (pass_regmove): Replace declaration with that of...
3135 (make_pass_regmove): ...new function.
3136 (pass_split_all_insns): Replace declaration with that of...
3137 (make_pass_split_all_insns): ...new function.
3138 (pass_fast_rtl_byte_dce): Replace declaration with that of...
3139 (make_pass_fast_rtl_byte_dce): ...new function.
3140 (pass_lower_subreg2): Replace declaration with that of...
3141 (make_pass_lower_subreg2): ...new function.
3142 (pass_mode_switching): Replace declaration with that of...
3143 (make_pass_mode_switching): ...new function.
3144 (pass_sms): Replace declaration with that of...
3145 (make_pass_sms): ...new function.
3146 (pass_sched): Replace declaration with that of...
3147 (make_pass_sched): ...new function.
3148 (pass_ira): Replace declaration with that of...
3149 (make_pass_ira): ...new function.
3150 (pass_reload): Replace declaration with that of...
3151 (make_pass_reload): ...new function.
3152 (pass_clean_state): Replace declaration with that of...
3153 (make_pass_clean_state): ...new function.
3154 (pass_branch_prob): Replace declaration with that of...
3155 (make_pass_branch_prob): ...new function.
3156 (pass_value_profile_transformations): Replace declaration with that
3158 (make_pass_value_profile_transformations): ...new function.
3159 (pass_postreload_cse): Replace declaration with that of...
3160 (make_pass_postreload_cse): ...new function.
3161 (pass_gcse2): Replace declaration with that of...
3162 (make_pass_gcse2): ...new function.
3163 (pass_split_after_reload): Replace declaration with that of...
3164 (make_pass_split_after_reload): ...new function.
3165 (pass_branch_target_load_optimize1): Replace declaration with that
3167 (make_pass_branch_target_load_optimize1): ...new function.
3168 (pass_thread_prologue_and_epilogue): Replace declaration with that
3170 (make_pass_thread_prologue_and_epilogue): ...new function.
3171 (pass_stack_adjustments): Replace declaration with that of...
3172 (make_pass_stack_adjustments): ...new function.
3173 (pass_peephole2): Replace declaration with that of...
3174 (make_pass_peephole2): ...new function.
3175 (pass_if_after_reload): Replace declaration with that of...
3176 (make_pass_if_after_reload): ...new function.
3177 (pass_regrename): Replace declaration with that of...
3178 (make_pass_regrename): ...new function.
3179 (pass_cprop_hardreg): Replace declaration with that of...
3180 (make_pass_cprop_hardreg): ...new function.
3181 (pass_reorder_blocks): Replace declaration with that of...
3182 (make_pass_reorder_blocks): ...new function.
3183 (pass_branch_target_load_optimize2): Replace declaration with that
3185 (make_pass_branch_target_load_optimize2): ...new function.
3186 (pass_leaf_regs): Replace declaration with that of...
3187 (make_pass_leaf_regs): ...new function.
3188 (pass_split_before_sched2): Replace declaration with that of...
3189 (make_pass_split_before_sched2): ...new function.
3190 (pass_compare_elim_after_reload): Replace declaration with that of...
3191 (make_pass_compare_elim_after_reload): ...new function.
3192 (pass_sched2): Replace declaration with that of...
3193 (make_pass_sched2): ...new function.
3194 (pass_stack_regs): Replace declaration with that of...
3195 (make_pass_stack_regs): ...new function.
3196 (pass_stack_regs_run): Replace declaration with that of...
3197 (make_pass_stack_regs_run): ...new function.
3198 (pass_df_finish): Replace declaration with that of...
3199 (make_pass_df_finish): ...new function.
3200 (pass_compute_alignments): Replace declaration with that of...
3201 (make_pass_compute_alignments): ...new function.
3202 (pass_duplicate_computed_gotos): Replace declaration with that of...
3203 (make_pass_duplicate_computed_gotos): ...new function.
3204 (pass_variable_tracking): Replace declaration with that of...
3205 (make_pass_variable_tracking): ...new function.
3206 (pass_free_cfg): Replace declaration with that of...
3207 (make_pass_free_cfg): ...new function.
3208 (pass_machine_reorg): Replace declaration with that of...
3209 (make_pass_machine_reorg): ...new function.
3210 (pass_cleanup_barriers): Replace declaration with that of...
3211 (make_pass_cleanup_barriers): ...new function.
3212 (pass_delay_slots): Replace declaration with that of...
3213 (make_pass_delay_slots): ...new function.
3214 (pass_split_for_shorten_branches): Replace declaration with that of...
3215 (make_pass_split_for_shorten_branches): ...new function.
3216 (pass_split_before_regstack): Replace declaration with that of...
3217 (make_pass_split_before_regstack): ...new function.
3218 (pass_convert_to_eh_region_ranges): Replace declaration with that
3220 (make_pass_convert_to_eh_region_ranges): ...new function.
3221 (pass_shorten_branches): Replace declaration with that of...
3222 (make_pass_shorten_branches): ...new function.
3223 (pass_set_nothrow_function_flags): Replace declaration with that of...
3224 (make_pass_set_nothrow_function_flags): ...new function.
3225 (pass_dwarf2_frame): Replace declaration with that of...
3226 (make_pass_dwarf2_frame): ...new function.
3227 (pass_final): Replace declaration with that of...
3228 (make_pass_final): ...new function.
3229 (pass_rtl_seqabstr): Replace declaration with that of...
3230 (make_pass_rtl_seqabstr): ...new function.
3231 (pass_release_ssa_names): Replace declaration with that of...
3232 (make_pass_release_ssa_names): ...new function.
3233 (pass_early_inline): Replace declaration with that of...
3234 (make_pass_early_inline): ...new function.
3235 (pass_inline_parameters): Replace declaration with that of...
3236 (make_pass_inline_parameters): ...new function.
3237 (pass_update_address_taken): Replace declaration with that of...
3238 (make_pass_update_address_taken): ...new function.
3239 (pass_convert_switch): Replace declaration with that of...
3240 (make_pass_convert_switch): ...new function.
3241 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
3242 to a subclass of simple_ipa_opt_pass along with...
3243 (pass_data_ipa_tree_profile): ...new pass_data instance and...
3244 (make_pass_ipa_tree_profile): ...new function.
3245 * tree-sra.c (pass_sra_early): Convert from a global struct to a
3246 subclass of gimple_opt_pass along with...
3247 (pass_data_sra_early): ...new pass_data instance and...
3248 (make_pass_sra_early): ...new function.
3249 (pass_sra): Convert from a global struct to a subclass of
3250 gimple_opt_pass along with...
3251 (pass_data_sra): ...new pass_data instance and...
3252 (make_pass_sra): ...new function.
3253 (pass_early_ipa_sra): Convert from a global struct to a subclass of
3254 gimple_opt_pass along with...
3255 (pass_data_early_ipa_sra): ...new pass_data instance and...
3256 (make_pass_early_ipa_sra): ...new function.
3257 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
3258 subclass of gimple_opt_pass along with...
3259 (pass_data_ccp): ...new pass_data instance and...
3260 (make_pass_ccp): ...new function.
3261 (pass_fold_builtins): Convert from a global struct to a subclass of
3262 gimple_opt_pass along with...
3263 (pass_data_fold_builtins): ...new pass_data instance and...
3264 (make_pass_fold_builtins): ...new function.
3265 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
3266 subclass of gimple_opt_pass along with...
3267 (pass_data_copy_prop): ...new pass_data instance and...
3268 (make_pass_copy_prop): ...new function.
3269 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
3270 global struct to a subclass of gimple_opt_pass along with...
3271 (pass_data_rename_ssa_copies): ...new pass_data instance and...
3272 (make_pass_rename_ssa_copies): ...new function.
3273 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
3274 subclass of gimple_opt_pass along with...
3275 (pass_data_dce): ...new pass_data instance and...
3276 (make_pass_dce): ...new function.
3277 (pass_dce_loop): Convert from a global struct to a subclass of
3278 gimple_opt_pass along with...
3279 (pass_data_dce_loop): ...new pass_data instance and...
3280 (make_pass_dce_loop): ...new function.
3281 (pass_cd_dce): Convert from a global struct to a subclass of
3282 gimple_opt_pass along with...
3283 (pass_data_cd_dce): ...new pass_data instance and...
3284 (make_pass_cd_dce): ...new function.
3285 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
3286 subclass of gimple_opt_pass along with...
3287 (pass_data_dominator): ...new pass_data instance and...
3288 (make_pass_dominator): ...new function.
3289 (pass_phi_only_cprop): Convert from a global struct to a subclass of
3290 gimple_opt_pass along with...
3291 (pass_data_phi_only_cprop): ...new pass_data instance and...
3292 (make_pass_phi_only_cprop): ...new function.
3293 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
3294 subclass of gimple_opt_pass along with...
3295 (pass_data_dse): ...new pass_data instance and...
3296 (make_pass_dse): ...new function.
3297 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
3298 a subclass of gimple_opt_pass along with...
3299 (pass_data_forwprop): ...new pass_data instance and...
3300 (make_pass_forwprop): ...new function.
3301 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
3302 struct to a subclass of gimple_opt_pass along with...
3303 (pass_data_tree_ifcombine): ...new pass_data instance and...
3304 (make_pass_tree_ifcombine): ...new function.
3305 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
3306 subclass of gimple_opt_pass along with...
3307 (pass_data_ch): ...new pass_data instance and...
3308 (make_pass_ch): ...new function.
3309 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
3310 subclass of gimple_opt_pass along with...
3311 (pass_data_tree_loop): ...new pass_data instance and...
3312 (make_pass_tree_loop): ...new function.
3313 (pass_tree_loop_init): Convert from a global struct to a subclass of
3314 gimple_opt_pass along with...
3315 (pass_data_tree_loop_init): ...new pass_data instance and...
3316 (make_pass_tree_loop_init): ...new function.
3317 (pass_lim): Convert from a global struct to a subclass of
3318 gimple_opt_pass along with...
3319 (pass_data_lim): ...new pass_data instance and...
3320 (make_pass_lim): ...new function.
3321 (pass_tree_unswitch): Convert from a global struct to a subclass of
3322 gimple_opt_pass along with...
3323 (pass_data_tree_unswitch): ...new pass_data instance and...
3324 (make_pass_tree_unswitch): ...new function.
3325 (pass_predcom): Convert from a global struct to a subclass of
3326 gimple_opt_pass along with...
3327 (pass_data_predcom): ...new pass_data instance and...
3328 (make_pass_predcom): ...new function.
3329 (pass_vectorize): Convert from a global struct to a subclass of
3330 gimple_opt_pass along with...
3331 (pass_data_vectorize): ...new pass_data instance and...
3332 (make_pass_vectorize): ...new function.
3333 (pass_graphite): Convert from a global struct to a subclass of
3334 gimple_opt_pass along with...
3335 (pass_data_graphite): ...new pass_data instance and...
3336 (make_pass_graphite): ...new function.
3337 (pass_graphite_transforms): Convert from a global struct to a subclass
3338 of gimple_opt_pass along with...
3339 (pass_data_graphite_transforms): ...new pass_data instance and...
3340 (make_pass_graphite_transforms): ...new function.
3341 (pass_check_data_deps): Convert from a global struct to a subclass of
3342 gimple_opt_pass along with...
3343 (pass_data_check_data_deps): ...new pass_data instance and...
3344 (make_pass_check_data_deps): ...new function.
3345 (pass_iv_canon): Convert from a global struct to a subclass of
3346 gimple_opt_pass along with...
3347 (pass_data_iv_canon): ...new pass_data instance and...
3348 (make_pass_iv_canon): ...new function.
3349 (pass_scev_cprop): Convert from a global struct to a subclass of
3350 gimple_opt_pass along with...
3351 (pass_data_scev_cprop): ...new pass_data instance and...
3352 (make_pass_scev_cprop): ...new function.
3353 (pass_record_bounds): Convert from a global struct to a subclass of
3354 gimple_opt_pass along with...
3355 (pass_data_record_bounds): ...new pass_data instance and...
3356 (make_pass_record_bounds): ...new function.
3357 (pass_complete_unroll): Convert from a global struct to a subclass of
3358 gimple_opt_pass along with...
3359 (pass_data_complete_unroll): ...new pass_data instance and...
3360 (make_pass_complete_unroll): ...new function.
3361 (pass_complete_unrolli): Convert from a global struct to a subclass of
3362 gimple_opt_pass along with...
3363 (pass_data_complete_unrolli): ...new pass_data instance and...
3364 (make_pass_complete_unrolli): ...new function.
3365 (pass_parallelize_loops): Convert from a global struct to a subclass
3366 of gimple_opt_pass along with...
3367 (pass_data_parallelize_loops): ...new pass_data instance and...
3368 (make_pass_parallelize_loops): ...new function.
3369 (pass_loop_prefetch): Convert from a global struct to a subclass of
3370 gimple_opt_pass along with...
3371 (pass_data_loop_prefetch): ...new pass_data instance and...
3372 (make_pass_loop_prefetch): ...new function.
3373 (pass_iv_optimize): Convert from a global struct to a subclass of
3374 gimple_opt_pass along with...
3375 (pass_data_iv_optimize): ...new pass_data instance and...
3376 (make_pass_iv_optimize): ...new function.
3377 (pass_tree_loop_done): Convert from a global struct to a subclass of
3378 gimple_opt_pass along with...
3379 (pass_data_tree_loop_done): ...new pass_data instance and...
3380 (make_pass_tree_loop_done): ...new function.
3381 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
3382 struct to a subclass of gimple_opt_pass along with...
3383 (pass_data_cse_reciprocals): ...new pass_data instance and...
3384 (make_pass_cse_reciprocals): ...new function.
3385 (pass_cse_sincos): Convert from a global struct to a subclass of
3386 gimple_opt_pass along with...
3387 (pass_data_cse_sincos): ...new pass_data instance and...
3388 (make_pass_cse_sincos): ...new function.
3389 (pass_optimize_bswap): Convert from a global struct to a subclass of
3390 gimple_opt_pass along with...
3391 (pass_data_optimize_bswap): ...new pass_data instance and...
3392 (make_pass_optimize_bswap): ...new function.
3393 (pass_optimize_widening_mul): Convert from a global struct to a
3394 subclass of gimple_opt_pass along with...
3395 (pass_data_optimize_widening_mul): ...new pass_data instance and...
3396 (make_pass_optimize_widening_mul): ...new function.
3397 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
3398 subclass of gimple_opt_pass along with...
3399 (pass_data_phiopt): ...new pass_data instance and...
3400 (make_pass_phiopt): ...new function.
3401 (pass_cselim): Convert from a global struct to a subclass of
3402 gimple_opt_pass along with...
3403 (pass_data_cselim): ...new pass_data instance and...
3404 (make_pass_cselim): ...new function.
3405 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
3406 subclass of gimple_opt_pass along with...
3407 (pass_data_phiprop): ...new pass_data instance and...
3408 (make_pass_phiprop): ...new function.
3409 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
3410 subclass of gimple_opt_pass along with...
3411 (pass_data_pre): ...new pass_data instance and...
3412 (make_pass_pre): ...new function.
3413 (pass_fre): Convert from a global struct to a subclass of
3414 gimple_opt_pass along with...
3415 (pass_data_fre): ...new pass_data instance and...
3416 (make_pass_fre): ...new function.
3417 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
3418 subclass of gimple_opt_pass along with...
3419 (pass_data_reassoc): ...new pass_data instance and...
3420 (make_pass_reassoc): ...new function.
3421 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
3422 subclass of gimple_opt_pass along with...
3423 (pass_data_sink_code): ...new pass_data instance and...
3424 (make_pass_sink_code): ...new function.
3425 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
3426 subclass of gimple_opt_pass along with...
3427 (pass_data_strlen): ...new pass_data instance and...
3428 (make_pass_strlen): ...new function.
3429 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
3430 struct to a subclass of gimple_opt_pass along with...
3431 (pass_data_build_alias): ...new pass_data instance and...
3432 (make_pass_build_alias): ...new function.
3433 (pass_build_ealias): Convert from a global struct to a subclass of
3434 gimple_opt_pass along with...
3435 (pass_data_build_ealias): ...new pass_data instance and...
3436 (make_pass_build_ealias): ...new function.
3437 (pass_ipa_pta): Convert from a global struct to a subclass of
3438 simple_ipa_opt_pass along with...
3439 (pass_data_ipa_pta): ...new pass_data instance and...
3440 (make_pass_ipa_pta): ...new function.
3441 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
3442 subclass of gimple_opt_pass along with...
3443 (pass_data_uncprop): ...new pass_data instance and...
3444 (make_pass_uncprop): ...new function.
3445 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
3446 global struct to a subclass of gimple_opt_pass along with...
3447 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
3448 (make_pass_late_warn_uninitialized): ...new function.
3449 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
3450 to a subclass of gimple_opt_pass along with...
3451 (pass_data_init_datastructures): ...new pass_data instance and...
3452 (make_pass_init_datastructures): ...new function.
3453 (pass_early_warn_uninitialized): Convert from a global struct to a
3454 subclass of gimple_opt_pass along with...
3455 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
3456 (make_pass_early_warn_uninitialized): ...new function.
3457 (pass_update_address_taken): Convert from a global struct to a
3458 subclass of gimple_opt_pass along with...
3459 (pass_data_update_address_taken): ...new pass_data instance and...
3460 (make_pass_update_address_taken): ...new function.
3461 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
3462 struct to a subclass of gimple_opt_pass along with...
3463 (pass_data_release_ssa_names): ...new pass_data instance and...
3464 (make_pass_release_ssa_names): ...new function.
3465 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
3466 subclass of gimple_opt_pass along with...
3467 (pass_data_stdarg): ...new pass_data instance and...
3468 (make_pass_stdarg): ...new function.
3469 * tree-switch-conversion.c (pass_convert_switch): Convert from a
3470 global struct to a subclass of gimple_opt_pass along with...
3471 (pass_data_convert_switch): ...new pass_data instance and...
3472 (make_pass_convert_switch): ...new function.
3473 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
3474 to a subclass of gimple_opt_pass along with...
3475 (pass_data_tail_recursion): ...new pass_data instance and...
3476 (make_pass_tail_recursion): ...new function.
3477 (pass_tail_calls): Convert from a global struct to a subclass of
3478 gimple_opt_pass along with...
3479 (pass_data_tail_calls): ...new pass_data instance and...
3480 (make_pass_tail_calls): ...new function.
3481 * tree-vect-generic.c (pass_lower_vector): Convert from a global
3482 struct to a subclass of gimple_opt_pass along with...
3483 (pass_data_lower_vector): ...new pass_data instance and...
3484 (make_pass_lower_vector): ...new function.
3485 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
3486 gimple_opt_pass along with...
3487 (pass_data_lower_vector_ssa): ...new pass_data instance and...
3488 (make_pass_lower_vector_ssa): ...new function.
3489 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
3490 to a subclass of gimple_opt_pass along with...
3491 (pass_data_slp_vectorize): ...new pass_data instance and...
3492 (make_pass_slp_vectorize): ...new function.
3493 (pass_ipa_increase_alignment): Convert from a global struct to a
3494 subclass of simple_ipa_opt_pass along with...
3495 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
3496 (make_pass_ipa_increase_alignment): ...new function.
3497 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
3498 gimple_opt_pass along with...
3499 (pass_data_vrp): ...new pass_data instance and...
3500 (make_pass_vrp): ...new function.
3501 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
3502 subclass of simple_ipa_opt_pass along with...
3503 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
3504 (make_pass_ipa_free_lang_data): ...new function.
3505 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
3506 gimple_opt_pass along with...
3507 (pass_data_tsan): ...new pass_data instance and...
3508 (make_pass_tsan): ...new function.
3509 (pass_tsan_O0): Convert from a global struct to a subclass of
3510 gimple_opt_pass along with...
3511 (pass_data_tsan_O0): ...new pass_data instance and...
3512 (make_pass_tsan_O0): ...new function.
3513 * var-tracking.c (pass_variable_tracking): Convert from a global
3514 struct to a subclass of rtl_opt_pass along with...
3515 (pass_data_variable_tracking): ...new pass_data instance and...
3516 (make_pass_variable_tracking): ...new function.
3517 * web.c (pass_web): Convert from a global struct to a subclass of
3518 rtl_opt_pass along with...
3519 (pass_data_web): ...new pass_data instance and...
3520 (make_pass_web): ...new function.
3521 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
3522 declaration with that of...
3523 (make_pass_mode_switch_use): ...new function.
3524 (pass_resolve_sw_modes): Replace declaration with that of...
3525 (make_pass_resolve_sw_modes): ...new function.
3526 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
3527 from a global struct to a subclass of rtl_opt_pass along with...
3528 (pass_data_mode_switch_use): ...new pass_data instance and...
3529 (make_pass_mode_switch_use): ...new function.
3530 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
3531 from a global struct to a subclass of rtl_opt_pass along with...
3532 (pass_data_resolve_sw_modes): ...new pass_data instance and...
3533 (make_pass_resolve_sw_modes): ...new function.
3534 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
3535 struct to a subclass of rtl_opt_pass along with...
3536 (pass_data_insert_vzeroupper): ...new pass_data instance and...
3537 (make_pass_insert_vzeroupper): ...new function.
3538 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
3539 global struct to a subclass of rtl_opt_pass along with...
3540 (pass_data_work_around_errata): ...new pass_data instance and...
3541 (make_pass_work_around_errata): ...new function.
3542 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
3543 struct to a subclass of rtl_opt_pass along with...
3544 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
3545 (make_pass_mips_machine_reorg2): ...new function.
3547 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3549 * passes.c (pass_manager::operator new): New.
3551 2013-08-05 David Malcolm <dmalcolm@redhat.com>
3553 Handwritten part of conversion of passes to C++ classes.
3555 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
3556 (toplev.o): Add dep on PASS_MANAGER_H.
3557 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
3558 of early local pases to reflect this moving from a global to a
3559 member of gcc::pass_manager.
3560 (cgraph_add_new_function): Likewise.
3561 * lto-cgraph.c (lto_output_node): Update for conversion of
3562 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
3563 * passes.c (opt_pass::clone): New.
3564 (opt_pass::gate): New.
3565 (opt_pass::execute): New.
3566 (opt_pass::opt_pass): New.
3567 (pass_manager::execute_early_local_passes): New.
3568 (pass_manager::execute_pass_mode_switching): new.
3569 (finish_optimization_passes): Convert to...
3570 (pass_manager::finish_optimization_passes): ...this.
3571 (finish_optimization_passes): Update for conversion of passes to
3573 (register_dump_files_1): Use has_gate since we cannot portably
3574 check a vtable entry against NULL.
3575 (dump_one_pass): Likewise.
3576 (ipa_write_summaries_2): Likewise.
3577 (ipa_write_optimization_summaries_1): Likewise.
3578 (ipa_read_summaries_1): Likewise.
3579 (ipa_read_optimization_summaries_1): Likewise.
3580 (execute_ipa_stmt_fixups): Likewise.
3581 (pass_manager::pass_manager): Rewrite pass-creation, invoking
3582 pass-creation functions rather than wiring up globals, and
3583 storing the results in fields of pass_manager generated using
3585 (pass_manager::dump_profile_report): Update for conversion of
3586 passes to C++ classes.
3587 (pass_manager::execute_ipa_summary_passes): Likewise.
3588 (execute_one_ipa_transform_pass): Likewise.
3589 (execute_one_pass): Use has_gate and has_execute since we cannot
3590 portably check a vtable entry against NULL.
3591 * pass_manager.h (pass_manager::finish_optimization_passes): New.
3592 (pass_manager): Use pass-instances.def to add fields for the
3593 various pass instances.
3594 * toplev.c (finalize): Update for move of
3595 finish_optimization_passes to a method of gcc::pass_manager.
3596 * toplev.h (finish_optimization_passes): Move to method of class
3598 * tree-pass.h (struct pass_data): New.
3599 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
3600 (opt_pass::gate): Convert to virtual function.
3601 (opt_pass::~opt_pass): New.
3602 (opt_pass::clone): New.
3603 (opt_pass::execute): Convert to virtual function.
3604 (opt_pass::opt_pass): New.
3605 (opt_pass::ctxt_): new.
3606 (gimple_opt_pass): Convert to subclass of opt_pass.
3607 (gimple_opt_pass::gimple_opt_pass): New.
3608 (rtl_opt_pass): Convert to subclass of opt_pass.
3609 (rtl_opt_pass::rtl_opt_pass): New.
3610 (ipa_opt_pass_d): Convert to subclass of opt_pass.
3611 (ipa_opt_pass_d::ipa_opt_pass_d): New.
3612 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
3613 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
3614 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
3615 invocation of pass_mode_switching to reflect this moving from a
3616 global to a member of gcc::pass_manager.
3617 (ix86_option_override): Rework how pass_insert_vzeroupper is
3618 added to the pass_manager to reflect autogenerated changes.
3619 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
3621 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
3623 PR rtl-optimization/57708
3624 * recog.c (peep2_find_free_register): Validate all regs in a
3627 2013-08-05 Jan Hubicka <jh@suse.cz>
3630 * cgraph.c (verify_cgraph_node): Accept local flags from other
3632 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
3633 (function_and_variable_visibility): Likewise.
3634 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
3636 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3638 * graph.c (init_graph_slim_pretty_print): Remove.
3639 (print_graph_cfg): Do not call it. Use local pretty printer.
3640 (start_graph_dump): Likewise.
3642 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3644 * gimple-pretty-print.c (buffer): Remove.
3645 (initialized): Likewise.
3646 (maybe_init_pretty_print): Likewise.
3647 (print_gimple_stmt): Do not call it. Use non-static local
3648 pretty_printer variable.
3649 (print_gimple_expr): Likewise.
3650 (print_gimple_seq): Likewise.
3651 (gimple_dump_bb): Likewise.
3653 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
3655 * asan.c (asan_pp): Remove.
3656 (asan_pp_initialized): Likewise.
3657 (asan_pp_initialize): Likewise.
3658 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
3659 (asan_emit_stack_protection): Tidy. Use local pretty printer.
3660 (asan_add_global): Likewise.
3662 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
3664 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
3665 * diagnostic.h (diagnostic_flush_buffer): Adjust.
3666 * pretty-print.c (pp_formatted_text_data): Likewise.
3667 (pp_indent): Rename from pp_base_indent.
3668 (pp_format): Rename from pp_base_format.
3669 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
3670 (pp_format_verbatim): Rename from pp_base_format_verbatim.
3671 (pp_flush): Rename from pp_base_flush.
3672 (pp_set_line_maximum_length): Rename from
3673 pp_base_set_line_maximum_length.
3674 (pp_clear_output_area): Rename from pp_base_clear_output_area.
3675 (pp_set_prefix): Rename from pp_base_set_prefix.
3676 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
3677 (pp_emit_prefix): Rename from pp_base_emit_prefix.
3678 (pp_append_text): Rename from pp_base_append_text.
3679 (pp_formatted_text): Rename from pp_base_formatted_text.
3680 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
3681 (pp_remaining_character_count_for_line): Rename from
3682 pp_base_remaining_character_count_for_line.
3683 (pp_newline): Rename from pp_base_newline.
3684 (pp_character): Rename from pp_base_character.
3685 (pp_string): Rename from pp_base_string.
3686 (pp_maybe_space): Rename from pp_base_maybe_space.
3687 * asan.c (asan_pp_string): Adjust.
3688 (asan_emit_stack_protection): Likewise.
3689 (asan_add_global): Likewise.
3690 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
3691 * tree-mudflap.c (mf_varname_tree): Likewise.
3692 * tree-pretty-print.c (pp_tree_identifier): Rename from
3693 pp_base_tree_identifier.
3694 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
3695 Declare as function.
3697 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3699 * pretty-print.h (pp_bar_bar): New.
3700 (pp_ampersand_ampersand): Likewise.
3701 (pp_less_equal): Likewise.
3702 (pp_greater_equal): Likewise.
3703 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
3704 printer functions instead of pp_string or operators and punctuators.
3705 (dump_gimple_call): Likewise.
3706 (dump_gimple_omp_for): Likewise.
3707 (dump_gimple_transaction): Likewise.
3708 (dump_gimple_phi): Likewise.
3709 (pp_gimple_stmt_1): Likewise.
3710 * sched-vis.c (print_insn): Likewise.
3711 * tree-mudflap.c (mf_varname_tree): Likewise.
3712 * tree-pretty-print.c (dump_block_node): Likewise.
3713 (dump_generic_node): Likewise.
3715 2013-08-02 Jan Hubicka <jh@suse.cz>
3717 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
3719 * lto-streamer-out.c (tree_is_indexable): Results decls and
3720 parm decls are not indexable.
3721 (DFS_write_tree_body): Do not follow args and results.
3722 (hash_tree): Likewise.
3723 (output_functions): Rearrange so struct function is needed
3724 only when real body is output; be able to also ouptut abstract
3725 functions; output DECL_ARGUMENTS and DECL_RESULT.
3726 (lto_output): When not in WPA, ale store abstract functions.
3727 (write_symbol): Do not care about RESULT_DECL.
3728 (output_symbol_p): Handle correctly sbtract decls.
3729 * lto-streamer-in.c (input_function): Rearrange so struct
3730 function can be NULL at entry; allow streaming of
3731 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
3732 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
3733 sanity check during LTO.
3734 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
3735 RESULT_DECl and DECL_ARGUMENTS.
3736 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
3739 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
3741 * pretty-print.h (pp_underscore): New.
3743 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
3744 printer functions instead of pp_character.
3745 (dump_binary_rhs): Likewise.
3746 (dump_ternary_rhs): Likewise.
3747 (dump_gimple_call_args): Likewise.
3748 (pp_points_to_solution): Likewise.
3749 (dump_gimple_call): Likewise.
3750 (dump_gimple_switch): Likewise.
3751 (dump_gimple_cond): Likewise.
3752 (dump_gimple_bind): Likewise.
3753 (dump_gimple_try): Likewise.
3754 (dump_gimple_omp_for): Likewise.
3755 (dump_gimple_omp_continue): Likewise.
3756 (dump_gimple_omp_single): Likewise.
3757 (dump_gimple_omp_sections): Likewise.
3758 (dump_gimple_omp_block): Likewise.
3759 (dump_gimple_omp_critical): Likewise.
3760 (dump_gimple_transaction): Likewise.
3761 (dump_gimple_asm): Likewise.
3762 (dump_gimple_phi): Likewise.
3763 (dump_gimple_omp_parallel): Likewise.
3764 (dump_gimple_omp_task): Likewise.
3765 (dump_gimple_omp_atomic_load): Likewise.
3766 (dump_gimple_omp_atomic_store): Likewise.
3767 (dump_gimple_mem_ops): Likewise.
3768 (pp_gimple_stmt_1): Likewise.
3769 (pp_cfg_jump): Likewise.
3770 (dump_implicit_edges): Likewise.
3771 (gimple_dump_bb_for_graph): Likewise.
3772 * graph.c (draw_cfg_node): Likewise.
3773 * langhooks.c (lhd_print_error_function): Likewise.
3774 * sched-vis.c (print_exp): Likewise.
3775 (print_value): Likewise.
3776 (print_pattern): Likewise.
3777 (print_insn): Likewise.
3778 (rtl_dump_bb_for_graph): Likewise.
3779 * tree-pretty-print.c (dump_function_declaration): Likewise.
3780 (dump_array_domain): Likewise.
3781 (dump_omp_clause): Likewise.
3782 (dump_location): Likewise.
3783 (dump_generic_node): Likewise.
3784 (print_struct_decl): Likewise.
3785 * diagnostic.c (diagnostic_show_locus): Use pp_space.
3787 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
3789 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
3790 candidate table when replacing a candidate statement.
3791 (replace_rhs_if_not_dup): Likewise.
3792 (replace_one_candidate): Likewise.
3794 2013-08-02 Jan Hubicka <jh@suse.cz>
3795 Martin Liska <marxin.liska@gmail.com>
3797 * cgraphunit.c (add_new_function): Fix logic when adding from
3799 (assemble_thunk): Rename to ...
3800 (expand_thunk); .. this one; export; get it working with
3801 general functions; make produced gimple valid.
3802 * cgraph.h (expand_thunk): Declare.
3804 2013-08-02 Jan Hubicka <jh@suse.cz>
3806 * ipa-cp.c (gather_context_independent_values): Use
3807 ipa_get_param_move_cost.
3808 (get_replacement_map): Remove PARAM; move parameter folding
3810 (create_specialized_node): Update.
3811 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
3812 assert that we have gimple body; update move_cost.
3813 (count_formal_params): Assert that we have gimple body.
3814 (ipa_dump_param): New function.
3815 (ipa_alloc_node_params): Break out from ...
3816 (ipa_initialize_node_params): ... here.
3817 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
3818 (ipa_write_node_info): Stream move costs.
3819 (ipa_read_node_info): Read move costs.
3820 (ipa_update_after_lto_read): Do not recompute node params.
3821 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
3822 (ipa_get_param): Check we are not in WPA.
3823 (ipa_get_param_move_cost): New.
3824 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
3825 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
3826 parm numbers to be present.
3828 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3830 PR rtl-optimization/58048
3831 * lra-constraints.c (process_alt_operands): Don't check asm
3832 operand on register.
3834 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
3836 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
3837 the implied StoreLoad barrier for atomic operations if before.
3839 2013-08-02 Jan Hubicka <jh@suse.cz>
3840 Martin Liska <marxin.liska@gmail.com>
3842 * cgraph.c (cgraph_function_body_availability): Do not check
3844 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
3845 symtab_node_availability): Declare.
3846 * ipa.c (can_replace_by_local_alias): New.
3847 (function_and_variable_visibility): Use it.
3848 * symtab.c (symtab_for_node_and_aliases,
3849 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
3851 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
3853 PR rtl-optimization/57963
3854 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
3855 (lra_constraints): Use them.
3857 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
3859 * config/arm/types.md (define_attr "type"): Add "load_acq"
3861 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
3863 (cortex_a53_store1): Likewise.
3865 2013-08-01 Jan Hubicka <jh@suse.cz>
3867 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
3868 partitions are not needed.
3870 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
3872 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
3873 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
3874 MAYBE_NON_Q_CLASS_P where appropriate.
3876 2013-08-01 Jan Hubicka <jh@suse.cz>
3878 * cgraph.h (release_function_body): Declare.
3879 * tree.c (free_lang_data_in_decl): Free, parameters and return values
3880 of unused delcarations.
3882 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3884 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
3885 RTL form when subtracting a constant.
3887 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3889 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
3890 Generate canonical plus rtx with negated immediate instead of minus
3892 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
3894 2013-08-01 Jan Hubicka <jh@suse.cz>
3896 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
3897 (cgraph_release_function_body): Likewise.
3898 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
3899 * cgraph.h (cgrpah_node): Rename abstract_and_needed
3900 to used_as_abstract_origin.
3901 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
3902 symbols used as abstract origins.
3903 * cgraphunit.c (analyze_functions): Update.
3904 * ipa.c (symtab_remove_unreachable_nodes): Recompute
3905 used_as_abstract_origin.
3906 * tree-inline.c (tree_function_versioning): Update
3907 used_as_abstract_origin; be ready for DECL_RESULT and
3908 DECL_ARGUMENTS to be NULL.
3910 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
3911 for abstract functions.
3912 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
3915 2013-08-01 Jan Hubicka <jh@suse.cz>
3917 * profile.c (compute_value_histograms): Fix thinko.
3919 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3921 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
3922 aarch-common-protos.h to extra_headers.
3923 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3924 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
3925 * config/aarch64/t-aarch64 (aarch-common.o): Define.
3927 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
3929 * config/aarch64/aarch64.md (define_attr "type"): Delete.
3930 Include "../arm/types.md". Define "type" attribute for all patterns.
3931 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
3934 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
3936 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
3937 to support power8 load fusion.
3938 (fusion_gpr_mem_load): Likewise.
3940 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
3942 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
3943 declarations for power8 load fusion.
3944 (emit_fusion_gpr_load): Likewise.
3946 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
3947 tuning for power8, turn on fusion mode by default. Turn on sign
3948 extending fusion mode if normal fusion mode is on, and we are at
3950 (fusion_gpr_load_p): New function, return true if we can fuse an
3951 addis instruction with a dependent load to a GPR.
3952 (emit_fusion_gpr_load): Emit the instructions for power8 load
3955 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
3956 (VSX load fusion peepholes): New peepholes to fuse together an
3957 addi instruction with a VSX load instruction.
3959 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
3960 peepholes to fuse an addis instruction with a load to a GPR base
3961 register. If we are supporting sign extending fusions, convert
3962 sign extending loads to zero extending loads and add an explicit
3965 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3967 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
3968 aarch-common-protos.h to extra_headers.
3969 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
3970 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
3971 (arm_early_store_addr_dep): Likewise.
3972 (arm_no_early_alu_shift_dep): Likewise.
3973 (arm_no_early_alu_shift_value_dep): Likewise.
3974 (arm_no_early_mul_dep): Likewise.
3975 (arm_no_early_store_addr_dep): Likewise.
3976 (arm_mac_accumulator_is_mul_result): Likewise.
3977 (arm_mac_accumulator_is_result): Likewise.
3978 * config/arm/aarch-common.c: ... here. New file.
3979 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
3981 (arm_early_store_addr_dep): Likewise.
3982 (arm_no_early_alu_shift_dep): Likewise.
3983 (arm_no_early_alu_shift_value_dep): Likewise.
3984 (arm_no_early_mul_dep): Likewise.
3985 (arm_no_early_store_addr_dep): Likewise.
3986 (arm_mac_accumulator_is_mul_result): Likewise.
3987 (arm_mac_accumulator_is_result): Likewise.
3988 * config/arm/aarch-common-protos.h: ... here. New file.
3989 * config/arm/t-arm (aarch-common.o): Define.
3991 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
3993 * config/arm/arm.md: Include new file "types.md".
3994 (define_attr "type"): Move from here to ...
3995 (define_attr "mul32"): Likewise.
3996 (define_attr "mul64"): Likewise.
3997 * config/arm/types.md: ... here. New file.
3999 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
4001 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
4002 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
4004 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4006 * gen-pass-instances.awk: Fix offset of substr().
4008 2013-07-31 David Malcolm <dmalcolm@redhat.com>
4010 * Makefile.in (pass-instances.def): New.
4011 (passes.o): Replace dependency on passes.def with one on
4014 * gen-pass-instances.awk: New.
4016 * passes.c (pass_manager::pass_manager): Use pass-instances.def
4017 rather than passes.def, updating local definition of NEXT_PASS
4018 macro to add an extra NUM parameter (currently unused).
4020 2013-07-30 David Malcolm <dmalcolm@redhat.com>
4022 * Makefile.in (PASS_MANAGER_H): New.
4023 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
4024 (passes.o): Likewise.
4025 (statistics.o): Likewise.
4026 (cgraphunit.o): Likewise.
4027 (context.o): Depend on PASS_MANAGER_H.
4029 * pass_manager.h: New.
4031 * cgraphunit.c (cgraph_add_new_function): Update for moves
4032 of globals to fields of pass_manager.
4033 (analyze_function): Likewise.
4034 (expand_function): Likewise.
4035 (ipa_passes): Likewise.
4036 (compile): Likewise.
4038 * context.c (context::context): New.
4039 * context.h (context::context): New.
4040 (context::get_passes): New.
4041 (context::passes_): New.
4043 * lto-cgraph.c (input_node): Update for moves of globals to
4044 fields of pass_manager.
4046 * passes.c (all_passes): Remove, in favor of a field of the
4047 same name within the new class pass_manager.
4048 (all_small_ipa_passes): Likewise.
4049 (all_lowering_passes): Likewise.
4050 (all_regular_ipa_passes): Likewise.
4051 (all_late_ipa_passes): Likewise.
4052 (all_lto_gen_passes): Likewise.
4053 (passes_by_id): Likewise.
4054 (passes_by_id_size): Likewise.
4055 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4056 the new class pass_manager.
4057 (set_pass_for_id): Convert to...
4058 (pass_manager::set_pass_for_id): ...method.
4059 (get_pass_for_id): Convert to...
4060 (pass_manager::get_pass_for_id): ...method.
4061 (register_one_dump_file): Move body of implementation into...
4062 (pass_manager::register_one_dump_file): ...here.
4063 (register_dump_files_1): Convert to...
4064 (pass_manager::register_dump_files_1): ...method.
4065 (register_dump_files): Convert to...
4066 (pass_manager::register_dump_files): ...method.
4067 (create_pass_tab): Update for moves of globals to fields of
4069 (dump_passes): Move body of implementation into...
4070 (pass_manager::dump_passes): ...here.
4071 (register_pass): Move body of implementation into...
4072 (pass_manager::register_pass): ...here.
4073 (init_optimization_passes): Convert into...
4074 (pass_manager::pass_manager): ...constructor for new
4075 pass_manager class, and initialize the pass_lists array.
4076 (check_profile_consistency): Update for moves of globals to
4077 fields of pass_manager.
4078 (dump_profile_report): Move body of implementation into...
4079 (pass_manager::dump_profile_report): ...here.
4080 (ipa_write_summaries_1): Update for moves of pass lists from
4081 being globals to fields of pass_manager.
4082 (ipa_write_optimization_summaries): Likewise.
4083 (ipa_read_summaries): Likewise.
4084 (ipa_read_optimization_summaries): Likewise.
4085 (execute_all_ipa_stmt_fixups): Likewise.
4087 * statistics.c (statistics_fini): Update for moves of globals to
4088 fields of pass_manager.
4090 * toplev.c (general_init): Replace call to
4091 init_optimization_passes with construction of the pass_manager
4094 * tree-pass.h (all_passes): Remove, in favor of a field of the
4095 same name within the new class pass_manager.
4096 (all_small_ipa_passes): Likewise.
4097 (all_lowering_passes): Likewise.
4098 (all_regular_ipa_passes): Likewise.
4099 (all_lto_gen_passes): Likewise.
4100 (all_late_ipa_passes): Likewise.
4101 (passes_by_id): Likewise.
4102 (passes_by_id_size): Likewise.
4103 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
4104 the new class pass_manager.
4105 (get_pass_for_id): Remove.
4107 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4109 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
4112 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
4114 * arm.md (mulhi3): New expand pattern.
4116 2013-07-30 Jan Hubicka <jh@suse.cz>
4117 Martin Liska <marxin.liska@gmail.com>
4119 * profile.c (compute_value_histograms): Do not ICE when
4120 there is mismatch only on some counters.
4122 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4124 PR rtl-optimization/57637
4125 * function.c (move_insn_for_shrink_wrap): Also check the
4126 GEN set of the LIVE problem for the liveness analysis
4127 if it exists, otherwise give up.
4129 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4131 PR tree-optimization/57993
4132 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
4133 replaced statement in the candidate table.
4134 (phi_add_costs): Return infinite cost when the hidden basis does
4135 not dominate all phis on which the candidate is dependent.
4136 (replace_one_candidate): Record replaced statement in the
4139 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
4141 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
4142 (ashlv2si3): New expander.
4143 (*ashlv2si3_i): New define_insn_and_split.
4144 * predicates.md (float_operation): Allow patterns with three
4147 PR rtl-optimization/58021
4148 * mode-switching.c (create_pre_exit): Always split off preceding
4149 insns if we are not at the basic block head.
4151 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
4153 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
4154 (UCLIBC_DYNAMIC_LINKER): New macro.
4155 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
4157 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
4158 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
4160 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
4161 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
4162 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
4163 for SF and DF modes. Use ieee_quad_format for TF mode.
4164 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
4165 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
4166 (mips_option_override): Handle `-mnan=legacy'.
4167 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
4168 `-mabs=2008' and `-mnan=2008'.
4169 (OPTION_DEFAULT_SPECS): Add "nan" default.
4170 (ASM_SPEC): Handle `-mnan='.
4171 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
4172 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
4173 comment accordingly.
4174 (neg<mode>2): Likewise.
4175 * config/mips/mips.opt (mabs, mnan): New options.
4176 * doc/install.texi (Configuration): Document `--with-nan=' option.
4177 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
4179 (MIPS Options): Document them.
4180 * config.gcc <mips*-*-*>: Handle `--with-nan='.
4181 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
4182 * configure: Regenerate.
4183 * config.in: Regenerate.
4185 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4187 * config/i386/i386.md (float post-reload splitters): Do not check
4188 for subregs of SSE registers.
4190 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
4191 H.J. Lu <hongjiu.lu@intel.com>
4195 * config/i386/i386.md (post-reload splitter
4196 to avoid partial SSE reg dependency stalls): New pattern.
4198 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
4200 * config/s390/s390.md ("movcc"): Swap load and store instructions.
4202 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
4204 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
4205 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
4207 2013-07-26 Cary Coutant <ccoutant@google.com>
4209 * dwarf2out.c (die_checksum_ordered): Don't include template
4210 instantiations in signature.
4211 (is_template_parameter): New function.
4212 (is_template_instantiation): New function.
4213 (generate_skeleton_bottom_up): Don't include template instantiations
4215 (generate_skeleton): Likewise.
4216 (break_out_comdat_types): Move recursive call to break out nested
4218 (prune_unused_types_mark_generic_parms_dies): Call
4219 is_template_parameter.
4221 2013-07-26 Ian Bolton <ian.bolton@arm.com>
4223 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
4224 uses vector registers.
4225 * config/aarch64/iterators.md: Add attributes rtn and vas.
4227 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4228 Richard Earnshaw <richard.earnshaw@arm.com>
4230 * combine.c (simplify_comparison): Re-canonicalize operands
4232 * config/arm/arm.md (movcond_addsi): New splitter.
4234 2013-07-25 Sterling Augustine <saugustine@google.com>
4236 * dwarf2out.c (size_of_pubnames): Move code to...
4237 (include_pubname_in_output): ...here. New.
4238 (want_pubnames): Rearrange.
4239 (output_pubnames): Call include_pubname_in_output. Move assertion.
4241 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4243 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
4245 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
4248 * doc/extend.texi: Remove obsolete builtins. Fix
4249 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
4251 2013-07-25 Jan Hubicka <jh@suse.cz>
4253 * cgraph.c (release_function_body): Break out from ...
4254 (cgraph_release_function_body): ... this one; also release DECL_RESULT
4256 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
4257 old_tree in the map.
4258 (create_specialized_node): Update.
4259 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
4261 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
4262 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
4263 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
4264 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
4267 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4269 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
4270 addsi3_carryin_alt2_<optab>): Correct output template.
4272 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4274 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
4275 Adjust for arm_restrict_it.
4276 Remove trailing whitespace.
4278 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
4280 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
4281 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
4283 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
4285 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
4287 PR rtl-optimization/57960
4288 * lra-constraints.c (process_alt_operands): Use the right mode
4289 when checking strict_low.
4291 2013-07-25 Jan Hubicka <jh@suse.cz>
4293 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
4294 * cgraph.c (cgraph_remove_node): Do not release function body
4295 when in cgraph streaming.
4296 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
4297 in other partitions are not considered reachable; fix handling of
4300 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4302 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
4304 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4309 * config/arm/arm.md ("*sibcall_insn): Replace use of
4310 Ss with US. Adjust output for v5 and v4t.
4311 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
4313 * config/arm/constraints.md ("Ss"): Rename to US.
4315 2013-07-25 Terry Guo <terry.guo@arm.com>
4317 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
4318 shift_add/shift_sub0/shift_sub1 RTXs.
4320 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
4321 Anton Blanchard <anton@au1.ibm.com>
4323 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
4324 (altivec_vpks<VI_char>ss): Likewise.
4325 (altivec_vpks<VI_char>us): Likewise.
4326 (altivec_vpku<VI_char>us): Likewise.
4327 (altivec_vpku<VI_char>um): Likewise.
4329 2013-07-24 David Malcolm <dmalcolm@redhat.com>
4331 Introduce context class.
4333 * Makefile.in (CONTEXT_H): New.
4334 (OBJS): Add context.o.
4335 (toplev.o): Add CONTEXT_H to dependencies.
4338 * toplev.c (general_init): Create the singleton gcc::context instance.
4344 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
4346 PR rtl-optimization/57968
4347 * mode-switching.c (create_pre_exit): Allow instructions that
4348 don't set a return register to need a non-exit mode.
4350 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4351 Anton Blanchard <anton@au1.ibm.com>
4353 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
4354 operands to vperm for little endian.
4355 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
4356 of lvsl to create the control mask for a vperm for little endian.
4358 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4359 Anton Blanchard <anton@au1.ibm.com>
4361 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
4362 two operands for little-endian.
4364 2013-07-23 Steve Ellcey <sellcey@mips.com>
4366 * config/mips/mips.c (mips_case_values_threshold): New.
4367 (TARGET_CASE_VALUES_THRESHOLD): Define.
4369 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4370 Anton Blanchard <anton@au1.ibm.com>
4372 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
4373 selection of field for vector splat in little endian mode.
4375 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4377 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
4378 expanders to rs6000.md.
4379 (ior<mode>3): Likewise.
4380 (and<mode>3): Likewise.
4381 (one_cmpl<mode>2): Likewise.
4382 (nor<mode>3): Likewise.
4383 (andc<mode>3): Likewise.
4384 (eqv<mode>3): Likewise.
4385 (nand<mode>3): Likewise.
4386 (orc<mode>3): Likewise.
4388 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
4391 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
4392 to split multi-word logical operations.
4393 (rs6000_split_logical_di): Likewise.
4394 (rs6000_split_logical): Likewise.
4396 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
4397 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
4398 and allow TImode operations in 32-bit.
4399 (vsx_and<mode>3_64bit): Likewise.
4400 (vsx_ior<mode>3_32bit): Likewise.
4401 (vsx_ior<mode>3_64bit): Likewise.
4402 (vsx_xor<mode>3_32bit): Likewise.
4403 (vsx_xor<mode>3_64bit): Likewise.
4404 (vsx_one_cmpl<mode>2_32bit): Likewise.
4405 (vsx_one_cmpl<mode>2_64bit): Likewise.
4406 (vsx_nor<mode>3_32bit): Likewise.
4407 (vsx_nor<mode>3_64bit): Likewise.
4408 (vsx_andc<mode>3_32bit): Likewise.
4409 (vsx_andc<mode>3_64bit): Likewise.
4410 (vsx_eqv<mode>3_32bit): Likewise.
4411 (vsx_eqv<mode>3_64bit): Likewise.
4412 (vsx_nand<mode>3_32bit): Likewise.
4413 (vsx_nand<mode>3_64bit): Likewise.
4414 (vsx_orc<mode>3_32bit): Likewise.
4415 (vsx_orc<mode>3_64bit): Likewise.
4417 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
4418 logical types in GPRs.
4420 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
4421 logical insns to rs6000.md, and allow TImode operations in
4423 (altivec_ior<mode>3): Likewise.
4424 (altivec_xor<mode>3): Likewise.
4425 (altivec_one_cmpl<mode>2): Likewise.
4426 (altivec_nor<mode>3): Likewise.
4427 (altivec_andc<mode>3): Likewise.
4429 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
4430 attributes for moving the 128-bit logical operations into
4432 (BOOL_REGS_OUTPUT): Likewise.
4433 (BOOL_REGS_OP1): Likewise.
4434 (BOOL_REGS_OP2): Likewise.
4435 (BOOL_REGS_UNARY): Likewise.
4436 (BOOL_REGS_AND_CR0): Likewise.
4437 (one_cmpl<mode>2): Add support for DI logical operations on
4438 32-bit, splitting the operations to 32-bit.
4442 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
4443 changes to combine the 32/64-bit code, allow logical operations on
4444 TI mode in 32-bit, and to use similar match_operator patterns like
4445 scalar mode uses. Combine the Altivec and VSX code for logical
4446 operations, and move it here.
4447 (ior<mode>3, 128-bit types): Likewise.
4448 (xor<mode>3, 128-bit types): Likewise.
4449 (one_cmpl<mode>3, 128-bit types): Likewise.
4450 (nor<mode>3, 128-bit types): Likewise.
4451 (andc<mode>3, 128-bit types): Likewise.
4452 (eqv<mode>3, 128-bit types): Likewise.
4453 (nand<mode>3, 128-bit types): Likewise.
4454 (orc<mode>3, 128-bit types): Likewise.
4455 (and<mode>3_internal): Likewise.
4456 (bool<mode>3_internal): Likewise.
4457 (boolc<mode>3_internal1): Likewise.
4458 (boolc<mode>3_internal2): Likewise.
4459 (boolcc<mode>3_internal1): Likewise.
4460 (boolcc<mode>3_internal2): Likewise.
4461 (eqv<mode>3_internal1): Likewise.
4462 (eqv<mode>3_internal2): Likewise.
4463 (one_cmpl1<mode>3_internal): Likewise.
4465 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4467 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4468 Rename flag_stack_usage to flag_stack_usage_info.
4470 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
4472 * config/microblaze/sync.md: New file.
4473 * config/microblaze/microblaze.md: Include sync.md
4474 * config/microblaze/microblaze.c: Add print_operand 'y'.
4475 * config/microblaze/constraints.md: Add memory_contraint
4476 'Q' which is a single register.
4478 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
4480 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
4482 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4484 * reload.c (find_reloads): Exit loop once we find this operand
4485 cannot be reloaded somehow for this alternative.
4487 * reload.c (find_reloads): Exit loop once we find a hard register.
4489 * rtlanal.c (computed_jump_p): Exit loop once we find label
4492 * i386.c (ix86_pad_returns): Exit loop after setting replace.
4494 * cfgloopmanip.c (remove_path): Exit loop after setting
4497 * gensupport.c (subst_dup): Avoid loop if code is not
4498 MATCH_DUP nor MATCH_OP_DUP.
4500 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
4502 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
4504 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4506 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
4507 true for SP_REGNUM if mode == ptr_mode.
4508 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
4509 with value R0_REGNUM + 31.
4511 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4513 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
4514 pad pointer-typed argument downward.
4516 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4518 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
4519 and __ILP32__ when the ILP32 model is in use.
4521 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4523 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
4524 (aarch64_load_symref_appropriately): In the case of
4525 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
4526 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
4527 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
4528 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
4529 if 'mode' doesn't equal to 'ptr_mode'.
4530 (aarch64_output_mi_thunk): Add an assertion on the alignment of
4531 'vcall_offset'; change to call aarch64_emit_move differently depending
4532 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
4533 to calculate the upper bound of 'vcall_offset'.
4534 (aarch64_cannot_force_const_mem): Change to also return true if
4536 (aarch64_legitimize_reload_address): In the case of large
4537 displacements, add new local variable 'xmode' and an assertion
4538 based on it; change to use 'xmode' to generate the new rtx and
4540 (aarch64_asm_trampoline_template): Change to generate the template
4541 differently depending on TARGET_ILP32 or not; change to use
4542 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
4543 (aarch64_trampoline_size): Removed.
4544 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
4545 and replace immediate literals with it. Change to use 'ptr_mode'
4546 instead of 'DImode' and call convert_memory_address if the mode
4547 of 'fnaddr' doesn't equal to 'ptr_mode'.
4548 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
4550 (aarch64_elf_asm_destructor): Likewise.
4551 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
4552 on TARGET_ILP32 instead of aarch64_trampoline_size.
4553 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
4554 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
4555 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
4556 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
4557 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
4558 (storewb_pair<GPI:mode>_<P:mode>): ... this.
4559 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
4560 depending on the value of 'mode'.
4561 (add_losym_<mode>): New.
4562 (ldr_got_small_<mode>): New, based on ldr_got_small.
4563 (ldr_got_small): Remove.
4564 (ldr_got_small_sidi): New.
4565 * config/aarch64/iterators.md (P): New.
4566 (PTR): Change to 'ptr_mode' in the condition.
4568 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
4570 * config.gcc (aarch64*-*-*): Support --with-abi.
4571 (aarch64*-*-elf): Support --with-multilib-list.
4572 (aarch64*-*-linux*): Likewise.
4573 (supported_defaults): Add abi to aarch64*-*-*.
4574 * configure.ac: Mention AArch64 for --with-multilib-list.
4575 * configure: Re-generated.
4576 * config/aarch64/biarchilp32.h: New file.
4577 * config/aarch64/biarchlp64.h: New file.
4578 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
4580 (MULTILIB_DEFAULTS): Ditto.
4581 (DRIVER_SELF_SPECS): Ditto.
4582 (ASM_SPEC): Update to also substitute -mabi.
4583 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
4584 file whose name depends on -mabi= and -mbig-endian.
4585 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
4587 (POINTER_SIZE): New define.
4588 (POINTERS_EXTEND_UNSIGNED): Ditto.
4589 (enum aarch64_abi_type): New enumeration tag.
4590 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
4591 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
4592 (TARGET_ILP32): New define.
4593 * config/aarch64/aarch64.opt (mabi): New.
4595 (ilp32, lp64): New values for -mabi.
4596 * config/aarch64/t-aarch64 (comma): New define.
4597 (MULTILIB_OPTIONS): Ditto.
4598 (MULTILIB_DIRNAMES): Ditto.
4599 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
4600 * doc/invoke.texi: Document -mabi for AArch64.
4602 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
4604 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
4606 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
4607 Anton Blanchard <anton@au1.ibm.com>
4609 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
4610 endianness when selecting field to splat.
4612 2013-07-22 Eric Christopher <echristo@gmail.com>
4614 * dwarf2out.c (die_odr_checksum): New function to use
4615 CHECKSUM_ macros and ULEB128 for DIE tag.
4616 (generate_type_signature): Use.
4618 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
4620 * config.gcc (sparc*-*-*): Accept leon3 processor.
4621 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
4622 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
4623 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
4624 * config/sparc/sparc.opt (enum processor_type): Add leon3.
4625 (mfix-ut699): Adjust comment.
4626 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
4627 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
4628 (CPP_CPU_SPEC): Likewise.
4629 (ASM_CPU_SPEC): Likewise.
4630 * config/sparc/sparc.c (leon3_cost): New constant.
4631 (sparc_option_override): Add leon3 support.
4632 (mem_ref): New function.
4633 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
4634 (sparc_do_work_around_errata): Look into the instruction in the delay
4635 slot and adjust accordingly. Add fix for the data cache nullify issues
4636 of the UT699. Change insertion position for the NOP.
4637 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
4638 (leon3_load): New reservation.
4639 (leon_store): Bump latency to 2.
4640 (grfpu): New automaton.
4641 (grfpu_alu): New unit.
4642 (grfpu_ds): Likewise.
4643 (leon_fp_alu): Adjust.
4644 (leon_fp_mult): Delete.
4645 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
4646 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
4647 * config/sparc/sparc.md (cpu): Add leon3.
4648 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
4650 (atomic_test_and_set): Likewise.
4653 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
4655 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
4656 default for R5900 targets.
4657 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
4658 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
4659 * config/mips/mips.c (mips_option_override): Report an error for
4660 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
4661 for -march=r5900 -mhard-float.
4663 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
4665 * df-problems.c (can_move_insns_across): Exit loop once we
4666 find a non-fixed, non-global register.
4668 * ipa-pure-const.c (propagate_nothrow): Exit loop after
4671 * omega.c (omega_eliminate_red): Break after setting red_found.
4672 (omega_problem_has_red_equations): Similarly after setting found.
4673 (omega_query_variable): Similarly after setting coupled.
4675 2013-07-22 Marek Polacek <polacek@redhat.com>
4677 * gimplify.c: Don't include gimple.h twice.
4679 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4681 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
4682 instead of TARGET_THUMB1.
4683 (Pz): New constraint.
4684 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
4686 (compare_negsi_si): Likewise.
4687 (compare_addsi2_op0): Likewise.
4688 (compare_addsi2_op1): Likewise.
4689 (addsi3_carryin_<optab>): Likewise.
4690 (addsi3_carryin_alt2_<optab>): Likewise.
4691 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
4692 for arm_restrict_it.
4693 (subsi3_carryin): Likewise.
4694 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
4695 (minmax_arithsi): Disable for arm_restrict_it.
4696 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
4697 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
4698 (satsi_<SAT:code>_shift): Likewise.
4699 (arm_shiftsi3): Add alternative for 16-bit encoding.
4700 (arm32_movhf): Disable for arm_restrict_it.
4701 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
4702 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
4704 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
4706 * config/arm/arm.md (attribute "insn"): Delete.
4707 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
4708 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
4709 (not_shiftsi): Update for attribute change.
4710 (not_shiftsi_compare0): Likewise.
4711 (not_shiftsi_compare0_scratch): Likewise.
4712 (arm_one_cmplsi2): Likewise.
4713 (thumb1_one_cmplsi2): Likewise.
4714 (notsi_compare0): Likewise.
4715 (notsi_compare0_scratch): Likewise.
4716 (thumb1_movdi_insn): Likewise.
4717 (arm_movsi_insn): Likewise.
4718 (movhi_insn_arch4): Likewise.
4719 (movhi_bytes): Likewise.
4720 (arm_movqi_insn): Likewise.
4721 (thumb1_movqi_insn): Likewise.
4722 (arm32_movhf): Likewise.
4723 (thumb1_movhf): Likewise.
4724 (arm_movsf_soft_insn): Likewise.
4725 (thumb1_movsf_insn): Likewise.
4726 (thumb_movdf_insn): Likewise.
4727 (movsicc_insn): Likewise.
4728 (movsfcc_soft_insn): Likewise.
4729 (and_scc): Likewise.
4730 (cond_move): Likewise.
4731 (if_move_not): Likewise.
4732 (if_not_move): Likewise.
4733 (if_shift_move): Likewise.
4734 (if_move_shift): Likewise.
4735 (if_shift_shift): Likewise.
4736 (if_not_arith): Likewise.
4737 (if_arith_not): Likewise.
4738 (cond_move_not): Likewise.
4739 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
4740 (neon_mov<mode>): Likewise.
4741 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
4742 (thumb2_movsi_vfp): Likewise.
4743 (movsf_vfp): Likewise.
4744 (thumb2_movsf_vfp): Likewise.
4745 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
4747 (cortexa7_older_only): Likewise.
4748 (cortexa7_younger): Likewise.
4749 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
4750 (1020alu_shift_op): Likewise.
4751 (1020alu_shift_reg_op): Likewise.
4752 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
4753 (alu_shift_op): Likewise.
4754 (alu_shift_reg_op): Likewise.
4755 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
4756 (11_alu_shift_op): Likewise.
4757 (11_alu_shift_reg_op): Likewise.
4758 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
4759 (9_alu_shift_reg_op): Likewise.
4760 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
4762 (cortex_a15_alu_shift): Likewise.
4763 (cortex_a15_alu_shift_reg): Likewise.
4764 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
4765 (cortex_a5_alu_shift): Likewise.
4766 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
4768 (cortex_a53_alu_shift): Likewise.
4769 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
4771 (cortex_a7_alu_reg): Likewise.
4772 (cortex_a7_alu_shift): Likewise.
4773 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4774 (cortex_a8_alu_shift): Likewise.
4775 (cortex_a8_alu_shift_reg): Likewise.
4776 (cortex_a8_mov): Likewise.
4777 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
4778 (cortex_a9_dp_shift): Likewise.
4779 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
4780 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
4781 (cortex_r4_mov): Likewise.
4782 (cortex_r4_alu_shift): Likewise.
4783 (cortex_r4_alu_shift_reg): Likewise.
4784 * config/arm/fa526.md (526_alu_op): Update for attribute change.
4785 (526_alu_shift_op): Likewise.
4786 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
4787 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
4788 (626te_alu_shift_op): Likewise.
4789 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
4790 (726te_alu_op): Likewise.
4791 (726te_alu_shift_op): Likewise.
4792 (726te_alu_shift_reg_op): Likewise.
4793 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
4794 (mp626_alu_shift_op): Likewise.
4795 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
4796 (pj4_alu_e1_conds): Likewise.
4797 (pj4_alu): Likewise.
4798 (pj4_alu_conds): Likewise.
4799 (pj4_shift): Likewise.
4800 (pj4_shift_conds): Likewise.
4801 (pj4_alu_shift): Likewise.
4802 (pj4_alu_shift_conds): Likewise.
4804 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4806 * config/arm/predicates.md (shiftable_operator_strict_it):
4808 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
4809 Disable cond_exec version for arm_restrict_it.
4810 (thumb2_smaxsi3): Convert to generate cond_exec.
4811 (thumb2_sminsi3): Likewise.
4812 (thumb32_umaxsi3): Likewise.
4813 (thumb2_uminsi3): Likewise.
4814 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
4815 (thumb2_neg_abssi2): Likewise.
4816 (thumb2_mov_scc): Add alternative for 16-bit encoding.
4817 (thumb2_movsicc_insn): Adjust alternatives.
4818 (thumb2_mov_negscc): Disable for arm_restrict_it.
4819 (thumb2_mov_negscc_strict_it): New pattern.
4820 (thumb2_mov_notscc_strict_it): New pattern.
4821 (thumb2_mov_notscc): Disable for arm_restrict_it.
4822 (thumb2_ior_scc): Likewise.
4823 (thumb2_ior_scc_strict_it): New pattern.
4824 (thumb2_cond_move): Adjust for arm_restrict_it.
4825 (thumb2_cond_arith): Disable for arm_restrict_it.
4826 (thumb2_cond_arith_strict_it): New pattern.
4827 (thumb2_cond_sub): Adjust for arm_restrict_it.
4828 (thumb2_movcond): Likewise.
4829 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
4830 (thumb2_zero_extendhisi2_v6): Likewise.
4831 (thumb2_zero_extendqisi2_v6): Likewise.
4832 (orsi_notsi_si): Likewise.
4833 (orsi_not_shiftsi_si): Likewise.
4835 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
4837 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
4838 instruction sequence is 1 byte shorter.
4840 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
4842 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
4843 it is not needed after split.
4845 2013-07-20 Iain Sandoe <iain@codesourcery.com>
4848 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
4849 second label for nonlocal goto receivers. Don't output pic base labels
4850 unless we're producing PIC; mark that action unreachable().
4851 (ix86_save_reg): If the function contains a nonlocal label, save the
4853 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
4854 * gcc/config/darwin.c (emitted_pic_label_num): New GTY.
4855 (update_pic_label_number_if_needed): New.
4856 (machopic_output_function_base_name): Adjust for nonlocal receiver
4858 (machopic_should_output_picbase_label): New.
4859 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
4860 (nonlocal_goto_receiver): New insn and split.
4862 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
4864 * config/aarch64/aarch64-builtins.c
4865 (aarch64_fold_builtin): Fold abs in all modes.
4866 * config/aarch64/aarch64-simd-builtins.def
4867 (abs): Enable for all modes.
4868 * config/aarch64/arm_neon.h
4869 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
4870 (vabs_f64): Add missing intrinsic.
4872 2013-07-19 Ian Bolton <ian.bolton@arm.com>
4874 * config/aarch64/arm_neon.h (vabs_s64): New function
4876 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
4879 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
4880 * config/avr/avr.md (adjust_len): Add `round'.
4881 * config/avr/avr-protos.h (avr_out_round): New prototype.
4882 (avr_out_plus): Add `out_label' argument.
4883 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
4884 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
4885 Handle the case where `insn' is just a pattern.
4886 (avr_out_bitop): Handle the case where `insn' is just a pattern.
4887 (avr_out_round): New function.
4888 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
4890 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
4892 * config/microblaze/microblaze.c (microblaze_expand_prologue):
4893 Add check for flag_stack_usage to handle -fstack-usage support
4895 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
4897 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
4898 interaction for new Power8 flags and VSX.
4900 2013-07-18 Sriraman Tallam <tmsriram@google.com>
4903 * tree-inline.c (expand_call_inline): Emit errors during
4904 early_inlining only if optimization is not turned on.
4906 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4910 * passes.c (init_optimization_passes): Move the construction of
4911 the pass hierarchy into a new passes.def file.
4913 * Makefile.in (passes.o): Add dependency on passes.def.
4915 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4917 * passes.c (init_optimization_passes): Introduce macros for
4918 constructing the tree of passes (INSERT_PASSES_AFTER,
4919 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
4920 TERMINATE_PASS_LIST).
4922 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
4923 Wei Mi <wmi@google.com>
4925 PR rtl-optimization/57878
4926 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
4928 (reload_pseudo_compare_func): Check nregs first for reload
4931 2013-07-18 David Malcolm <dmalcolm@redhat.com>
4933 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
4935 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
4937 * read-rtl.c (validate_const_int): Once an invalid character is
4938 seen, quit the loop.
4940 * gengtype.c (write_roots): Similarly once we find the "deletable"
4941 or "if_marked" option.
4943 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4945 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
4946 "xtab" and "sat". Move value "clz" from here to ...
4947 (attriubte "type"): ... here.
4948 (satsi_<SAT:code>): Delete "insn" attribute.
4949 (satsi_<SAT:code>_shift): Likewise.
4950 (arm_zero_extendqisi2addsi): Likewise.
4951 (arm_extendqisi2addsi): Likewise.
4952 (clzsi2): Update for attribute changes.
4953 (rbitsi2): Likewise.
4954 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
4956 (arm_usatsihi): Likewise.
4957 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
4959 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
4961 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
4962 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
4963 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
4964 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
4965 in alphabetical order.
4966 (attribute "core_cycles"): Update for attribute changes.
4967 (arm_addsi3): Likewise.
4968 (addsi3_compare0): Likewise.
4969 (addsi3_compare0_scratch): Likewise.
4970 (addsi3_compare_op1): Likewise.
4971 (addsi3_compare_op2): Likewise.
4972 (compare_addsi2_op0): Likewise.
4973 (compare_addsi2_op1): Likewise.
4974 (addsi3_carryin_shift_<optab>): Likewise.
4975 (subsi3_carryin_shift): Likewise.
4976 (rsbsi3_carryin_shift): Likewise.
4977 (arm_subsi3_insn): Likewise.
4978 (subsi3_compare0): Likewise.
4979 (subsi3_compare): Likewise.
4980 (arm_andsi3_insn): Likewise.
4981 (thumb1_andsi3_insn): Likewise.
4982 (andsi3_compare0): Likewise.
4983 (andsi3_compare0_scratch): Likewise.
4984 (zeroextractsi_compare0_scratch
4985 (andsi_not_shiftsi_si): Likewise.
4986 (iorsi3_insn): Likewise.
4987 (iorsi3_compare0): Likewise.
4988 (iorsi3_compare0_scratch): Likewise.
4989 (arm_xorsi3): Likewise.
4990 (thumb1_xorsi3_insn): Likewise.
4991 (xorsi3_compare0): Likewise.
4992 (xorsi3_compare0_scratch): Likewise.
4993 (satsi_<SAT:code>_shift): Likewise.
4995 (arm_shiftsi3): Likewise.
4996 (shiftsi3_compare0): Likewise.
4997 (not_shiftsi): Likewise.
4998 (not_shiftsi_compare0): Likewise.
4999 (not_shiftsi_compare0_scratch): Likewise.
5000 (arm_one_cmplsi2): Likewise.
5001 (thumb_one_complsi2): Likewise.
5002 (notsi_compare0): Likewise.
5003 (notsi_compare0_scratch): Likewise.
5004 (thumb1_zero_extendhisi2): Likewise.
5005 (arm_zero_extendhisi2): Likewise.
5006 (arm_zero_extendhisi2_v6): Likewise.
5007 (arm_zero_extendhisi2addsi): Likewise.
5008 (thumb1_zero_extendqisi2): Likewise.
5009 (thumb1_zero_extendqisi2_v6): Likewise.
5010 (arm_zero_extendqisi2): Likewise.
5011 (arm_zero_extendqisi2_v6): Likewise.
5012 (arm_zero_extendqisi2addsi): Likewise.
5013 (thumb1_extendhisi2): Likewise.
5014 (arm_extendhisi2): Likewise.
5015 (arm_extendhisi2_v6): Likewise.
5016 (arm_extendqisi): Likewise.
5017 (arm_extendqisi_v6): Likewise.
5018 (arm_extendqisi2addsi): Likewise.
5019 (thumb1_extendqisi2): Likewise.
5020 (thumb1_movdi_insn): Likewise.
5021 (arm_movsi_insn): Likewise.
5022 (movsi_compare0): Likewise.
5023 (movhi_insn_arch4): Likewise.
5024 (movhi_bytes): Likewise.
5025 (arm_movqi_insn): Likewise.
5026 (thumb1_movqi_insn): Likewise.
5027 (arm32_movhf): Likewise.
5028 (thumb1_movhf): Likewise.
5029 (arm_movsf_soft_insn): Likewise.
5030 (thumb1_movsf_insn): Likewise.
5031 (movdf_soft_insn): Likewise.
5032 (thumb_movdf_insn): Likewise.
5033 (arm_cmpsi_insn): Likewise.
5034 (cmpsi_shiftsi): Likewise.
5035 (cmpsi_shiftsi_swp): Likewise.
5036 (arm_cmpsi_negshiftsi_si): Likewise.
5037 (movsicc_insn): Likewise.
5038 (movsfcc_soft_insn): Likewise.
5039 (arith_shiftsi): Likewise.
5040 (arith_shiftsi_compare0
5041 (arith_shiftsi_compare0_scratch
5042 (sub_shiftsi): Likewise.
5043 (sub_shiftsi_compare0
5044 (sub_shiftsi_compare0_scratch
5045 (and_scc): Likewise.
5046 (cond_move): Likewise.
5047 (if_plus_move): Likewise.
5048 (if_move_plus): Likewise.
5049 (if_move_not): Likewise.
5050 (if_not_move): Likewise.
5051 (if_shift_move): Likewise.
5052 (if_move_shift): Likewise.
5053 (if_shift_shift): Likewise.
5054 (if_not_arith): Likewise.
5055 (if_arith_not): Likewise.
5056 (cond_move_not): Likewise.
5057 (thumb1_ashlsi3): Set type attribute.
5058 (thumb1_ashrsi3): Likewise.
5059 (thumb1_lshrsi3): Likewise.
5060 (thumb1_rotrsi3): Likewise.
5061 (shiftsi3_compare0_scratch): Likewise.
5062 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
5063 (neon_mov<mode>): Likewise.
5064 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
5066 (thumb2_movsi_insn): Likewise.
5067 (thumb2_cmpsi_neg_shiftsi): Likewise.
5068 (thumb2_extendqisi_v6): Likewise.
5069 (thumb2_zero_extendhisi2_v6): Likewise.
5070 (thumb2_zero_extendqisi2_v6): Likewise.
5071 (thumb2_shiftsi3_short): Likewise.
5072 (thumb2_addsi3_compare0_scratch): Likewise.
5073 (orsi_not_shiftsi_si): Likewise.
5074 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
5075 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
5077 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
5078 (1020alu_shift_op): Likewise.
5079 (1020alu_shift_reg_op): Likewise.
5080 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
5081 (alu_shift_op): Likewise.
5082 (alu_shift_reg_op): Likewise.
5083 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
5084 (11_alu_shift_op): Likewise.
5085 (11_alu_shift_reg_op): Likewise.
5086 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
5087 (9_alu_shift_reg_op): Likewise.
5088 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
5090 (cortex_a15_alu_shift): Likewise.
5091 (cortex_a15_alu_shift_reg): Likewise.
5092 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
5094 (cortex_a5_alu_shift): Likewise.
5095 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
5097 (cortex_a53_alu_shift): Likewise.
5098 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
5100 (cortex_a7_alu_reg): Likewise.
5101 (cortex_a7_alu_shift): Likewise.
5102 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
5104 (cortex_a8_alu_shift): Likewise.
5105 (cortex_a8_alu_shift_reg): Likewise.
5106 (cortex_a8_mov): Likewise.
5107 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
5108 (cortex_a9_dp_shift): Likewise.
5109 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
5111 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
5113 (cortex_r4_mov): Likewise.
5114 (cortex_r4_alu_shift): Likewise.
5115 (cortex_r4_alu_shift_reg): Likewise.
5116 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
5117 (526_alu_shift_op): Likewise.
5118 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
5119 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
5120 (626te_alu_shift_op): Likewise.
5121 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
5122 (726te_alu_op): Likewise.
5123 (726te_alu_shift_op): Likewise.
5124 (726te_alu_shift_reg_op): Likewise.
5125 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
5126 (mp626_alu_shift_op): Likewise.
5127 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
5128 (pj4_alu_e1_conds): Likewise.
5129 (pj4_alu): Likewise.
5130 (pj4_alu_conds): Likewise.
5131 (pj4_shift): Likewise.
5132 (pj4_shift_conds): Likewise.
5133 (pj4_alu_shift): Likewise.
5134 (pj4_alu_shift_conds): Likewise.
5135 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
5137 (cortexa7_older_only): Likewise.
5138 (cortexa7_younger): Likewise.
5140 2013-07-18 David Malcolm <dmalcolm@redhat.com>
5142 * ipa-pure-const.c (generate_summary): Rename to...
5143 (pure_const_generate_summary): ... this.
5145 2013-07-17 Iain Sandoe <iain@codesourcery.com>
5147 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
5149 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
5152 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
5155 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5157 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
5158 enabled without -march=zEC12.
5159 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
5162 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5164 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
5165 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
5166 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
5167 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
5168 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
5169 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
5170 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
5172 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
5173 and ISA_HAS_NMADD3_NMSUB3.
5174 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
5175 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
5176 (nmsub4<mode>, nmsub3<mode>): Likewise.
5177 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
5179 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5181 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
5182 TARGET_MIPS5400 checking.
5184 2013-07-16 Jakub Jelinek <jakub@redhat.com>
5185 Peter Bergner <bergner@vnet.ibm.com>
5187 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
5188 registers in the comment.
5189 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
5190 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
5191 rather than FIRST_PSEUDO_REGISTERS.
5193 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
5195 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
5196 enable extra ISA flags with TARGET_HTM.
5198 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
5200 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
5203 2013-07-15 Cong Hou <congh@google.com>
5205 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
5206 in compare function for sorting.
5208 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
5210 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
5211 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
5212 * config/rs6000/rs6000.opt: Add -mhtm option.
5213 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
5214 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
5215 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
5216 __HTM__ if the HTM instructions are available.
5217 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
5218 htm_spr_reg_operand): New define_predicates.
5219 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
5220 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
5222 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
5223 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
5224 HTM builtin functions.
5225 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
5226 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
5227 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
5228 (rs6000_builtin_mask_calculate): Likewise.
5229 (rs6000_option_override_internal): Likewise.
5230 (bdesc_htm): Add new HTM builtin support.
5231 (htm_spr_num): New function.
5232 (htm_spr_regno): Likewise.
5233 (rs6000_htm_spr_icode): Likewise.
5234 (htm_expand_builtin): Likewise.
5235 (htm_init_builtins): Likewise.
5236 (rs6000_expand_builtin): Add support for HTM builtin functions.
5237 (rs6000_init_builtins): Likewise.
5238 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
5240 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
5241 (TARGET_HTM, MASK_HTM): Define macros.
5242 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
5243 (FIXED_REGISTERS): Likewise.
5244 (CALL_USED_REGISTERS): Likewise.
5245 (CALL_REALLY_USED_REGISTERS): Likewise.
5246 (REG_ALLOC_ORDER): Likewise.
5247 (enum reg_class): Likewise.
5248 (REG_CLASS_NAMES): Likewise.
5249 (REG_CLASS_CONTENTS): Likewise.
5250 (REGISTER_NAMES): Likewise.
5251 (ADDITIONAL_REGISTER_NAMES): Likewise.
5252 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
5253 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
5254 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
5255 * config/rs6000/htm.md: New file.
5256 * config/rs6000/htmintrin.h: New file.
5257 * config/rs6000/htmxlintrin.h: New file.
5259 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
5261 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5262 Define SYMBOL_TINY_GOT, update comment.
5263 * config/aarch64/aarch64.c
5264 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
5265 (aarch64_expand_mov_immediate): Likewise.
5266 (aarch64_print_operand): Likewise.
5267 (aarch64_classify_symbol): Likewise.
5268 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
5269 (ldr_got_tiny): Define.
5271 2013-07-13 Tobias Grosser <tobias@grosser.es>
5273 PR tree-optimization/54094
5274 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
5275 scheduling dimension for the parallelism check from the polyhedral
5276 information in the AST.
5277 * graphite-dependences.c (carries_deps): Do not assume the schedule is
5280 2013-07-13 Jason Merrill <jason@redhat.com>
5282 * print-tree.c (debug_vec_tree): Use debug_raw.
5283 (debug_raw (vec<tree, va_gc> &)): New.
5284 (debug_raw (vec<tree, va_gc> *)): New.
5285 * tree.h: Declare them.
5287 2013-07-13 Bin Cheng <bin.cheng@arm.com>
5289 * ifcvt.c (ifcvt_after_combine): New static variable.
5290 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
5292 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
5293 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
5294 rest_of_handle_if_after_reload): Pass new argument for if_convert.
5296 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
5298 * config/mips/mips.c (mips_expand_call): Remove empty statement.
5300 2013-07-12 Michael Matz <matz@suse.de>
5303 * convert.c (convert_to_real): Reject non-float inner types.
5305 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
5307 * config/aarch64/aarch64-protos.h
5308 (aarch64_simd_immediate_valid_for_move): Remove.
5309 * config/aarch64/aarch64.c (simd_immediate_info): New member.
5310 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
5312 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
5314 2013-07-11 Steve Ellcey <sellcey@mips.com>
5316 * config/mips/mips.c (mips_conditional_register_usage): Do not
5317 use t[0-7] registers in MIPS16 mode when optimizing for size.
5319 2013-07-11 Sriraman Tallam <tmsriram@google.com>
5321 * config/i386/i386.c (dispatch_function_versions): Fix array
5322 indexing of function_version_info to match actual_versions.
5324 2013-07-11 Teresa Johnson <tejohnson@google.com>
5326 * vec.h (struct va_gc): Move release out-of-line.
5327 (va_gc::release): Call ggc_free on released vec.
5329 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5331 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
5332 Require GOT register as additional operand in UNSPEC.
5333 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
5334 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5335 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
5336 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
5337 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
5338 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
5339 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
5341 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
5344 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
5345 name seen by assembler/linker rather if available.
5347 2013-07-11 Andreas Schwab <schwab@suse.de>
5349 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
5351 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
5353 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
5355 2013-07-10 Joseph Myers <joseph@codesourcery.com>
5357 * doc/tm.texi.in: Move hook documentation to ....
5358 * target.def: ... here.
5360 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
5362 * doc/tm.texi: Regenerate.
5364 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
5367 * doc/invoke.texi: Document -Wconditionally-supported.
5369 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5372 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
5375 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5378 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
5379 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
5380 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
5381 Remove duplicate devices.
5382 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
5383 * config/avr/t-multilib: Regenerate.
5384 * config/avr/avr-tables.opt: Regenerate.
5385 * doc/avr-mmcu.texi: Regenerate.
5387 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
5390 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
5392 2013-07-10 Graham Stott <graham.stott@btinternet.com>
5394 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
5395 the cost of MULT when optimizing for size.
5397 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5399 * config/cr16/cr16-protos.h: Don't include target.h.
5401 2013-07-09 Joseph Myers <joseph@codesourcery.com>
5403 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
5404 adjust register size for TDmode and TFmode for VSX registers.
5406 2013-07-08 Kai Tietz <ktietz@redhat.com>
5409 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
5410 hook_bool_const_tree_true.
5412 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5414 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
5415 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
5416 * config/s390/s390.md: Define FPR*_REGNUM constants.
5417 Fix FPR2_REGNUM constant (18 -> 17).
5418 ("*trunc<BFP:mode><DFP_ALL:mode>2")
5419 ("*trunc<DFP_ALL:mode><BFP:mode>2")
5420 ("trunc<BFP:mode><DFP_ALL:mode>2")
5421 ("trunc<DFP_ALL:mode><BFP:mode>2")
5422 ("*extend<BFP:mode><DFP_ALL:mode>2")
5423 ("*extend<DFP_ALL:mode><BFP:mode>2")
5424 ("extend<BFP:mode><DFP_ALL:mode>2")
5425 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
5428 2013-07-08 Graham Stott <graham.stott@btinternet.com>
5430 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
5432 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5434 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
5435 and cfun_fpr_bit_p to cfun_fpr_save_p.
5436 (s390_frame_area, s390_register_info, s390_frame_info)
5437 (s390_emit_prologue, s390_emit_epilogue)
5438 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
5440 * config/s390/s390.h: Define *_REGNUM macros for floating point
5443 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
5445 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
5447 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
5449 PR rtl-optimization/57786
5450 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
5451 and break out of the loop when it is set to false.
5453 2013-07-08 Jakub Jelinek <jakub@redhat.com>
5456 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
5457 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
5458 (const_int 63)) 0)).
5459 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
5460 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
5461 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
5463 PR rtl-optimization/57829
5464 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
5465 mask bits outside of mode are just sign-extension from mode to HWI.
5467 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5469 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
5470 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
5471 adjust_address instead of change_address to keep info about alignment.
5472 (emit_strmov): Remove.
5473 (emit_memmov): New function.
5474 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
5475 (expand_movmem_epilogue): Likewise and return updated rtx for
5477 (expand_constant_movmem_prologue): Likewise and return updated rtx for
5478 destination and source.
5479 (decide_alignment): Refactor, handle vector_loop.
5480 (ix86_expand_movmem): Likewise.
5481 (ix86_expand_setmem): Likewise.
5482 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
5484 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
5486 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
5487 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
5489 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5491 * config/i386/sse.md (sse_movlhps): Change alternative 3
5492 of operand 2 to "m".
5494 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
5497 * config/i386/sse.md (iptr): New mode attribute.
5498 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
5499 (<sse>_vm<plusminus_insn><mode>3): Ditto.
5500 (<sse>_vmmul<mode>3): Ditto.
5501 (<sse>_vmdiv<mode>3): Ditto.
5502 (sse_vmrcpv4sf2): Ditto.
5503 (<sse>_vmsqrt<mode>2): Ditto.
5504 (sse_vmrsqrtv4sf2): Ditto.
5505 (<sse>_vm<code><mode>3): Ditto.
5506 (avx_vmcmp<mode>3): Ditto.
5507 (<sse>_vmmaskcmp<mode>3): Ditto.
5508 (<sse>_comi): Ditto.
5509 (<sse>_ucomi): Ditto.
5510 (*xop_vmfrcz_<mode>): Ditto.
5511 (*fmai_fmadd_<mode>): Ditto.
5512 (*fmai_fmsub_<mode>): Ditto.
5513 (*fmai_fnmadd_<mode>): Ditto.
5514 (*fmai_fnmsub_<mode>): Ditto.
5515 (*fma4i_vmfmadd_<mode>): Ditto.
5516 (*fma4i_vmfmsub_<mode>): Ditto.
5517 (*fma4i_vmfnmadd_<mode>): Ditto.
5518 (*fma4i_vmfnmsub_<mode>): Ditto.
5519 (*xop_vmfrcz_<mode>): Ditto.
5520 (sse_cvtps2pi): Ditto.
5521 (sse_cvttps2pi): Ditto.
5522 (sse_cvtss2si): Ditto.
5523 (sse_cvtss2si_2): Ditto.
5524 (sse_cvtss2siq_2): Ditto.
5525 (sse_cvttss2si): Ditto.
5526 (sse_cvttss2siq): Ditto.
5527 (sse_cvtsd2si): Ditto.
5528 (sse_cvtsd2si_2): Ditto.
5529 (sse_cvtsd2siq_2): Ditto.
5530 (sse_cvttsd2si): Ditto.
5531 (sse_cvttsd2siq): Ditto.
5532 (sse_cvtsd2ss): Ditto.
5533 (sse_cvtss2sd): Ditto.
5534 (avx2_pbroadcast<mode>): Ditto.
5535 (avx2_pbroadcast<mode>_1): Ditto.
5536 (*avx_vperm_broadcast_v4sf): Ditto.
5538 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
5539 (sse_movlhps): Ditto.
5540 (sse_storehps): Ditto.
5541 (sse_loadhps): Ditto.
5542 (sse_storelps): Ditto.
5543 (sse_loadlps): Ditto.
5544 (*vec_concatv4sf): Ditto.
5545 (*vec_interleave_highv2df): Ditto.
5546 (*vec_interleave_lowv2df): Ditto.
5547 (*vec_extractv2df_1_sse): Ditto.
5548 (*vec_extractv2df_0_sse): Ditto.
5549 (sse2_storelpd): Ditto.
5550 (sse2_loadlpd): Ditto.
5551 (sse2_movsd): Ditto.
5552 (*vec_concatv4si): Ditto.
5553 (vec_concatv2di): Ditto.
5555 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
5556 for Intel asm dialect.
5557 (mmx_punpcklwd): Ditto.
5558 (mmx_punpckldq): Ditto.
5560 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
5561 for intel assembler dialect.
5563 2013-07-06 Jakub Jelinek <jakub@redhat.com>
5566 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
5567 for BUILT_IN_C{LZ,LRSB}*.
5568 * tree.h (CASE_INT_FN): Add FN##IMAX case.
5569 * tree-vrp.c (extract_range_basic): Handle
5570 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
5571 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
5572 fall thru to code calling set_value*.
5573 * builtins.c (expand_builtin): Remove *IMAX cases.
5574 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
5575 if width is bigger than 2*HWI.
5577 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
5579 PR rtl-optimization/55342
5580 * lra-int.h (lra_subreg_reload_pseudos): New.
5581 * lra.c: Add undoing optional reloads to the block diagram.
5582 (lra_subreg_reload_pseudos): New.
5583 (lra_optional_reload_pseudos): Change comments.
5584 (lra): Init and clear lra_subreg_reload_pseudos. Clear
5585 lra_optional_reload_pseudos after undo transformations.
5586 * lra-assigns.c (pseudo_prefix_title): New.
5587 (lra_setup_reg_renumber): Use it.
5588 (spill_for): Ditto. Check subreg reload pseudos too.
5589 (assign_by_spills): Consider subreg reload pseudos too.
5590 * lra-constraints.c (simplify_operand_subreg): Use
5591 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
5592 (curr_insn_transform): Recognize and do optional reloads.
5593 (undo_optional_reloads): New.
5594 (lra_undo_inheritance): Call undo_optional_reloads.
5596 2013-07-05 Thomas Quinot <quinot@adacore.com>
5598 * tree-complex.c (expand_complex_operations_1): Fix typo.
5600 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
5602 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
5603 (tune_params): New member 'const vec_costs'.
5604 * config/aarch64/aarch64.c (generic_vector_cost): New.
5605 (generic_tunings): New member 'generic_vector_cost'.
5606 (aarch64_builtin_vectorization_cost): New.
5607 (aarch64_add_stmt_cost): New.
5608 (TARGET_VECTORIZE_ADD_STMT_COST): New.
5609 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
5611 2013-07-03 Jakub Jelinek <jakub@redhat.com>
5614 * config/i386/predicates.md (vsib_address_operand): Disallow
5615 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
5617 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
5620 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
5621 expand_builtin_setjmp_receiver.
5622 (expand_label): Adjust, call expand_builtin_setjmp_receiver
5623 with NULL for the label parameter.
5624 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
5625 the frame-pointer. Adjust comments.
5626 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
5627 only if LABEL is non-NULL.
5629 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
5631 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
5632 (ARM_ABI_AAPCS64): Ditto.
5634 (ARM_DEFAULT_ABI): Ditto.
5636 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
5638 * config/aarch64/aarch64-builtins.c
5639 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
5640 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
5642 * config/aarch64/aarch64-simd.md
5643 (aarch64_ld1<VALL:mode>): New.
5644 (aarch64_st1<VALL:mode>): Likewise.
5645 * config/aarch64/arm_neon.h
5646 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
5648 2013-07-02 Sriraman Tallam <tmsriram@google.com>
5650 * config/i386/i386.c (gate_insert_vzeroupper): Check if
5652 (ix86_option_override_internal):Turn on all -mavx target flags by
5653 default as they are dependent on AVX anyway.
5655 2013-07-02 Cary Coutant <ccoutant@google.com>
5657 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
5659 (loc_checksum_ordered): Likewise.
5660 (hash_loc_operands): Remove inline keyword.
5662 2013-07-02 Jakub Jelinek <jakub@redhat.com>
5664 PR tree-optimization/57741
5665 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
5666 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
5667 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
5668 Allow REAL_CST step_exprs if flag_associative_math.
5669 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
5671 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5673 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
5675 2013-07-02 Ian Bolton <ian.bolton@arm.com>
5677 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
5679 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5681 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
5683 (iorsi3_insn): Likewise.
5684 (arm_xorsi3): Likewise.
5686 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
5688 * arm.md (attribute "wtype"): Delete. Move attribute values from here
5690 (attribute "type"): ... here, and prefix with "wmmx_".
5691 (attribute "core_cycles"): Update for attribute changes.
5692 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
5693 (tbcstv4hi): Likewise.
5694 (tbcstv2si): Likewise.
5695 (iwmmxt_iordi3): Likewise.
5696 (iwmmxt_xordi3): Likewise.
5697 (iwmmxt_anddi3): Likewise.
5698 (iwmmxt_nanddi3): Likewise.
5699 (iwmmxt_arm_movdi): Likewise.
5700 (iwmmxt_movsi_insn): Likewise.
5701 (mov<mode>_internal): Likewise.
5702 (and<mode>3_iwmmxt): Likewise.
5703 (ior<mode>3_iwmmxt): Likewise.
5704 (xor<mode>3_iwmmxt): Likewise.
5705 (add<mode>3_iwmmxt): Likewise.
5706 (ssaddv8qi3): Likewise.
5707 (ssaddv4hi3): Likewise.
5708 (ssaddv2si3): Likewise.
5709 (usaddv8qi3): Likewise.
5710 (usaddv4hi3): Likewise.
5711 (usaddv2si3): Likewise.
5712 (sub<mode>3_iwmmxt): Likewise.
5713 (sssubv8qi3): Likewise.
5714 (sssubv4hi3): Likewise.
5715 (sssubv2si3): Likewise.
5716 (ussubv8qi3): Likewise.
5717 (ussubv4hi3): Likewise.
5718 (ussubv2si3): Likewise.
5719 (mulv4hi3_iwmmxt): Likewise.
5720 (smulv4hi3_highpart): Likewise.
5721 (umulv4hi3_highpart): Likewise.
5722 (iwmmxt_wmacs): Likewise.
5723 (iwmmxt_wmacsz): Likewise.
5724 (iwmmxt_wmacu): Likewise.
5725 (iwmmxt_wmacuz): Likewise.
5726 (iwmmxt_clrdi): Likewise.
5727 (iwmmxt_clrv8qi): Likewise.
5728 (iwmmxt_clr4hi): Likewise.
5729 (iwmmxt_clr2si): Likewise.
5730 (iwmmxt_uavgrndv8qi3): Likewise.
5731 (iwmmxt_uavgrndv4hi3): Likewise.
5732 (iwmmxt_uavgv8qi3): Likewise.
5733 (iwmmxt_uavgv4hi3): Likewise.
5734 (iwmmxt_tinsrb): Likewise.
5735 (iwmmxt_tinsrh): Likewise.
5736 (iwmmxt_tinsrw): Likewise.
5737 (iwmmxt_textrmub): Likewise.
5738 (iwmmxt_textrmsb): Likewise.
5739 (iwmmxt_textrmuh): Likewise.
5740 (iwmmxt_textrmsh): Likewise.
5741 (iwmmxt_textrmw): Likewise.
5742 (iwmxxt_wshufh): Likewise.
5743 (eqv8qi3): Likewise.
5744 (eqv4hi3): Likewise.
5745 (eqv2si3): Likewise.
5746 (gtuv8qi3): Likewise.
5747 (gtuv4hi3): Likewise.
5748 (gtuv2si3): Likewise.
5749 (gtv8qi3): Likewise.
5750 (gtv4hi3): Likewise.
5751 (gtv2si3): Likewise.
5752 (smax<mode>3_iwmmxt): Likewise.
5753 (umax<mode>3_iwmmxt): Likewise.
5754 (smin<mode>3_iwmmxt): Likewise.
5755 (umin<mode>3_iwmmxt): Likewise.
5756 (iwmmxt_wpackhss): Likewise.
5757 (iwmmxt_wpackwss): Likewise.
5758 (iwmmxt_wpackdss): Likewise.
5759 (iwmmxt_wpackhus): Likewise.
5760 (iwmmxt_wpackwus): Likewise.
5761 (iwmmxt_wpackdus): Likewise.
5762 (iwmmxt_wunpckihb): Likewise.
5763 (iwmmxt_wunpckihh): Likewise.
5764 (iwmmxt_wunpckihw): Likewise.
5765 (iwmmxt_wunpckilb): Likewise.
5766 (iwmmxt_wunpckilh): Likewise.
5767 (iwmmxt_wunpckilw): Likewise.
5768 (iwmmxt_wunpckehub): Likewise.
5769 (iwmmxt_wunpckehuh): Likewise.
5770 (iwmmxt_wunpckehuw): Likewise.
5771 (iwmmxt_wunpckehsb): Likewise.
5772 (iwmmxt_wunpckehsh): Likewise.
5773 (iwmmxt_wunpckehsw): Likewise.
5774 (iwmmxt_wunpckelub): Likewise.
5775 (iwmmxt_wunpckeluh): Likewise.
5776 (iwmmxt_wunpckeluw): Likewise.
5777 (iwmmxt_wunpckelsb): Likewise.
5778 (iwmmxt_wunpckelsh): Likewise.
5779 (iwmmxt_wunpckelsw): Likewise.
5780 (ror<mode>3): Likewise.
5781 (ashr<mode>3_iwmmxt): Likewise.
5782 (lshr<mode>3_iwmmxt): Likewise.
5783 (ashl<mode>3_iwmmxt): Likewise.
5784 (ror<mode>3_di): Likewise.
5785 (ashr<mode>3_di): Likewise.
5786 (lshr<mode>3_di): Likewise.
5787 (ashl<mode>3_di): Likewise.
5788 (iwmmxt_wmadds): Likewise.
5789 (iwmmxt_wmaddu): Likewise.
5790 (iwmmxt_tmia): Likewise.
5791 (iwmmxt_tmiaph): Likewise.
5792 (iwmmxt_tmiabb): Likewise.
5793 (iwmmxt_tmiatb): Likewise.
5794 (iwmmxt_tmiabt): Likewise.
5795 (iwmmxt_tmiatt): Likewise.
5796 (iwmmxt_tmovmskb): Likewise.
5797 (iwmmxt_tmovmskh): Likewise.
5798 (iwmmxt_tmovmskw): Likewise.
5799 (iwmmxt_waccb): Likewise.
5800 (iwmmxt_wacch): Likewise.
5801 (iwmmxt_waccw): Likewise.
5802 (iwmmxt_waligni): Likewise.
5803 (iwmmxt_walignr): Likewise.
5804 (iwmmxt_walignr0): Likewise.
5805 (iwmmxt_walignr1): Likewise.
5806 (iwmmxt_walignr2): Likewise.
5807 (iwmmxt_walignr3): Likewise.
5808 (iwmmxt_wsadb): Likewise.
5809 (iwmmxt_wsadh): Likewise.
5810 (iwmmxt_wsadbz): Likewise.
5811 (iwmmxt_wsadhz): Likewise.
5812 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
5813 (iwmmxt_wabsdiffb): Likewise.
5814 (iwmmxt_wabsdiffh): Likewise.
5815 (iwmmxt_wabsdiffw): Likewise.
5816 (iwmmxt_waddsubhx): Likewise
5817 (iwmmxt_wsubaddhx): Likewise.
5818 (addc<mode>3): Likewise.
5819 (iwmmxt_avg4): Likewise.
5820 (iwmmxt_avg4r): Likewise.
5821 (iwmmxt_wmaddsx): Likewise.
5822 (iwmmxt_wmaddux): Likewise.
5823 (iwmmxt_wmaddsn): Likewise.
5824 (iwmmxt_wmaddun): Likewise.
5825 (iwmmxt_wmulwsm): Likewise.
5826 (iwmmxt_wmulwum): Likewise.
5827 (iwmmxt_wmulsmr): Likewise.
5828 (iwmmxt_wmulumr): Likewise.
5829 (iwmmxt_wmulwsmr): Likewise.
5830 (iwmmxt_wmulwumr): Likewise.
5831 (iwmmxt_wmulwl): Likewise.
5832 (iwmmxt_wqmulm): Likewise.
5833 (iwmmxt_wqmulwm): Likewise.
5834 (iwmmxt_wqmulmr): Likewise.
5835 (iwmmxt_wqmulwmr): Likewise.
5836 (iwmmxt_waddbhusm): Likewise.
5837 (iwmmxt_waddbhusl): Likewise.
5838 (iwmmxt_wqmiabb): Likewise.
5839 (iwmmxt_wqmiabt): Likewise.
5840 (iwmmxt_wqmiatb): Likewise.
5841 (iwmmxt_wqmiatt): Likewise.
5842 (iwmmxt_wqmiabbn): Likewise.
5843 (iwmmxt_wqmiabtn): Likewise.
5844 (iwmmxt_wqmiatbn): Likewise.
5845 (iwmmxt_wqmiattn): Likewise.
5846 (iwmmxt_wmiabb): Likewise.
5847 (iwmmxt_wmiabt): Likewise.
5848 (iwmmxt_wmiatb): Likewise.
5849 (iwmmxt_wmiatt): Likewise.
5850 (iwmmxt_wmiabbn): Likewise.
5851 (iwmmxt_wmiabtn): Likewise.
5852 (iwmmxt_wmiatbn): Likewise.
5853 (iwmmxt_wmiattn): Likewise.
5854 (iwmmxt_wmiawbb): Likewise.
5855 (iwmmxt_wmiawbt): Likewise.
5856 (iwmmxt_wmiawtb): Likewise.
5857 (iwmmxt_wmiawtt): Likewise.
5858 (iwmmxt_wmiawbbn): Likewise.
5859 (iwmmxt_wmiawbtn): Likewise.
5860 (iwmmxt_wmiawtbn): Likewise.
5861 (iwmmxt_wmiawttn): Likewise.
5862 (iwmmxt_wmerge): Likewise.
5863 (iwmmxt_tandc<mode>3): Likewise.
5864 (iwmmxt_torc<mode>3): Likewise.
5865 (iwmmxt_torvsc<mode>3): Likewise.
5866 (iwmmxt_textrc<mode>3): Likewise.
5867 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
5868 (wmmxt_pack): Likewise.
5869 (wmmxt_mult_c1): Likewise.
5870 (wmmxt_mult_c2): Likewise.
5871 (wmmxt_alu_c1): Likewise.
5872 (wmmxt_alu_c2): Likewise.
5873 (wmmxt_alu_c3): Likewise.
5874 (wmmxt_transfer_c1): Likewise.
5875 (wmmxt_transfer_c2): Likewise.
5876 (wmmxt_transfer_c3): Likewise.
5877 (marvell_f_iwmmxt_wstr): Likewise.
5878 (marvell_f_iwmmxt_wldr): Likewise.
5880 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
5882 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
5884 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5887 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5888 * lra-constraints.c (need_for_split_p): Check call used hard regs
5889 living through calls.
5891 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
5892 call used regs for call insn.
5894 2013-06-28 Jakub Jelinek <jakub@redhat.com>
5897 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
5898 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
5900 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
5902 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
5904 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
5906 * lra-constraints.c (need_for_split_p): Check call used hard regs
5907 living through calls.
5909 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
5912 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
5913 to tie with any other modes. Eliminate Altivec vector mode tests,
5914 since these are a subset of ALTIVEC or VSX vector modes. Simplify
5915 code, to return 0 if testing MODE2 for a condition, if we've
5916 already tested MODE1 for the same condition.
5918 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5920 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
5923 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5925 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
5926 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
5928 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
5930 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
5932 (aarch64_symbolic_constant_p): Remove.
5933 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
5934 static. Fix line length and white space.
5935 (aarch64_symbolic_constant_p): Remove.
5936 * config/aarch64/predicates.md (aarch64_valid_symref):
5937 Use aarch64_classify_symbol_expression.
5939 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5941 * config/arm/constraints.md (Ts): New constraint.
5942 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
5944 (compare_scc): Use "Ts" constraint for operand 0.
5945 (ior_scc_scc): Likewise.
5946 (and_scc_scc): Likewise.
5947 (and_scc_scc_nodom): Likewise.
5948 (ior_scc_scc_cmp): Likewise for operand 7.
5949 (and_scc_scc_cmp): Likewise.
5950 * config/arm/thumb2.md (thumb2_movsi_insn):
5951 Add alternatives for 16-bit encodings.
5952 (thumb2_movhi_insn): Likewise.
5953 (thumb2_movsicc_insn): Likewise.
5954 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
5955 (thumb2_negscc): Use "Ts" constraint.
5956 Move mvn instruction outside cond_exec block.
5957 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
5958 for 16-bit encodings.
5960 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5962 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
5964 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
5965 (mulsi3subsi): Likewise.
5966 (mulsidi3adddi): Likewise.
5967 (mulsidi3_v6): Likewise.
5968 (umulsidi3_v6): Likewise.
5969 (umulsidi3adddi_v6): Likewise.
5970 (smulsi3_highpart_v6): Likewise.
5971 (umulsi3_highpart_v6): Likewise.
5972 (mulhisi3tb): Likewise.
5973 (mulhisi3bt): Likewise.
5974 (mulhisi3tt): Likewise.
5975 (maddhisi4): Likewise.
5976 (maddhisi4tb): Likewise.
5977 (maddhisi4tt): Likewise.
5978 (maddhidi4): Likewise.
5979 (maddhidi4tb): Likewise.
5980 (maddhidi4tt): Likewise.
5981 (zeroextractsi_compare0_scratch): Likewise.
5982 (insv_zero): Likewise.
5983 (insv_t2): Likewise.
5984 (anddi_notzesidi_di): Likewise.
5985 (anddi_notsesidi_di): Likewise.
5986 (andsi_notsi_si): Likewise.
5987 (iordi_zesidi_di): Likewise.
5988 (xordi_zesidi_di): Likewise.
5989 (andsi_iorsi3_notsi): Likewise.
5991 (smax_m1): Likewise.
5993 (not_shiftsi): Likewise.
5994 (unaligned_loadsi): Likewise.
5995 (unaligned_loadhis): Likewise.
5996 (unaligned_loadhiu): Likewise.
5997 (unaligned_storesi): Likewise.
5998 (unaligned_storehi): Likewise.
5999 (extv_reg): Likewise.
6000 (extzv_t2): Likewise.
6002 (udivsi3): Likewise.
6003 (arm_zero_extendhisi2addsi): Likewise.
6004 (arm_zero_extendqisi2addsi): Likewise.
6005 (compareqi_eq0): Likewise.
6006 (arm_extendhisi2_v6): Likewise.
6007 (arm_extendqisi2addsi): Likewise.
6008 (arm_movt): Likewise.
6009 (thumb2_ldrd): Likewise.
6010 (thumb2_ldrd_base): Likewise.
6011 (thumb2_ldrd_base_neg): Likewise.
6012 (thumb2_strd): Likewise.
6013 (thumb2_strd_base): Likewise.
6014 (thumb2_strd_base_neg): Likewise.
6015 (arm_negsi2): Add alternative for 16-bit encoding.
6016 (arm_one_cmplsi2): Likewise.
6018 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6020 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
6021 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
6022 (movdfcc): Likewise.
6023 * config/arm/vfp.md (*thumb2_movsf_vfp):
6024 Disable predication for arm_restrict_it.
6025 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
6026 (*thumb2_movdfcc_vfp): Likewise.
6027 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
6028 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
6029 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
6030 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
6031 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
6032 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
6033 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
6034 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
6035 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
6036 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
6037 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
6038 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
6039 Disable predication for arm_restrict_it.
6041 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
6043 * config/i386/bmiintrin.h (_bextr_u32): New.
6044 (_bextr_u64): Ditto.
6046 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
6048 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
6049 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
6050 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
6051 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
6052 * config/mips/n32-elf.h: ...this new file.
6054 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
6057 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
6058 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
6060 2013-06-27 Catherine Moore <clm@codesourcery.com>
6062 * config/mips/mips-tables.opt: Regenerate.
6063 * config/mips/mips-cpus.def: Add m14ke and m14kec.
6064 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
6065 * doc/invoke.texi: Add -m14kc.
6067 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6070 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
6071 constraints of operand 1 and 2.
6074 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
6075 to match RTL canonicalization. Swap predicates and
6076 constraints of operand 1 and 2.
6078 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
6080 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
6081 Process OP_INOUT regs for splitting too.
6083 2013-06-27 Jakub Jelinek <jakub@redhat.com>
6085 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
6086 decl before the loop, initialize to NULL.
6087 (vectorizable_load): Initialize ptr_incr to NULL.
6089 2013-06-27 Martin Jambor <mjambor@suse.cz>
6092 * ipa-ref.h (ipa_maybe_record_reference): Declare.
6093 * ipa-ref.c (ipa_maybe_record_reference): New function.
6094 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
6095 * ipa-cp.c (create_specialized_node): Record potential references from
6097 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
6099 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6101 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
6102 parameter 'mode' of type 'enum machine_mode mode'; change to pass
6103 'mode' to force_reg.
6104 (aarch64_add_offset): Update calls to aarch64_force_temporary.
6105 (aarch64_expand_mov_immediate): Likewise.
6107 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
6109 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
6110 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
6112 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6114 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
6115 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6116 (struct machine_function): Add tbegin_p.
6117 (s390_canonicalize_comparison): Fold CC mode compares to
6118 conditional jump if possible.
6119 (s390_emit_jump): Return the emitted jump.
6120 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
6121 Handle CCRAWmode compares.
6122 (s390_option_override): Default to -mhtm if available.
6123 (s390_reg_clobbered_rtx): Handle floating point regs as well.
6124 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
6125 FPRs instead of df_regs_ever_live_p.
6126 (s390_optimize_nonescaping_tx): New function.
6127 (s390_init_frame_layout): Extend clobbered_regs array to cover
6129 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
6130 (s390_expand_tbegin): New function.
6131 (enum s390_builtin): New enum definition.
6132 (code_for_builtin): New array definition.
6133 (s390_init_builtins): New function.
6134 (s390_expand_builtin): New function.
6135 (TARGET_INIT_BUILTINS): Define.
6136 (TARGET_EXPAND_BUILTIN): Define.
6137 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
6138 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
6139 (s390_alc_comparison): Likewise.
6140 * config/s390/s390-modes.def: Add CCRAWmode.
6141 * config/s390/s390.h (processor_flags): Add PF_TX.
6142 (TARGET_CPU_HTM): Define macro.
6143 (TARGET_HTM): Define macro.
6144 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
6145 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
6146 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
6147 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
6148 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
6149 (TBEGIN_MASK, TBEGINC_MASK): New constants.
6150 ("*cc_to_int"): Move up.
6151 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
6152 constants other than 0.
6153 ("*ccraw_to_int"): New insn and splitter definition.
6154 ("tbegin", "tbegin_nofloat", "tbegin_retry")
6155 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
6156 ("tx_assist"): New expander.
6157 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
6158 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
6159 * config/s390/s390.opt: Add -mhtm option.
6160 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
6161 * config/s390/htmxlintrin.h: New file.
6162 * config/s390/htmintrin.h: New file.
6163 * config/s390/s390intrin.h: New file.
6164 * doc/extend.texi: Document htm builtins.
6165 * config.gcc: Add the new header files to extra_headers.
6167 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
6169 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
6170 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
6172 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
6173 Pat Haugen <pthaugen@us.ibm.com>
6174 Peter Bergner <bergner@vnet.ibm.com>
6176 * config/rs6000/power8.md: New.
6177 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
6178 setting for power8 entry.
6179 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
6180 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
6181 test for Power4/Power5 only.
6182 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
6184 (force_new_group): Adjust comment.
6185 * config/rs6000/rs6000.md: Include power8.md.
6187 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
6189 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
6190 * config/arm/arm-protos.h (arm_max_conditional_execute): New
6192 (tune_params): Update comment.
6193 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
6194 (arm_max_conditional_execute): New function.
6195 (thumb2_final_prescan_insn): Use max_insn_skipped and
6196 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
6198 2013-06-25 Jakub Jelinek <jakub@redhat.com>
6200 PR tree-optimization/57705
6201 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
6202 SSA_NAME step, provided that it is not defined inside the loop.
6203 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
6204 (get_initial_def_for_induction): Handle SSA_NAME IV step.
6206 2013-06-25 Martin Jambor <mjambor@suse.cz>
6209 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
6210 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
6212 (ipa_note_param_call): Initialize member_ptr flag.
6213 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
6214 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
6215 (ipa_write_indirect_edge_info): Stream member_ptr flag.
6216 (ipa_read_indirect_edge_info): Likewise.
6218 2013-06-25 Richard Biener <rguenther@suse.de>
6221 * passes.c (init_optimization_passes): Move pass_fold_builtins
6222 and pass_dce earlier with -Og.
6224 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
6226 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
6227 <BIT_FIELD_REF>: Remove trailing TAB.
6228 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
6231 2013-06-24 Martin Jambor <mjambor@suse.cz>
6233 PR tree-optimization/57358
6234 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
6235 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
6236 (ipa_analyze_params_uses): Generate pessimistic info when true.
6238 2013-06-24 Martin Jambor <mjambor@suse.cz>
6240 PR tree-optimization/57539
6241 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
6242 global.inlined_to of the new node to it. All callers changed.
6243 * ipa-inline-transform.c (clone_inlined_nodes): New variable
6244 inlining_into, pass it to cgraph_clone_node.
6245 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
6246 ipa_free_edge_args_substructures.
6247 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
6248 rdesc linked list. Do not assert rdesc edges have inlined caller.
6249 Assert we have found an rdesc in the rdesc list.
6251 2013-06-24 Richard Biener <rguenther@suse.de>
6253 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
6254 (pointer_set_lookup): Declare.
6255 (class pointer_map): New template class implementing a
6256 generic pointer to T map.
6257 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
6258 pointer_map<T>::contains, pointer_map<T>::insert,
6259 pointer_map<T>::traverse): New functions.
6260 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
6261 (pointer_set_lookup): New function.
6262 (pointer_set_contains): Use pointer_set_lookup.
6263 (pointer_set_insert): Likewise.
6264 (insert_aux): Remove.
6265 (struct pointer_map_t): Embed a pointer_set_t.
6266 (pointer_map_create): Adjust.
6267 (pointer_map_destroy): Likewise.
6268 (pointer_map_contains): Likewise.
6269 (pointer_map_insert): Likewise.
6270 (pointer_map_traverse): Likewise.
6271 * tree-streamer.h (struct streamer_tree_cache_d): Use a
6272 pointer_map<unsigned> instead of a pointer_map_t.
6273 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
6274 (streamer_tree_cache_lookup): Likewise.
6275 (streamer_tree_cache_create): Likewise.
6276 (streamer_tree_cache_delete): Likewise.
6277 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
6278 pointer_map<unsigned> instead of a pointer_map_t.
6279 (lto_init_tree_ref_encoder): Adjust.
6280 (lto_destroy_tree_ref_encoder): Likewise.
6281 * lto-section-out.c (lto_output_decl_index): Likewise.
6282 (lto_record_function_out_decl_state): Likewise.
6283 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
6285 2013-06-24 Richard Biener <rguenther@suse.de>
6287 PR tree-optimization/57488
6288 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
6290 2013-06-24 Alan Modra <amodra@gmail.com>
6292 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
6293 (gen_easy_altivec_constant): Likewise.
6294 * config/rs6000/predicates.md (easy_vector_constant_add_self,
6295 easy_vector_constant_msb): Likewise.
6297 2013-06-23 Jakub Jelinek <jakub@redhat.com>
6300 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
6301 add missing return true.
6303 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
6306 * config/sh/predicates.md (general_extend_operand): Invoke
6307 general_movsrc_operand for memory operands.
6308 (general_movsrc_operand): Allow reg+reg addressing, do not use
6309 general_operand for memory operands.
6311 2013-06-23 Sriraman Tallam <tmsriram@google.com>
6313 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
6314 when current target options does not apply.
6315 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
6316 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
6317 * config/i386/bmiintrin.h: Pass appropriate target
6318 attributes to header.
6319 * config/i386/mmintrin.h: Ditto.
6320 * config/i386/nmmintrin.h: Ditto.
6321 * config/i386/avx2intrin.h: Ditto.
6322 * config/i386/fxsrintrin.h: Ditto.
6323 * config/i386/tbmintrin.h: Ditto.
6324 * config/i386/xsaveintrin.h: Ditto.
6325 * config/i386/f16cintrin.h: Ditto.
6326 * config/i386/xtestintrin.h: Ditto.
6327 * config/i386/xsaveoptintrin.h: Ditto.
6328 * config/i386/bmi2intrin.h: Ditto.
6329 * config/i386/lzcntintrin.h: Ditto.
6330 * config/i386/smmintrin.h: Ditto.
6331 * config/i386/wmmintrin.h: Ditto.
6332 * config/i386/x86intrin.h: Remove all header include guards.
6333 * config/i386/prfchwintrin.h: Ditto.
6334 * config/i386/pmmintrin.h: Ditto.
6335 * config/i386/tmmintrin.h: Ditto.
6336 * config/i386/xmmintrin.h: Ditto.
6337 * config/i386/popcntintrin.h: Ditto.
6338 * config/i386/rdseedintrin.h: Ditto.
6339 * config/i386/ammintrin.h: Ditto.
6340 * config/i386/emmintrin.h: Ditto.
6341 * config/i386/immintrin.h: Remove all header include guards.
6342 * config/i386/fma4intrin.h: Ditto.
6343 * config/i386/lwpintrin.h: Ditto.
6344 * config/i386/xopintrin.h: Ditto.
6345 * config/i386/ia32intrin.h: Ditto.
6346 * config/i386/avxintrin.h: Ditto.
6347 * config/i386/rtmintrin.h: Ditto.
6348 * config/i386/fmaintrin.h: Ditto.
6349 * config/i386/mm3dnow.h: Ditto.
6351 2013-06-22 Sriraman Tallam <tmsriram@google.com>
6353 * common/config/i386/i386-common.c: Handle LZCNT.
6355 2013-06-22 Andi Kleen <ak@linux.intel.com>
6357 * doc/extend.texi: Use __atomic_store_n instead of
6358 __atomic_store in HLE example.
6360 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
6362 * config/sh/sh.c: Remove <cstdlib> workaround.
6364 2013-06-21 Andi Kleen <ak@linux.intel.com>
6366 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
6368 2013-06-21 Andi Kleen <ak@linux.intel.com>
6370 * doc/extend.texi: Document that __atomic_clear and
6371 __atomic_test_and_set should only be used with bool.
6373 2013-06-20 Jan Hubicka <jh@suse.cz>
6375 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
6377 * tree.c (decls_same_for_odr): New function.
6378 (same_for_edr): New function.
6379 (types_same_for_odr): New function.
6380 (get_binfo_at_offset): Use it.
6381 * tree.h (types_same_for_odr): Declare.
6383 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
6384 Jason Merrill <jason@redhat.com>
6386 * system.h: Include <cstdlib> as well as <stdlib.h>.
6388 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
6391 * config/i386/i386.c (construct_container): Report error if
6392 long double is used with disabled x87 float returns.
6394 2013-06-20 Jan Hubicka <jh@suse.cz>
6396 * lto-cgraph.c (input_symtab): Do not set cgraph state.
6398 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
6400 PR rtl-optimization/57425
6401 PR rtl-optimization/57569
6402 * alias.c (write_dependence_p): Remove parameters mem_mode and
6403 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
6404 Changed all callers.
6405 (canon_anti_dependence): Get comments and semantics in sync.
6406 Add parameter mem_canonicalized. Changed all callers.
6407 * rtl.h (canon_anti_dependence): Update prototype.
6409 2013-06-20 Richard Biener <rguenther@suse.de>
6411 * data-streamer-in.c (streamer_read_uhwi): Optimize single
6412 byte case, inline streamer_read_uchar and defer section
6415 2013-06-20 Richard Biener <rguenther@suse.de>
6417 PR tree-optimization/57584
6418 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
6419 SSA names into the expanded expression that take part in
6420 abnormal coalescing.
6422 2013-06-19 Sharad Singhai <singhai@google.com>
6424 * gcov.c (print_usage): Handle new option.
6425 (process_args): Ditto.
6426 (get_gcov_intermediate_filename): New function.
6427 (output_intermediate_file): New function.
6428 (output_gcov_file): New function
6429 (generate_results): Handle new option.
6430 (release_function): Relase demangled name.
6431 (read_graph_file): Handle demangled name.
6432 (output_lines): Ditto.
6433 * doc/gcov.texi: Document gcov intermediate format.
6435 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
6438 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
6439 (lra_emit_add): Use the functions. Add comment about Y as an
6442 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
6445 * collect2.c (collect_atexit): New.
6446 (collect_exit): Delete.
6447 (main): Register collect_atexit with atexit.
6448 (collect_wait): Change collect_exit to exit.
6450 * collect2.h (collect_exit): Delete.
6451 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
6453 2013-06-19 Wei Mi <wmi@google.com>
6455 PR rtl-optimization/57518
6456 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
6457 if regno is used in paradoxical subreg.
6458 (update_equiv_regs): Check pdx_subregs[regno] before
6459 set a reg to be equivalent with a mem.
6461 2013-06-19 Matthias Klose <doko@ubuntu.com>
6464 * file-find.h (find_a_file): Add a mode parameter.
6465 * file-find.c (find_a_file): Likewise.
6466 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
6467 with X_OK for the executables.
6468 * collect2.c (main): Call find_a_file with X_OK.
6470 2013-06-19 Steve Ellcey <sellcey@mips.com>
6473 * config/mips/mips.md (casesi_internal_mips16_<mode>):
6474 Use NEXT_INSN instead of next_real_insn.
6476 2013-06-19 Jan Hubicka <jh@suse.cz>
6478 * cgraph.h (const_value_known_p): Replace by ...
6479 (ctor_for_folding): .. this one.
6480 * cgraphunit.c (process_function_and_variable_attributes): Use it.
6481 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
6482 * expr.c (expand_expr_real_1): Likewise.
6483 (string_constant): Likewise.
6484 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
6485 * ipa.c (process_references): Likewise.
6486 (symtab_remove_unreachable_nodes): Likewise.
6487 * ipa-inline-analysis.c (param_change_prob): Likewise.
6488 * gimple-fold.c (canonicalize_constructor_val): Likewise.
6489 (get_base_constructor): Likwise.
6490 * varpool.c (varpool_remove_node): Likewise.
6491 (varpool_remove_initializer): LIkewise.
6492 (dump_varpool_node): LIkwise.
6493 (const_value_known_p): Rewrite to ...
6494 (ctor_for_folding): ... this one.
6496 2013-06-19 Jakub Jelinek <jakub@redhat.com>
6499 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
6500 PERSONALITY in $PATH derived prefixes.
6502 2013-06-19 Jeff Law <law@redhat.com>
6504 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
6507 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
6508 (simplify_bitwise_binary): Use it to simpify certain binary ops on
6511 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
6513 * config/arm/vfp.md: Move VFP instruction classification documentation
6515 * config/arm/arm.md: ... here. Update instruction classification
6518 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
6520 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
6521 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
6522 pattern. Use more efficient sequences on ARMv5 and Thumb2.
6524 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
6527 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6528 with NEXT_INSN. Use tablejump_p to check for jump table data
6531 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
6534 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
6535 that now in C++ the value is correct per the C++ standards.
6537 2013-06-19 Richard Biener <rguenther@suse.de>
6539 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
6542 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6545 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6548 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6549 with next_active_insn.
6551 2013-06-18 Sriraman Tallam <tmsriram@google.com>
6553 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
6554 functions are inlined during failures to flag an error.
6555 * tree-inline.c (expand_call_inline): Allow the error to be flagged
6556 in early inline pass.
6558 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
6560 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
6563 2013-06-18 Julian Brown <julian@codesourcery.com>
6565 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
6566 Permit virtual register pre-reload if !strict.
6567 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
6569 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
6571 * config/arm/neon.md (movmisalign<mode>): Use
6572 neon_perm_struct_or_reg_operand instead of
6573 neon_struct_or_register_operand.
6574 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
6575 neon_permissive_struct_operand instead of neon_struct_operand.
6576 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
6577 neon_vector_mem_operand.
6578 * config/arm/predicates.md (neon_struct_operand): Adjust call to
6579 neon_vector_mem_operand.
6580 (neon_permissive_struct_operand): New.
6581 (neon_struct_or_register_operand): Rename to...
6582 (neon_perm_struct_or_reg_operand): This. Adjust call to
6583 neon_vector_mem_operand.
6585 2013-06-18 Richard Biener <rguenther@suse.de>
6587 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
6588 * lto-streamer.h: Include pointer-set.h.
6589 (struct lto_decl_slot): Remove.
6590 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
6591 Remove next_index entry.
6592 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6593 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6594 (lto_init_tree_ref_encoder): Adjust.
6595 (lto_destroy_tree_ref_encoder): Likewise.
6596 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
6597 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
6598 (lto_output_decl_index): Adjust.
6599 (lto_new_out_decl_state): Likewise.
6600 (lto_record_function_out_decl_state): Likewise.
6601 * lto-streamer-out.c (copy_function): Likewise.
6603 2013-06-18 Richard Biener <rguenther@suse.de>
6605 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
6606 * cgraphunit.c: Include cfgloop.h.
6607 (init_lowered_empty_function): Initialize the loop tree.
6608 (assemble_thunk): Insert new BBs into loops.
6610 2013-06-18 Richard Biener <rguenther@suse.de>
6612 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
6613 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
6614 the map from cache entry to cache index optional.
6615 (streamer_tree_cache_replace_tree): Adjust accordingly.
6616 (streamer_tree_cache_append): Likewise.
6617 (streamer_tree_cache_delete): Likewise.
6618 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
6619 streamer cache map from cache entry to cache index.
6620 * lto-streamer-out.c (create_output_block): Adjust.
6622 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
6624 * config/arm/arm.md (attribute "insn"): Move multiplication and
6625 division attributes to...
6626 (attribute "type"): ... here. Remove mult.
6627 (attribute "mul32"): New attribute.
6628 (attribute "mul64"): Add umaal.
6629 (*arm_mulsi3): Update attributes.
6630 (*arm_mulsi3_v6): Likewise.
6631 (*thumb_mulsi3): Likewise.
6632 (*thumb_mulsi3_v6): Likewise.
6633 (*mulsi3_compare0): Likewise.
6634 (*mulsi3_compare0_v6): Likewise.
6635 (*mulsi_compare0_scratch): Likewise.
6636 (*mulsi_compare0_scratch_v6): Likewise.
6637 (*mulsi3addsi): Likewise.
6638 (*mulsi3addsi_v6): Likewise.
6639 (*mulsi3addsi_compare0): Likewise.
6640 (*mulsi3addsi_compare0_v6): Likewise.
6641 (*mulsi3addsi_compare0_scratch): Likewise.
6642 (*mulsi3addsi_compare0_scratch_v6): Likewise.
6643 (*mulsi3subsi): Likewise.
6644 (*mulsidi3adddi): Likewise.
6645 (*mulsi3addsi_v6): Likewise.
6646 (*mulsidi3adddi_v6): Likewise.
6647 (*mulsidi3_nov6): Likewise.
6648 (*mulsidi3_v6): Likewise.
6649 (*umulsidi3_nov6): Likewise.
6650 (*umulsidi3_v6): Likewise.
6651 (*umulsidi3adddi): Likewise.
6652 (*umulsidi3adddi_v6): Likewise.
6653 (*smulsi3_highpart_nov6): Likewise.
6654 (*smulsi3_highpart_v6): Likewise.
6655 (*umulsi3_highpart_nov6): Likewise.
6656 (*umulsi3_highpart_v6): Likewise.
6657 (mulhisi3): Likewise.
6658 (*mulhisi3tb): Likewise.
6659 (*mulhisi3bt): Likewise.
6660 (*mulhisi3tt): Likewise.
6661 (maddhisi4): Likewise.
6662 (*maddhisi4tb): Likewise.
6663 (*maddhisi4tt): Likewise.
6664 (maddhidi4): Likewise.
6665 (*maddhidi4tb): Likewise.
6666 (*maddhidi4tt): Likewise.
6668 (udivsi3): Likewise.
6669 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
6670 (thumb2_mulsi_short_compare0): Likewise.
6671 (thumb2_mulsi_short_compare0_scratch): Likewise.
6672 * config/arm/arm1020e.md (1020mult1): Update attribute change.
6673 (1020mult2): Likewise.
6674 (1020mult3): Likewise.
6675 (1020mult4): Likewise.
6676 (1020mult5): Likewise.
6677 (1020mult6): Likewise.
6678 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
6680 (cortex_a15_mult64): Likewise.
6681 (cortex_a15_sdiv): Likewise.
6682 (cortex_a15_udiv): Likewise.
6683 * config/arm/arm1026ejs.md (mult1): Update attribute change.
6689 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
6690 (pj4_ir_div): Likewise.
6691 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
6692 (11_mult2): Likewise.
6693 (11_mult3): Likewise.
6694 (11_mult4): Likewise.
6695 (11_mult5): Likewise.
6696 (11_mult6): Likewise.
6697 (11_mult7): Likewise.
6698 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
6699 (cortex_a8_mla): Likewise.
6700 (cortex_a8_mull): Likewise.
6701 (cortex_a8_smulwy): Likewise.
6702 (cortex_a8_smlald): Likewise.
6703 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
6704 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
6705 (cortex_r4_mul_3): Likewise.
6706 (cortex_r4_mla_4): Likewise.
6707 (cortex_r4_mla_3): Likewise.
6708 (cortex_r4_smlald): Likewise.
6709 (cortex_r4_mull): Likewise.
6710 (cortex_r4_sdiv): Likewise.
6711 (cortex_r4_udiv): Likewise.
6712 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
6713 (cortex_a7_idiv): Likewise.
6714 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
6715 (9_mult2): Likewise.
6716 (9_mult3): Likewise.
6717 (9_mult4): Likewise.
6718 (9_mult5): Likewise.
6719 (9_mult6): Likewise.
6720 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
6721 (cortex_a53_sdiv): Likewise.
6722 (cortex_a53_udiv): Likewise.
6723 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
6724 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
6725 (mp626_mult2): Likewise.
6726 (mp626_mult3): Likewise.
6727 (mp626_mult4): Likewise.
6728 * config/arm/fa526.md (526_mult1): Update attribute change.
6729 (526_mult2): Likewise.
6730 * config/arm/arm-generic.md (mult): Update attribute change.
6731 (mult_ldsched_strongarm): Likewise.
6732 (mult_ldsched): Likewise.
6733 (multi_cycle): Likewise.
6734 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
6735 * config/arm/fa606te.md (606te_mult1): Update attribute change.
6736 (606te_mult2): Likewise.
6737 (606te_mult3): Likewise.
6738 (606te_mult4): Likewise.
6739 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
6740 (cortex_a9_mac16): Likewise.
6741 (cortex_a9_multiply): Likewise.
6742 (cortex_a9_mac): Likewise.
6743 (cortex_a9_multiply_long): Likewise.
6744 * config/arm/fa626te.md (626te_mult1): Update attribute change.
6745 (626te_mult2): Likewise.
6746 (626te_mult3): Likewise.
6747 (626te_mult4): Likewise.
6749 2013-06-18 Richard Biener <rguenther@suse.de>
6752 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
6754 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6757 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
6758 with next_active_insn.
6760 2013-06-18 Alan Modra <amodra@gmail.com>
6762 * config/rs6000/rs6000.h (enum data_align): New.
6763 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
6764 (DATA_ABI_ALIGNMENT): Define.
6765 (CONSTANT_ALIGNMENT): Correct comment.
6766 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
6767 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
6769 2013-06-17 David Malcolm <dmalcolm@redhat.com>
6771 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
6772 ATTRIBUTE_UNUSED marking.
6774 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
6776 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
6777 alternative and update.
6778 (aarch64_dup_lanedi): Delete.
6779 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
6780 * config/aarch64/aarch64-simd-builtins.def: Update.
6782 2013-06-17 Richard Biener <rguenther@suse.de>
6784 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
6785 (lto_input_scc): Declare.
6786 (lto_input_tree_1): Likewise.
6787 (struct lto_stats_d): Add num_tree_bodies_output and
6788 num_pickle_refs_output.
6789 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
6790 (lto_read_tree_1): Split out from ...
6791 (lto_read_tree): ... this.
6792 (lto_input_scc): New function.
6793 (lto_input_tree_1): Split out from ...
6794 (lto_input_tree): ... this. Handle LTO_tree_scc.
6795 (lto_data_in_create): Create the streamer cache without hashes.
6796 * lto-streamer-out.c (create_output_block): Create the streamer
6797 cache with hashes when not doing WPA.
6798 (lto_write_tree_1): Split out from ...
6799 (lto_write_tree): ... this.
6800 (get_symbol_initial_value): New function.
6801 (lto_output_tree_1): Split out from ...
6802 (lto_output_tree): ... this. Write trees as series of SCCs
6803 using a DFS walk via DFS_write_tree.
6804 (struct sccs, struct scc_entry): New types.
6805 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
6806 (DFS_write_tree_body): New function.
6807 (DFS_write_tree): Likewise.
6808 (hash_tree): Likewise.
6809 (scc_entry_compare): Likewise.
6810 (hash_scc): Likewise.
6811 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
6812 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
6813 TREE_CHAIN as regular reference.
6814 (streamer_read_integer_cst): Remove.
6815 (streamer_get_pickled_tree): Adjust.
6816 * tree-streamer-out.c (streamer_write_chain): Disable streaming
6817 of DECL_EXTERNALs in BLOCK_VARS for now.
6818 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
6820 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
6821 Add hash value argument and record that if hashes are recorded
6823 (streamer_tree_cache_insert_1): Adjust.
6824 (streamer_tree_cache_insert): Likewise.
6825 (streamer_tree_cache_insert_at): Rename to ...
6826 (streamer_tree_cache_replace_tree): ... this and adjust.
6827 (streamer_tree_cache_append): Adjust.
6828 (record_common_node): Likewise.
6829 (streamer_tree_cache_create): Add argument whether to
6830 record hash values together with trees.
6831 (streamer_tree_cache_delete): Adjust.
6832 * tree-streamer.h (struct streamer_tree_cache_d): Add
6834 (streamer_read_integer_cst): Remove.
6835 (streamer_tree_cache_insert): Adjust.
6836 (streamer_tree_cache_append): Likewise.
6837 (streamer_tree_cache_insert_at): Rename to ...
6838 (streamer_tree_cache_replace_tree): ... this and adjust.
6839 (streamer_tree_cache_create): Add argument whether to record hashes.
6840 (streamer_tree_cache_get): Rename to ...
6841 (streamer_tree_cache_get_tree): ... this.
6842 (streamer_tree_cache_get_hash): New function.
6843 * tree.c (cache_integer_cst): New function.
6844 * tree.h (cache_integer_cst): Declare.
6845 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
6846 * lto-symtab.c (lto_varpool_replace_node): Only release
6847 DECL_INITIAL of non-prevailing decls.
6848 * varpool.c (varpool_remove_initializer): Do not release
6849 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
6851 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6853 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
6854 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
6855 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
6856 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
6857 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
6858 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
6859 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
6860 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
6861 instead of TARGET_64BIT.
6862 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
6863 Require ISA_HAS_<D>DIV.
6865 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
6867 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
6868 (mips*-*-linux*): Move default with_llsc setting to where other
6870 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
6872 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
6873 Likewise. Remove default with_tune setting. Move default float
6874 setting to its own block. Handle with_llsc in the same block as above.
6876 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
6878 PR rtl-optimization/57425
6879 PR rtl-optimization/57569
6880 * alias.c (write_dependence_p): Add new parameters mem_mode,
6881 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
6882 Changed all callers.
6883 (canon_anti_dependence): New function.
6884 * cse.c (check_dependence): Use canon_anti_dependence.
6885 * cselib.c (cselib_invalidate_mem): Likewise.
6886 * rtl.h (canon_anti_dependence): Declare.
6888 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
6890 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
6891 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
6892 ".set mips3" for 64-bit targets.
6894 2013-06-15 Dehao Chen <dehao@google.com>
6896 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
6897 * gimple-low.c (gimple_check_call_matching_types): Likewise.
6898 (gimple_check_call_args): Likewise.
6899 * value-prof.c (check_ic_target): Likewise.
6900 * ipa-inline.c (early_inliner): Likewise.
6901 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
6902 * cgraph.c (cgraph_create_edge_1): Likewise.
6903 (cgraph_make_edge_direct): Likewise.
6905 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
6908 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
6909 rs6000_output_move_128bit to handle emitting quad memory
6910 operations. Set attribute length to 8 bytes.
6912 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
6914 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
6916 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
6917 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
6918 (aarch64_<su>mlsl<mode>): Likewise.
6920 2013-06-14 Mike Stump <mikestump@comcast.net>
6922 * Makefile.in (TARGET_H): Add insn-codes.h.
6924 2013-06-14 Alan Modra <amodra@gmail.com>
6928 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
6929 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
6930 bitfield expansion when EXPAND_MEMORY.
6931 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
6933 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
6935 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
6936 test for clearing quad memory on 32-bit later.
6938 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6940 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
6941 (fold_negate_expr): Likewise.
6942 (fold_real_zero_addition_p): Handle vectors.
6943 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
6945 2013-06-14 Alan Modra <amodra@gmail.com>
6947 * varasm.c (force_const_mem): Revert 2013-06-07 change.
6949 2013-06-13 Jan Hubicka <jh@suse.cz>
6951 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
6952 Local comdats are not externally visible.
6953 * symtab.c (dump_symtab_base): Dump externally visible.
6954 (verify_symtab_base): Verify back links in the symtab hash.
6956 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6958 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
6959 CONVERT_EXPR as equal nodes.
6961 2013-06-13 Bin Cheng <bin.cheng@arm.com>
6963 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
6965 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
6967 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
6968 Generalize to complex and vector.
6969 * tree.c (build_all_ones_cst): New function.
6970 * tree.h (build_all_ones_cst): Declare it.
6972 2013-06-13 Alan Modra <amodra@gmail.com>
6974 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
6975 * config/rs6000/rs6000.md (signbittf2): New insn.
6976 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
6977 (abstf2_internal, cmptf_internal2): Likewise.
6978 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
6980 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
6981 Pat Haugen <pthaugen@us.ibm.com>
6982 Peter Bergner <bergner@vnet.ibm.com>
6984 * config/rs6000/rs6000.c (emit_load_locked): Add support for
6985 power8 byte, half-word, and quad-word atomic instructions.
6986 (emit_store_conditional): Likewise.
6987 (rs6000_expand_atomic_compare_and_swap): Likewise.
6988 (rs6000_expand_atomic_op): Likewise.
6990 * config/rs6000/sync.md (larx): Add new modes for power8.
6992 (AINT): New mode iterator to include TImode as well as normal
6993 integer modes on power8.
6994 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
6995 that VSX registers are not considered. Use AINT mode iterator
6996 instead of INT1 to allow inclusion of quad word atomic operations
6998 (load_locked<mode>): Likewise.
6999 (store_conditional<mode>): Likewise.
7000 (atomic_compare_and_swap<mode>): Likewise.
7001 (atomic_exchange<mode>): Likewise.
7002 (atomic_nand<mode>): Likewise.
7003 (atomic_fetch_<fetchop_name><mode>): Likewise.
7004 (atomic_nand_fetch<mode>): Likewise.
7005 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
7007 (ATOMIC): On power8, add QImode, HImode modes.
7008 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
7009 modes that promote to SImode.
7010 (load_lockedti): Convert TImode arguments to PTImode, so that we
7011 get a guaranteed even/odd register pair.
7012 (load_lockedpti): Likewise.
7013 (store_conditionalti): Likewise.
7014 (store_conditionalpti): Likewise.
7016 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
7017 atomic load/store instructions.
7020 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
7022 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
7024 (insn_count): New attribute, with most cases extracted from...
7025 (length): ...here. Redefine most cases in terms of insn_count.
7026 (single_insn): Delete.
7027 (can_delay): Use insn_count to check for single instructions.
7028 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
7029 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
7030 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
7031 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
7032 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
7033 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
7034 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
7035 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
7036 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
7037 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
7038 rather than "length".
7039 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
7040 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
7041 Use "insn_count" rather than "length".
7042 * config/mips/mips-dsp.md
7043 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
7044 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
7047 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
7049 PR tree-optimization/57361
7050 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
7052 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
7054 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
7056 (aarch64_simd_combine<mode>): New instruction expansion.
7057 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
7059 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
7060 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
7062 2013-06-12 Jan Hubicka <jh@suse.cz>
7064 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
7065 decl has when in streaming stage.
7066 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
7067 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
7069 2013-06-12 Roland Stigge <stigge@antcom.de>
7072 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
7074 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7076 PR tree-optimization/57537
7077 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
7078 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
7080 2013-06-12 Richard Biener <rguenther@suse.de>
7082 * data-streamer.h (streamer_write_char_stream): CSE
7083 obs->current_pointer.
7084 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
7085 streamer_write_char_stream manually and optimize the resulting loop.
7086 (streamer_write_hwi_stream): Likewise.
7088 2013-06-12 Jan Hubicka <jh@suse.cz>
7090 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
7091 * cgraph.h (varpool_create_empty_node): Declare.
7092 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
7094 * symtab.c (symtab_unregister_node): Be lax about missin entries
7096 (symtab_get_node): Update comment.
7097 * varpool.c (varpool_create_empty_node): Break out from ...
7098 (varpool_node_for_decl): ... here.
7099 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
7101 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
7103 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
7104 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
7105 part. Use straight-line flow at the end.
7106 <COMPONENT_REF>: Remove superfluous else.
7107 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
7109 2013-06-12 Jakub Jelinek <jakub@redhat.com>
7112 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
7113 target hook even for !TREE_PUBLIC decls. If no resolution info
7114 is available, return false for common and external decls.
7116 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
7118 * config/rl78/constraints.md (U): New constraint.
7119 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
7122 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
7125 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
7126 to allow returning address to AT_PLATFORM name.
7128 2013-06-11 Jan Hubicka <jh@suse.cz>
7130 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
7131 * cgraph.h (symtab_node_base): Add weakref flag.
7132 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
7133 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
7134 (output_weakrefs): Use weakref flag.
7135 * fold-const.c (simple_operand_p): Handle WEAK.
7136 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
7137 * ipa.c (varpool_externally_visible_p): Drop weakref.
7138 (function_and_variable_visibility): Update comment; fix weakref
7139 sanity checks; do not clear DECL_WEAK on them.
7140 * lto-cgraph.c (lto_output_node): update.
7141 (lto_output_varpool_node): Update.
7142 (input_overwrite_node): Update.
7143 (input_node): Update.
7144 (input_varpool_node): Update.
7145 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
7146 (lto_symtab_merge_symbols): Add sanity check.
7147 (lto_symtab_prevailing_decl): Do not special case weakrefs.
7148 * passes.c (rest_of_decl_compilation): Set static flag, too.
7149 * symtab.c (dump_symtab_base): Dump weakref.
7150 (verify_symtab_base): Sanity check weakrefs.
7151 (symtab_make_decl_local): Remove duplicated code.
7152 (symtab_alias_ultimate_target): Simplify.
7153 * varpool.c (varpool_create_variable_alias): Set weakref flag.
7155 2013-06-11 DJ Delorie <dj@redhat.com>
7157 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
7158 (rl78_unwind_word_mode): New.
7160 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7162 * final.c (debug_prefix_maps): Make static.
7164 2013-06-11 David Malcolm <dmalcolm@redhat.com>
7166 * function.c (initial_trampoline): Remove stray copy.
7168 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
7170 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
7172 2013-06-11 Martin Jambor <mjambor@suse.cz>
7174 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
7175 within bounds at the beginning of the function.
7177 2013-06-11 Alan Modra <amodra@gmail.com>
7179 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
7181 (get_named_section): Don't NULL !DECL_P decl.
7183 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
7185 * doc/invoke.texi (core-avx2): Document.
7187 (atom): Updated with MOVBE.
7189 2013-06-11 Richard Biener <rguenther@suse.de>
7191 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
7193 2013-06-11 Anton Blanchard <anton@samba.org>
7195 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
7196 correct shift value in little-endian mode.
7198 2013-06-11 Jakub Jelinek <jakub@redhat.com>
7201 * varasm.c (get_variable_align): Move #endif to the right place.
7203 2013-06-10 Cary Coutant <ccoutant@google.com>
7205 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
7206 for hash so that hash table traversal order is deterministic.
7208 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
7209 Pat Haugen <pthaugen@us.ibm.com>
7210 Peter Bergner <bergner@vnet.ibm.com>
7212 * config/rs6000/vector.md (GPR move splitter): Do not split moves
7213 of vectors in GPRS if they are direct moves or quad word load or
7216 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
7218 (direct_move_p): Likewise.
7219 (quad_load_store_p): Likewise.
7221 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
7222 classes into bins based on the physical register type.
7223 (reg_class_to_reg_type): Likewise.
7224 (IS_STD_REG_TYPE): Likewise.
7225 (IS_FP_VECT_REG_TYPE): Likewise.
7226 (reload_fpr_gpr): Arrays to determine what insn to use if we can
7227 use direct move instructions.
7228 (reload_gpr_vsx): Likewise.
7229 (reload_vsx_gpr): Likewise.
7230 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
7231 information that is a simplification of register classes. Also
7232 precalculate direct move reload helpers.
7233 (direct_move_p): New function to return true if the operation can
7234 be done as a direct move instruciton.
7235 (quad_load_store_p): New function to return true if the operation
7236 is a quad memory operation.
7237 (rs6000_legitimize_address): If quad memory, only allow register
7238 indirect for TImode addresses.
7239 (rs6000_legitimate_address_p): Likewise.
7240 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
7241 (rs6000_reload_register_type): Likewise.
7242 (register_to_reg_type): Return register type.
7243 (rs6000_secondary_reload_simple_move): New helper function for
7244 secondary reload and secondary memory needed to identify anything
7245 that is a simple move, and does not need reloading.
7246 (rs6000_secondary_reload_direct_move): New helper function for
7247 secondary reload to identify cases that can be done with several
7248 instructions via the direct move instructions.
7249 (rs6000_secondary_reload_move): New helper function for secondary
7250 reload to identify moves between register types that can be done.
7251 (rs6000_secondary_reload): Add support for quad memory operations
7252 and for direct move.
7253 (rs6000_secondary_memory_needed): Likewise.
7254 (rs6000_debug_secondary_memory_needed): Change argument names.
7255 (rs6000_output_move_128bit): New function to return the move to
7256 use for 128-bit moves, including knowing about the various
7257 limitations of quad memory operations.
7259 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
7260 memory operations. call rs6000_output_move_128bit for the actual
7261 instruciton(s) to generate.
7262 (vsx_movti_64bit): Likewise.
7264 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
7265 (UNSPEC_P8V_MTVSRWZ): Likewise.
7266 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
7267 (UNSPEC_P8V_MTVSRD): Likewise.
7268 (UNSPEC_P8V_XXPERMDI): Likewise.
7269 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
7270 (UNSPEC_FUSION_GPR): Likewise.
7271 (FMOVE128_GPR): New iterator for direct move.
7272 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
7275 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
7276 loads and direct move instructions.
7277 (zero_extendsidi2_lfiwzx): Likewise.
7278 (extendsidi2_lfiwax): Likewise.
7279 (extendsidi2_nocell): Likewise.
7280 (floatsi<mode>2_lfiwax): Likewise.
7282 (floatunssi<mode>2_lfiwzx): Likewise.
7284 (fix_trunc<mode>_stfiwx): Likewise.
7285 (fixuns_trunc<mode>_stfiwx): Likewise.
7286 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
7287 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
7288 (parity<mode>2_cmpb): Set length/type attr.
7289 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
7290 for 'mr.' to fast_compare.
7291 (bpermd_<mode>): Change type attr to popcnt.
7292 (p8_fmrgow_<mode>): New insns for power8 direct move support.
7293 (p8_mtvsrwz_1): Likewise.
7294 (p8_mtvsrwz_2): Likewise.
7295 (reload_fpr_from_gpr<mode>): Likewise.
7296 (p8_mtvsrd_1): Likewise.
7297 (p8_mtvsrd_2): Likewise.
7298 (p8_xxpermdi_<mode>): Likewise.
7299 (reload_vsx_from_gpr<mode>): Likewise.
7300 (reload_vsx_from_gprsf): Likewise.
7301 (p8_mfvsrd_3_<mode>): LIkewise.
7302 (reload_gpr_from_vsx<mode>): Likewise.
7303 (reload_gpr_from_vsxsf): Likewise.
7304 (p8_mfvsrd_4_disf): Likewise.
7305 (multi-word GPR splits): Do not split direct moves or quad memory
7308 2013-06-10 David Malcolm <dmalcolm@redhat.com>
7310 * tree-into-ssa.c (interesting_blocks): Make static.
7312 2013-06-10 Jakub Jelinek <jakub@redhat.com>
7315 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
7316 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
7317 Use DATA_ABI_ALIGNMENT for that case instead if defined.
7318 (get_variable_align): New function.
7319 (get_variable_section, emit_bss, emit_common,
7320 assemble_variable_contents, place_block_symbol): Use
7321 get_variable_align instead of DECL_ALIGN.
7322 (assemble_noswitch_variable): Add align argument, use it
7323 instead of DECL_ALIGN.
7324 (assemble_variable): Adjust caller. Use get_variable_align
7325 instead of DECL_ALIGN.
7326 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
7328 (DATA_ABI_ALIGNMENT): Define.
7329 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
7330 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
7331 opt is false, only return the psABI mandated alignment increase.
7332 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
7333 (DATA_ABI_ALIGNMENT): ... this.
7334 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
7335 (DATA_ABI_ALIGNMENT): ... this.
7336 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
7337 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
7338 (DATA_ABI_ALIGNMENT): ... this.
7339 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
7340 * doc/tm.texi: Regenerated.
7342 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
7344 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
7345 cmp_code to construct REG_EQUAL note.
7347 2013-06-09 Jakub Jelinek <jakub@redhat.com>
7350 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
7351 that operands[2] doesn't overlap with operands[0].
7353 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
7354 Jan Hubicka <jh@suse.cz>
7356 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
7357 hack to mark symbols as used.
7359 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
7361 PR rtl-optimization/57559
7362 * lra-constraints.c (process_alt_operands): Don't discourage
7363 memory with known offset for offsetable memory constraint.
7364 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
7366 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
7368 * varasm.c (struct oc_local_state): Reorder fields.
7369 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
7370 and adjust accordingly.
7371 (output_constructor): Reorder initialization code and adjust call to
7372 output_constructor_bitfield.
7374 2013-06-07 Jan Hubicka <jh@suse.cz>
7376 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
7378 2013-06-07 David Malcolm <dmalcolm@redhat.com>
7380 * tree-object-size.c (unknown): Make const.
7382 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7384 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
7385 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
7386 for last alternative in the cpu_facility attribute.
7388 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7391 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
7392 (xordi3): Change operand 2 constraint to arm_xordi_operand.
7393 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
7394 * config/arm/constraints.md (Dg): New constraint.
7395 * config/arm/neon.md (xordi3_neon): Remove.
7396 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
7397 * config/arm/predicates.md (arm_xordi_operand): New predicate.
7399 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7401 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
7402 Clean up alternatives.
7404 2013-06-07 Alan Modra <amodra@gmail.com>
7406 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
7409 2013-06-07 Alan Modra <amodra@gmail.com>
7411 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
7413 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7415 * config/arm/constraints.md (Df): New constraint.
7416 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
7417 Correct length attribute for last two alternatives.
7419 2013-06-07 Alan Modra <amodra@gmail.com>
7421 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7422 override user -mfp-in-toc.
7423 (offsettable_ok_by_alignment): Consider just the current access
7424 rather than the whole object, unless BLKmode. Handle
7425 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
7426 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
7427 for -mcmodel=medium.
7428 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
7429 override user -mfp-in-toc or -msum-in-toc. Default to
7430 -mno-fp-in-toc for -mcmodel=medium.
7432 2013-06-06 DJ Delorie <dj@redhat.com>
7434 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
7435 TARGET_VALID_POINTER_MODE.
7437 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
7438 Pat Haugen <pthaugen@us.ibm.com>
7439 Peter Bergner <bergner@vnet.ibm.com>
7441 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7442 Document new power8 builtins.
7444 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
7445 condition code register, to allow 128-bit logical operations to be
7446 done in the VSX or GPR registers.
7447 (nor<mode>3): Use the canonical form for nor.
7448 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
7449 vclz*, and vpopcnt* vector instructions.
7450 (nand<mode>3): Likewise.
7451 (orc<mode>3): Likewise.
7452 (clz<mode>2): LIkewise.
7453 (popcount<mode>2): Likewise.
7455 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
7456 that only the GPRs are recognized.
7458 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7459 support for new power8 builtins.
7461 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
7463 (xscvdpspn): Likewise.
7469 (vpopcnt): Likewise.
7470 (vpopcntb): Likewise.
7471 (vpopcnth): Likewise.
7472 (vpopcntw): Likewise.
7473 (vpopcntd): Likewise.
7478 (eqv_v16qi3): Likewise.
7479 (eqv_v8hi3): Likewise.
7480 (eqv_v4si3): Likewise.
7481 (eqv_v2di3): Likewise.
7482 (eqv_v4sf3): Likewise.
7483 (eqv_v2df3): Likewise.
7485 (nand_v16qi3): Likewise.
7486 (nand_v8hi3): Likewise.
7487 (nand_v4si3): Likewise.
7488 (nand_v2di3): Likewise.
7489 (nand_v4sf3): Likewise.
7490 (nand_v2df3): Likewise.
7492 (orc_v16qi3): Likewise.
7493 (orc_v8hi3): Likewise.
7494 (orc_v4si3): Likewise.
7495 (orc_v2di3): Likewise.
7496 (orc_v4sf3): Likewise.
7497 (orc_v2df3): Likewise.
7499 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
7500 allow power8 quad mode in 64-bit.
7501 (rs6000_builtin_vectorized_function): Add support to vectorize
7502 ISA 2.07 count leading zeros, population count builtins.
7503 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
7504 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
7505 (builtin_function_type): Add vgbbd builtin function which takes an
7507 (altivec_expand_vec_perm_const): Add support for new power8 merge
7510 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
7511 that does not include TImdoe for use with 32-bit.
7512 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
7514 (UNSPEC_VSX_CVDPSPN): Likewise.
7515 (vsx_xscvdpspn): Likewise.
7516 (vsx_xscvspdpn): Likewise.
7517 (vsx_xscvdpspn_scalar): Likewise.
7518 (vsx_xscvspdpn_directmove): Likewise.
7519 (vsx_and<mode>3): Split logical operations into 32-bit and
7520 64-bit. Add support to do logical operations on TImode as well as
7521 VSX vector types. Allow logical operations to be done in either
7522 VSX registers or in general purpose registers in 64-bit mode. Add
7523 splitters if GPRs were used. For AND, add clobber of CCmode to
7524 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
7526 (vsx_and<mode>3_32bit): Likewise.
7527 (vsx_and<mode>3_64bit): Likewise.
7528 (vsx_ior<mode>3): Likewise.
7529 (vsx_ior<mode>3_32bit): Likewise.
7530 (vsx_ior<mode>3_64bit): Likewise.
7531 (vsx_xor<mode>3): Likewise.
7532 (vsx_xor<mode>3_32bit): Likewise.
7533 (vsx_xor<mode>3_64bit): Likewise.
7534 (vsx_one_cmpl<mode>2): Likewise.
7535 (vsx_one_cmpl<mode>2_32bit): Likewise.
7536 (vsx_one_cmpl<mode>2_64bit): Likewise.
7537 (vsx_nor<mode>3): Likewise.
7538 (vsx_nor<mode>3_32bit): Likewise.
7539 (vsx_nor<mode>3_64bit): Likewise.
7540 (vsx_andc<mode>3): Likewise.
7541 (vsx_andc<mode>3_32bit): Likewise.
7542 (vsx_andc<mode>3_64bit): Likewise.
7543 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
7544 and xxlorc instructions.
7545 (vsx_eqv<mode>3_64bit): Likewise.
7546 (vsx_nand<mode>3_32bit): Likewise.
7547 (vsx_nand<mode>3_64bit): Likewise.
7548 (vsx_orc<mode>3_32bit): Likewise.
7549 (vsx_orc<mode>3_64bit): Likewise.
7551 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
7553 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
7555 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
7556 (p8_vmrgow): Likewise.
7557 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
7558 GPRs to be split under VSX.
7559 (p8v_clz<mode>2): Add power8 count leading zero support.
7560 (p8v_popcount<mode>2): Add power8 population count support.
7561 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
7564 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
7567 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
7569 (vec_nand): Likewise.
7570 (vec_vclz): Likewise.
7571 (vec_vclzb): Likewise.
7572 (vec_vclzd): Likewise.
7573 (vec_vclzh): Likewise.
7574 (vec_vclzw): Likewise.
7575 (vec_vgbbd): Likewise.
7576 (vec_vmrgew): Likewise.
7577 (vec_vmrgow): Likewise.
7578 (vec_vpopcnt): Likewise.
7579 (vec_vpopcntb): Likewise.
7580 (vec_vpopcntd): Likewise.
7581 (vec_vpopcnth): Likewise.
7582 (vec_vpopcntw): Likewise.
7584 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7586 PR rtl-optimization/57468
7587 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
7590 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7592 PR rtl-optimization/57459
7593 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
7594 type when setting live regs.
7596 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
7598 * config/s390/s390.opt (mlra): New option.
7599 * config/s390/s390.c (s390_decompose_address): Check displacement
7600 for all registers for LRA.
7601 (s390_secondary_reload): Don't used secondary reloads for LRA.
7602 (s390_lra_p): New function.
7603 (TARGET_LRA_P): Define.
7604 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
7605 of attribute cpu_facility to zarch for the last alternative.
7606 (*cmpmem_short): Ditto.
7608 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
7610 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
7611 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
7612 (arm_expand_prologue): Likewise.
7614 2013-06-06 Teresa Johnson <tejohnson@google.com>
7617 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
7618 as this is now done by redirect_edge_and_branch_force.
7619 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
7620 barriers, and fix interaction with splitting.
7621 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
7622 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
7623 reflect changes made in the routine.
7624 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
7625 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
7626 since this is called in cfglayout mode, and replace partition fixup
7627 with assert as that is now done by force_nonfallthru_and_redirect.
7628 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
7629 already be marked with region crossing note.
7630 (insert_section_boundary_note): Make non-static, gate on flag
7631 has_bb_partition, rewrite to also check for multiple partitions.
7632 (rest_of_handle_reorder_blocks): Remove call to
7633 insert_section_boundary_note, now done later during free_cfg.
7634 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
7635 * bb-reorder.h (insert_section_boundary_note): Declare.
7636 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
7637 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
7638 invoke insert_section_boundary_note.
7639 (try_redirect_by_replacing_jump): Remove unnecessary
7640 check for region crossing note.
7641 (fixup_partition_crossing): New function.
7642 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
7643 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
7644 in non-cfglayout mode.
7645 (force_nonfallthru_and_redirect): Fixup partition boundaries,
7646 remove old code that tried to do this. Emit barrier correctly
7647 when we are in cfglayout mode.
7648 (last_bb_in_partition): New function.
7649 (rtl_split_edge): Correctly fixup partition boundaries.
7650 (commit_one_edge_insertion): Remove old code that tried to
7651 fixup region crossing edge since this is now handled in
7652 split_block, and set up insertion point correctly since
7653 block may now end in a jump.
7654 (verify_hot_cold_block_grouping): Guard against checking when not in
7655 linearized RTL mode.
7656 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
7658 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
7659 rtl_verify_flow_info, so not called in cfglayout mode.
7660 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
7661 (fixup_reorder_chain): Remove old code that attempted to fixup region
7662 crossing note as this is now handled in force_nonfallthru_and_redirect.
7663 (duplicate_insn_chain): Don't duplicate switch section notes.
7664 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
7666 * basic-block.h (emit_barrier_after_bb): Declare.
7668 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7670 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
7671 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
7672 arm_usatsihi): Adjust alternatives for arm_restrict_it.
7674 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7676 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
7678 * config/arm/ldmstm.md: Regenerate.
7680 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7682 * config/arm/sync.md (atomic_loaddi_1):
7683 Disable predication for arm_restrict_it.
7684 (arm_load_exclusive<mode>): Likewise.
7685 (arm_load_exclusivesi): Likewise.
7686 (arm_load_exclusivedi): Likewise.
7687 (arm_load_acquire_exclusive<mode>): Likewise.
7688 (arm_load_acquire_exclusivesi): Likewise.
7689 (arm_load_acquire_exclusivedi): Likewise.
7690 (arm_store_exclusive<mode>): Likewise.
7691 (arm_store_exclusive<mode>): Likewise.
7692 (arm_store_release_exclusivedi): Likewise.
7693 (arm_store_release_exclusive<mode>): Likewise.
7695 2013-06-06 Richard Biener <rguenther@suse.de>
7697 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
7699 (lto_tag_is_tree_code_p): Adjust.
7700 (lto_tag_is_gimple_code_p): Likewise.
7701 (lto_gimple_code_to_tag): Likewise.
7702 (lto_tag_to_gimple_code): Likewise.
7703 (lto_tree_code_to_tag): Likewise.
7704 (lto_tag_to_tree_code): Likewise.
7705 * data-streamer.h (streamer_write_hwi_in_range): Use
7706 uhwi streaming to stream the normalized range.
7707 (streamer_read_hwi_in_range): Likewise.
7709 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7711 * config/arm/arm.md (enabled_for_depr_it): New attribute.
7712 (predicable_short_it): Likewise.
7713 (predicated): Likewise.
7714 (enabled): Handle above.
7715 (define_cond_exec): Set predicated attribute to yes.
7717 2013-06-05 Mike Stump <mikestump@comcast.net>
7719 * gdbinit.in (__FUNCTION__): Add.
7721 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
7723 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
7724 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
7726 2013-06-05 Jan Hubicka <jh@suse.cz>
7728 * varasm.c (mark_decl_referenced): Revert the removal until targets
7731 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
7733 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
7734 instead of mark_decl_referenced.
7736 2013-06-05 Jan Hubicka <jh@suse.cz>
7738 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
7739 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
7740 and symtab_used_from_object_file_p.
7741 (cgraph_make_node_local_1): Clear forced_by_abi.
7742 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
7743 * cgraph.h (symtab_node_base): Add forced_by_abi.
7744 (decide_is_variable_needed): Remove.
7745 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
7746 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
7747 (decide_is_symbol_needed): ... this one; handle symbols in general;
7748 always analyze virtuals; honnor forced_by_abi.
7749 (cgraph_finalize_function): Update.
7750 (varpool_finalize_decl): Update.
7751 (symbol_defined_and_needed): Remove.
7752 (analyze_functions): Update.
7753 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
7754 output_refs, input_overwrite_node): Handle forced_by_abi.
7755 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
7756 (address_taken_from_non_vtable_p): ... this one.
7757 (comdat_can_be_unshared_p_1): New function.
7758 (cgraph_comdat_can_be_unshared_p): Rename to ...
7759 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
7760 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
7761 (function_and_variable_visibility): Clear forced_by_abi as needed.
7762 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
7763 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
7764 * symtab.c (dump_symtab_base): Dump forced_by_abi.
7765 * varpool.c (decide_is_variable_needed): Remove.
7767 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7769 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
7770 (arm_option_override): Override arm_restrict_it where appropriate.
7771 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
7772 * config/arm/arm.opt (mrestrict-it): New command-line option.
7773 * doc/invoke.texi: Document -mrestrict-it.
7775 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7777 * tsan.c (tsan_atomic_table): Make const.
7779 2013-06-05 Richard Biener <rguenther@suse.de>
7781 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
7782 index associated with the tree we are supposed to replace.
7783 * tree-streamer-out.c (pack_ts_base_value_fields): Output
7784 TREE_ASM_WRITTEN as zero for everything but SSA names.
7786 2013-06-05 David Malcolm <dmalcolm@redhat.com>
7788 * tree-ssa-structalias.c (call_stmt_vars): Make static.
7790 2013-06-04 Jan Hubicka <jh@suse.cz>
7792 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
7793 (input_node, input_varpool_node): Handle correctly external same
7795 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
7796 nodes at ltrans stage.
7798 2013-06-04 Jan Hubicka <jh@suse.cz>
7800 * ipa-inline.c (update_caller_keys): Fix availability test.
7801 (update_callee_keys): Likewise.
7802 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
7803 to follow ELF standard.
7805 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
7807 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
7808 (mips64r5900el-*-elf*): New configurations.
7809 * config/mips/mips-cpus.def (r5900): New processor.
7810 * config/mips/mips-tables.opt: Regenerate.
7811 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
7812 (mips_issue_rate): Handle PROCESSOR_R5900.
7813 (mips_reorg_process_insns): Force reorder mode for the R5900.
7814 * config/mips/mips.h (TARGET_MIPS5900): Define.
7815 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
7817 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
7819 * config/mips/mips.md (processor): Add r5900.
7820 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
7822 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7824 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
7825 into function to generate MOVI instruction.
7826 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
7827 (aarch64_preferred_simd_mode): Turn into wrapper.
7828 (aarch64_output_scalar_simd_mov_immediate): New function.
7829 * config/aarch64/aarch64-protos.h: Add prototype for above.
7831 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7833 * config/aarch64/aarch64.c (simd_immediate_info): Remove
7834 element_char member.
7835 (sizetochar): Return signed char.
7836 (aarch64_simd_valid_immediate): Remove elchar and other
7837 unnecessary variables.
7838 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
7839 Calculate element_char as required.
7840 * config/aarch64/aarch64-protos.h: Update and move prototype
7841 for aarch64_output_simd_mov_immediate.
7842 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
7845 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7847 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
7848 information completed by aarch64_simd_valid_immediate.
7849 (aarch64_legitimate_constant_p): Update arguments.
7850 (aarch64_simd_valid_immediate): Work with struct rather than many
7852 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
7853 (aarch64_simd_make_constant): Update arguments.
7854 (aarch64_output_simd_mov_immediate): Work with struct rather than
7855 many pointers. Output immediate directly rather than as operand.
7856 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7858 * config/aarch64/constraints.md (Dn): Update arguments.
7860 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7862 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
7864 (aarch64_simd_immediate_valid_for_move): Remove.
7865 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
7866 (aarch64_simd_make_constant): Update call.
7867 (aarch64_output_simd_mov_immediate): Update call.
7868 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
7870 * config/aarch64/constraints.md (Dn): Update call.
7872 2013-06-04 Ian Bolton <ian.bolton@arm.com>
7874 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
7875 return type to bool for prototype.
7876 (aarch64_legitimate_constant_p): Check for true instead of not -1.
7877 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
7878 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
7880 2013-06-04 Catherine Moore <clm@codesourcery.com>
7882 * config/mips/mips.opt (meva): New.
7883 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
7884 (ASM_SPEC): Handle -meva.
7885 * doc/invoke.texi (meva): Document.
7887 2013-06-04 Alan Modra <amodra@gmail.com>
7889 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
7892 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7894 * rtl.def: Add extra fourth optional field to define_cond_exec.
7895 * gensupport.c (process_one_cond_exec): Process attributes from
7897 * doc/md.texi: Document fourth field in define_cond_exec.
7899 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
7901 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
7902 out the processing order as in store_bit_field_1.
7904 2013-06-04 Jan Hubicka <jh@suse.cz>
7907 * cgraphunit.c (cgraph_process_same_body_aliases): Create
7908 non-VAR_DECL node if it does not exist yet.
7910 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
7912 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
7913 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
7914 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
7915 target_cpu_default setting.
7917 2013-06-03 Teresa Johnson <tejohnson@google.com>
7919 * dumpfile.c (opt_info_switch_p): Change -fopt-info
7920 default to -fopt-info=optimized instead of all.
7921 * doc/invoke.texi: Ditto.
7922 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
7923 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
7924 (execute_vect_slp): Emit BB vectorization success under
7925 MSG_OPTIMIZED_LOCATIONS.
7926 * tree-vect-slp.c (vect_slp_transform_bb): Change
7927 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
7928 * tree-vect-loop.c (vect_transform_loop): Ditto.
7930 2013-06-03 Jason Merrill <jason@redhat.com>
7933 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7934 Use TARGET_EXPR for C++.
7936 2013-06-03 Jakub Jelinek <jakub@redhat.com>
7938 PR rtl-optimization/57268
7939 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
7940 if DEBUG_INSN_P (insn).
7943 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
7945 PR rtl-optimization/57268
7946 * sched-deps.c (sched_analyze_2): Flush dependence lists if
7947 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
7949 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
7951 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
7952 (ix86_avoid_lea_for_addr): Likewise.
7953 (exact_dependency_1): Likewise.
7954 (ix86_adjust_cost): Likewise.
7955 (swap_top_of_ready_list): Fix formatting and !reload_completed check
7957 (do_reorder_for_imul): Fix typo, formatting and
7958 !reload_completed check removed.
7959 (ix86_sched_reorder): Fix typo and formatting.
7960 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
7963 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
7965 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
7967 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
7969 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
7970 <STRING_CST>: Likewise.
7971 <VECTOR_CST>: Likewise.
7973 2013-06-01 Janus Weil <janus@gcc.gnu.org>
7974 Mikael Morin <mikael@gcc.gnu.org>
7976 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
7977 * config.in: Regenerated.
7978 * configure: Regenerated.
7980 2013-06-01 Jan Hubicka <jh@suse.cz>
7983 * cgraphunit.c (compile): When weakref is not supported,
7984 set up transparent aliases before final output pass.
7985 * varasm.c (assemble_alias): Do not try to do it here.
7987 2013-06-01 Jan Hubicka <jh@suse.cz>
7990 * passes.c (for_per_function): Skip unanalyzed functions.
7992 2013-06-01 Jan Hubicka <jh@suse.cz>
7994 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
7995 (lto_symtab_merge_symbols_1): ... this one.
7996 (lto_symtab_merge_cgraph_nodes): Rename to ...
7997 (lto_symtab_merge_symbols): ... this one; simplify.
7998 * cgraph.c (same_body_aliases_done): Rename to ...
7999 (cpp_implicit_aliases_done): ... this one.
8000 (cgraph_create_function_alias): Update.
8001 (cgraph_same_body_alias): Update.
8002 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
8003 (verify_edge_corresponds_to_fndecl): Simplify.
8004 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
8005 (cgraph_node): Remove same_body_alias.
8006 (varpool_node): Remove alias_of and extra_name_alias.
8007 (same_body_aliases_done): Rename to ..
8008 (cpp_implicit_aliases_done): ... this one.
8009 (symtab_alias_ultimate_target): Add default parameter.
8010 (symtab_resolve_alias): New function.
8011 (fixup_same_cpp_alias_visibility): Declare.
8012 (cgraph_function_node): Add default parameter.
8013 (cgraph_node_asm_name): Likewise.
8014 (cgraph_function_or_thunk_node): Add default parameter; do
8015 not ICE when it is NULL.
8016 (varpool_variable_node): Likewise.
8017 * tree-emutls.c (create_emultls_var): Update.
8018 (ipa_lower_emutls): Update.
8019 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
8020 (cgraph_reset_node): Reset alias info.
8021 (cgraph_finalize_function): Update.
8022 (fixup_same_cpp_alias_visibility): Move to symtab.c.
8023 (analyze_function): Simplify.
8024 (cgraph_process_same_body_aliases): Simplify.
8025 (analyze_functions): Fixup same body aliases.
8026 (handle_alias_pairs): Simplify.
8027 (assemble_thunk): Update.
8028 (assemble_thunks_and_aliases): Update.
8029 (output_weakrefs): Rewrite.
8030 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
8031 (lto_output_varpool_node): Likewise.
8032 (compute_ltrans_boundary): Remve assert.
8033 (get_alias_symbol): New functoin.
8034 (input_node): Rewrite alias handling.
8035 (input_varpool_node): Likewise.
8036 * ipa-pure-const.c (propagate_pure_const): Fix formating.
8037 * ipa.c (process_references): Handle weakrefs correctly.
8038 (symtab_remove_unreachable_nodes): Likewise.
8039 * trans-mem.c (get_cg_data): Update.
8040 (ipa_tm_create_version_alias): Update.
8041 (ipa_tm_execute): Update.
8042 * symtab.c (dump_symtab_base): Dump aliases.
8043 (verify_symtab_base): Verify aliases.
8044 (symtab_node_availability): New function.
8045 (symtab_alias_ultimate_target): Simplify.
8046 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
8047 handle all the fixup cases.
8048 (symtab_resolve_alias): New function.
8049 * passes.c (ipa_write_summaries): Handle weakrefs.
8050 * varpool.c (varpool_analyze_node): Simplify.
8051 (assemble_aliases): Update.
8052 (varpool_create_variable_alias): Simplify.
8053 (varpool_extra_name_alias): Simplify.
8054 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
8055 (lto_symtab_merge_symbols): ... this one.
8057 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
8060 PR rtl-optimization/57268
8061 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8062 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8064 2013-06-01 Tobias Burnus <burnus@net-b.de>
8067 2013-05-31 Tobias Burnus <burnus@net-b.de>
8070 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
8073 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
8075 PR rtl-optimization/57268
8076 * sched-deps.c (sched_analyze_2): Flush dependence lists if
8077 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
8079 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8081 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
8082 unordered comparison operators when -fno-trapping-math is in effect
8084 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
8085 and implement unordered comparison operators properly on the e500.
8087 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
8089 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
8090 for constant scalar integers.
8091 (simplify_relational_operation_1): Likewise.
8093 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
8095 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
8096 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
8099 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8100 Igor Zamyatin <igor.zamyatin@intel.com>
8102 Silvermont (SLM) architecture performance tuning.
8103 * config/i386/i386.h (enum ix86_tune_indices): Add
8104 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
8105 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
8107 * config/i386/i386.c (initial_ix86_tune_features)
8108 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
8109 (ix86_lea_outperforms): Handle Silvermont tuning.
8110 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
8112 (ix86_use_lea_for_mov): Likewise.
8113 (ix86_avoid_lea_for_addr): Likewise.
8114 (ix86_lea_for_add_ok): Likewise.
8115 (exact_dependency_1): New function.
8116 (exact_store_load_dependency): Likewise.
8117 (ix86_adjust_cost): Handle Silvermont tuning.
8118 (do_reoder_for_imul): Likewise.
8119 (swap_top_of_ready_list): New function.
8120 (ix86_sched_reorder): Changed to handle Silvermont tuning.
8122 * config/i386/i386.md (peepholes that split memory operand in fp
8125 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8127 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
8128 Remove un-necessary braces.
8130 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
8132 * config/aarch64/aarch64.c (aarch64_classify_symbol):
8133 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
8135 2013-05-31 Tobias Burnus <burnus@net-b.de>
8138 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
8140 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8143 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
8144 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
8145 * config/arm/neon.md (iordi3_neon): Remove.
8146 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
8147 * config/arm/predicates.md (imm_for_neon_logic_operand):
8148 Move to earlier in the file.
8149 (neon_logic_op2): Likewise.
8150 (arm_iordi_operand_neon): New predicate.
8152 2013-05-31 Richard Biener <rguenther@suse.de>
8154 PR tree-optimization/57478
8155 PR tree-optimization/57453
8156 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
8159 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
8161 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
8162 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
8164 2013-05-30 Tobias Burnus <burnus@net-b.de>
8165 Thomas Koenig <tkoenig@gcc.gnu.org>
8168 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
8169 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
8171 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
8173 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
8175 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
8177 * target.def (register_usage_leveling_p): New hook.
8178 * targhooks.c (default_register_usage_leveling_p): New.
8179 * targhooks.h (default_register_usage_leveling_p): New prototype.
8180 * lra-assigns.c (register_usage_leveling_p): Use the hook.
8181 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
8182 * doc/tm.texi: Update.
8183 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
8185 2013-05-30 Ian Bolton <ian.bolton@arm.com>
8187 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
8188 (*insv_reg<mode>): New define_insn.
8190 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
8192 PR rtl-optimization/57439
8193 * postreload.c (move2add_valid_value_p): Check that we have
8194 a zero subreg_regno_offset when accessing the register in
8197 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
8198 Igor Zamyatin <igor.zamyatin@intel.com>
8200 Silvermont (SLM) architecture pipeline model, tuning and
8202 * config.gcc: Add slm config options and target.
8204 * config/i386/slm.md: New.
8206 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
8208 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): New case
8210 (ix86_target_macros_internal): Likewise.
8212 * gcc/config/i386/i386.c (slm_cost): New cost.
8213 (m_SLM): New macro flag.
8214 (initial_ix86_tune_features): Set m_SLM.
8215 (x86_accumulate_outgoing_args): Likewise.
8216 (x86_arch_always_fancy_math_387): Likewise.
8217 (processor_target_table): Add slm cost.
8218 (cpu_names): Add slm cpu name.
8219 (x86_option_override_internal): Set SLM ISA.
8220 (ix86_issue_rate): New case PROCESSOR_SLM.
8221 (ia32_multipass_dfa_lookahead): Likewise.
8222 (fold_builtin_cpu): Add slm.
8224 * config/i386/i386.h (TARGET_SLM): New target macro.
8225 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
8226 (processor_type): Add PROCESSOR_SLM.
8228 * config/i386/i386.md (cpu): Add new value "slm".
8229 (slm.md): Include slm.md.
8231 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
8232 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8234 * config/arm/arm-protos.h: Add and update function protos.
8235 * config/arm/arm.c (use_simple_return_p): New added.
8236 (thumb2_expand_return): Check simple_return flag.
8237 * config/arm/arm.md: Add simple_return and conditional simple_return.
8238 * config/arm/iterators.md: Add iterator for return and simple_return.
8240 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8242 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
8243 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
8244 (arm_emit_vfp_multi_reg_pop): Likewise.
8245 (thumb2_emit_ldrd_pop): Likewise.
8246 (arm_expand_epilogue): Add misc REG_CFA notes.
8247 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
8249 2013-05-29 Lawrence Crowl <crowl@google.com>
8251 * config/arm/t-arm: Update for below.
8253 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
8254 Change type to hash_table. Update dependent calls and types.
8256 * config/i386/t-cygming: Update for below.
8258 * config/i386/t-interix: Update for below.
8260 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
8261 Change type to hash_table. Update dependent calls and types.
8262 (i386_find_on_wrapper_list::wrappers): Likewise.
8264 * config/ia64/t-ia64: Update for below.
8266 * config/ia64/ia64.c (bundle_state_table):
8267 Change type to hash_table. Update dependent calls and types.
8269 * config/mips/mips.c (mips_reorg_process_insns::htab):
8270 Change type to hash_table. Update dependent calls and types.
8272 * config/sol2.c (solaris_comdat_htab):
8273 Change type to hash_table. Update dependent calls and types.
8275 * config/t-sol2: Update for above.
8277 2013-05-29 Teresa Johnson <tejohnson@google.com>
8279 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
8280 functions are not yet marked as defined.
8282 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
8283 Pat Haugen <pthaugen@us.ibm.com>
8284 Peter Bergner <bergner@vnet.ibm.com>
8286 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
8290 (vrotl<mode>3): Likewise.
8291 (vashl<mode>3): Likewise.
8292 (vlshr<mode>3): Likewise.
8293 (vashr<mode>3): Likewise.
8295 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
8296 support for power8 V2DI builtins.
8298 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
8299 power8 V2DI builtins.
8300 (vupkhsw): Likewise.
8301 (vupklsw): Likewise.
8302 (vaddudm): Likewise.
8307 (vpkudum): Likewise.
8308 (vpksdss): Likewise.
8309 (vpkudus): Likewise.
8310 (vpksdus): Likewise.
8315 (vsubudm): Likewise.
8316 (vcmpequd): Likewise.
8317 (vcmpgtsd): Likewise.
8318 (vcmpgtud): Likewise.
8319 (vcmpequd_p): Likewise.
8320 (vcmpgtsd_p): Likewise.
8321 (vcmpgtud_p): Likewise.
8322 (vupkhsw): Likewise.
8323 (vupklsw): Likewise.
8324 (vaddudm): Likewise.
8329 (vpksdss): Likewise.
8330 (vpksdus): Likewise.
8331 (vpkudum): Likewise.
8332 (vpkudus): Likewise.
8337 (vsubudm): Likewise.
8339 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
8340 support for power8 V2DI instructions.
8342 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
8343 power8 V2DI instructions. Combine pack and unpack insns to use an
8344 iterator for each mode. Check whether a particular mode supports
8345 Altivec instructions instead of just checking TARGET_ALTIVEC.
8346 (UNSPEC_VPKUWUM): Likewise.
8347 (UNSPEC_VPKSHSS): Likewise.
8348 (UNSPEC_VPKSWSS): Likewise.
8349 (UNSPEC_VPKUHUS): Likewise.
8350 (UNSPEC_VPKSHUS): Likewise.
8351 (UNSPEC_VPKUWUS): Likewise.
8352 (UNSPEC_VPKSWUS): Likewise.
8353 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
8354 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
8355 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
8356 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
8357 (UNSPEC_VUPKHSB): Likewise.
8358 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
8359 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
8360 (UNSPEC_VUPKHSH): Likewise.
8361 (UNSPEC_VUPKLSB): Likewise.
8362 (UNSPEC_VUPKLSH): Likewise.
8364 (VI_char): Likewise.
8365 (VI_scalar): Likewise.
8366 (VI_unit): Likewise.
8368 (VP_small): Likewise.
8369 (VP_small_lc): Likewise.
8370 (VU_char): Likewise.
8371 (add<mode>3): Likewise.
8372 (altivec_vaddcuw): Likewise.
8373 (altivec_vaddu<VI_char>s): Likewise.
8374 (altivec_vadds<VI_char>s): Likewise.
8375 (sub<mode>3): Likewise.
8376 (altivec_vsubcuw): Likewise.
8377 (altivec_vsubu<VI_char>s): Likewise.
8378 (altivec_vsubs<VI_char>s): Likewise.
8379 (altivec_vavgs<VI_char>): Likewise.
8380 (altivec_vcmpbfp): Likewise.
8381 (altivec_eq<mode>): Likewise.
8382 (altivec_gt<mode>): Likewise.
8383 (altivec_gtu<mode>): Likewise.
8384 (umax<mode>3): Likewise.
8385 (smax<mode>3): Likewise.
8386 (umin<mode>3): Likewise.
8387 (smin<mode>3): Likewise.
8388 (altivec_vpkuhum): Likewise.
8389 (altivec_vpkuwum): Likewise.
8390 (altivec_vpkshss): Likewise.
8391 (altivec_vpkswss): Likewise.
8392 (altivec_vpkuhus): Likewise.
8393 (altivec_vpkshus): Likewise.
8394 (altivec_vpkuwus): Likewise.
8395 (altivec_vpkswus): Likewise.
8396 (altivec_vpks<VI_char>ss): Likewise.
8397 (altivec_vpks<VI_char>us): Likewise.
8398 (altivec_vpku<VI_char>us): Likewise.
8399 (altivec_vpku<VI_char>um): Likewise.
8400 (altivec_vrl<VI_char>): Likewise.
8401 (altivec_vsl<VI_char>): Likewise.
8402 (altivec_vsr<VI_char>): Likewise.
8403 (altivec_vsra<VI_char>): Likewise.
8404 (altivec_vsldoi_<mode>): Likewise.
8405 (altivec_vupkhsb): Likewise.
8406 (altivec_vupkhs<VU_char>): Likewise.
8407 (altivec_vupkls<VU_char>): Likewise.
8408 (altivec_vupkhsh): Likewise.
8409 (altivec_vupklsb): Likewise.
8410 (altivec_vupklsh): Likewise.
8411 (altivec_vcmpequ<VI_char>_p): Likewise.
8412 (altivec_vcmpgts<VI_char>_p): Likewise.
8413 (altivec_vcmpgtu<VI_char>_p): Likewise.
8414 (abs<mode>2): Likewise.
8415 (vec_unpacks_hi_v16qi): Likewise.
8416 (vec_unpacks_hi_v8hi): Likewise.
8417 (vec_unpacks_lo_v16qi): Likewise.
8418 (vec_unpacks_hi_<VP_small_lc>): Likewise.
8419 (vec_unpacks_lo_v8hi): Likewise.
8420 (vec_unpacks_lo_<VP_small_lc>): Likewise.
8421 (vec_pack_trunc_v8h): Likewise.
8422 (vec_pack_trunc_v4si): Likewise.
8423 (vec_pack_trunc_<mode>): Likewise.
8425 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
8427 (vec_vmaxsd): Likewise.
8428 (vec_vmaxud): Likewise.
8429 (vec_vminsd): Likewise.
8430 (vec_vminud): Likewise.
8431 (vec_vpksdss): Likewise.
8432 (vec_vpksdus): Likewise.
8433 (vec_vpkudum): Likewise.
8434 (vec_vpkudus): Likewise.
8435 (vec_vrld): Likewise.
8436 (vec_vsld): Likewise.
8437 (vec_vsrad): Likewise.
8438 (vec_vsrd): Likewise.
8439 (vec_vsubudm): Likewise.
8440 (vec_vupkhsw): Likewise.
8441 (vec_vupklsw): Likewise.
8443 2013-05-29 Jan Hubicka <jh@suse.cz>
8445 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
8446 flags; reorder rest of fields in more consistent way.
8447 (varpool_node): Remove analyzed, finalized and alias.
8448 (cgraph_ndoe): Likewise.
8449 (symtab_alias_ultimate_target): New function.
8450 (cgraph_function_node): Move offline.
8451 (cgraph_reset_node): Declare.
8452 (cgraph_comdat_can_be_unshared_p): Remove.
8453 (varpool_remove_initializer): Declare.
8454 (varpool_first_defined_variable, varpool_next_defined_variable
8455 cgraph_first_defined_function, cgraph_next_defined_function): Update.
8456 (cgraph_function_with_gimple_body_p): Update.
8457 (varpool_all_refs_explicit_p): Update.
8458 (symtab_alias_target): New function.
8459 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
8460 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
8461 (cgraph_function_or_thunk_node): Simplify using
8462 symtab_alias_ultimate_target.
8463 (varpool_variable_node): Likewise.
8464 * cgraph.c (cgraph_create_function_alias): Update.
8465 (cgraph_add_thunk): Update.
8466 (cgraph_remove_node): Update.
8467 (dump_cgraph_node): Do not dump removed flags.
8468 (cgraph_function_body_availability): Update.
8469 (cgraph_propagate_frequency): Update.
8470 (verify_cgraph_node): Check sanity of local flag.
8471 (cgraph_function_node): Move here from cgraph.h; revamp for
8472 cgraph_function_or_thunk_node.
8473 * lto-symtab.c (lto_varpool_replace_node): Update.
8474 (lto_symtab_resolve_can_prevail_p): Update.
8475 (lto_symtab_merge_cgraph_nodes): Update.
8476 * ipa-cp.c (determine_versionability, initialize_node_lattices,
8477 propagate_constants_accross_call, devirtualization_time_bonus,
8478 ipcp_propagate_stage): Update.
8479 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
8480 * ipa-inline-transform.c (clone_inlined_nodes,
8481 preserve_function_body_p): Update.
8482 * ipa-reference.c (propagate): Update.
8483 (write_node_summary_p): Update.
8484 * toplev.c (wrapup_global_declaration_2): Update.
8485 * cgraphunit.c (cgraph_analyze_function): Rename to ...
8486 (analyze_function) ... this one.
8487 (cgraph_process_new_functions): Update.
8488 (cgraph_reset_node): Export.
8489 (cgraph_finalize_function): Update.
8490 (cgraph_add_new_function): Update.
8491 (process_function_and_variable_attributes): Update.
8492 (varpool_finalize_decl): Update.
8493 (symbol_finalized): Remove.
8494 (symbol_finalized_and_needed): Rename to ...
8495 (symbol_defined_and_needed): ... update.
8496 (cgraph_analyze_functions): Update.
8497 (handle_alias_pairs): Update.
8498 (mark_functions_to_output): Update.
8499 (assemble_thunk): Update.
8500 (output_in_order): Update.
8501 (output_weakrefs): Update.
8502 (finalize_compilation_unit): Update.
8503 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
8504 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
8505 input_node, input_varpool_node): Update.
8506 * dbxout.c (dbxout_expand_expr): Update.
8507 * cgraphclones.c (cgraph_clone_node): Update.
8508 (cgraph_copy_node_for_versioning): Update.
8509 (cgraph_materialize_clone): Update.
8510 (cgraph_materialize_all_clones): Update.
8511 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
8512 propagate_pure_const, propagate_nothrow): Update.
8513 * lto-streamer-out.c (lto_output, write_symbol): Update.
8514 * ipa-utils.c (ipa_reverse_postorder): Update.
8515 * ipa-inline.c (can_inline_edge_p): Update.
8516 (update_caller_keys, ipa_inline): Update.
8517 * dwarf2out.c (reference_to_unused,
8518 premark_types_used_by_global_vars_helper): Update.
8519 * tree-eh.c (tree_could_trap_p): Update.
8520 * ipa-split.c (consider_split, execute_split_functions): Update.
8521 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
8522 has_addr_references_p): Update; move ahead in file for better
8524 (process_references): Simplify.
8525 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
8527 (cgraph_comdat_can_be_unshared_p): Make static.
8528 (cgraph_externally_visible_p): Update.
8529 (varpool_externally_visible_p): Update.
8530 (function_and_variable_visibility): Update.
8531 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
8532 ipa_tm_mark_force_output_node): Update.
8533 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
8534 estimate_edge_devirt_benefit, inline_generate_summary,
8535 inline_write_summary): Update.
8536 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
8537 * ipa-prop.c (ipa_compute_jump_functions): Update.
8538 (ipa_print_node_params, ipa_prop_read_section,
8539 ipa_update_after_lto_read, read_replacements_section): Update.
8540 * varasm.c (mark_decl_referenced): Update.
8541 (assemble_alias, dump_tm_clone_pairs): Update.
8542 * tree-inline.c (copy_bb): Update.
8543 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
8545 * symtab.c (dump_symtab_base): Print new flags.
8546 (verify_symtab_base): Verify new flags.
8547 (symtab_alias_ultimate_target): New function.
8548 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
8549 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
8551 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
8553 * i386.c (ix86_get_function_versions_dispatcher,
8554 ix86_generate_version_dispatcher_body): Update.
8555 (fold_builtin_cpu): Use varpool_add_new_variable.
8556 * varpool.c (varpool_remove_initializer): Break out from ...
8557 (varpool_remove_node): ... this one.
8558 (dump_varpool_node, varpool_node_for_asm,
8559 cgraph_variable_initializer_availability, varpool_analyze_node,
8560 varpool_assemble_decl, varpool_remove_unreferenced_decls,
8561 varpool_finalize_named_section_flags, varpool_create_variable_alias):
8564 2013-05-29 Jan Hubicka <jh@suse.cz>
8566 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
8568 2013-05-29 Easwaran Raman <eraman@google.com>
8570 PR tree-optimization/57442
8571 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
8572 when control exits the main loop.
8574 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
8576 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
8578 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
8579 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
8580 * rx/t-rx: Add rx100 under multi library matches option for nofpu
8583 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8585 PR tree-optimization/57441
8586 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
8587 Don't limit size of incr_vec to number of candidates.
8589 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
8591 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
8592 and mips16 directories.
8593 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
8594 (MULTILIB_DIRNAMES): Ditto.
8595 (MULTILIB_EXCEPTIONS): Add new exceptions.
8596 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
8597 (MULTILIB_DIRNAMES): Ditto.
8598 (MULTILIB_EXCEPTIONS): Add new exceptions.
8600 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8601 Marcus Shawcroft <marcus.shawcroft@arm.com>
8603 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
8604 SYMBOL_TINY_ABSOLUTE.
8605 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
8606 SYMBOL_TINY_ABSOLUTE.
8607 (aarch64_expand_mov_immediate): Likewise.
8608 (aarch64_classify_symbol): Likewise.
8609 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
8610 Permit SYMBOL_TINY_ABSOLUTE.
8611 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
8613 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8614 Marcus Shawcroft <marcus.shawcroft@arm.com>
8616 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
8617 Refactor if/switch. Replace gcc_assert with if.
8619 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
8621 * config/i386/i386.c (initial_ix86_tune_features): Enable
8622 FP Reassociation for AMD bdver1 and bdver2.
8624 2013-05-29 Martin Jambor <mjambor@suse.cz>
8626 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
8627 and IMAGPART_EXPR do not occur within other handled_components.
8629 2013-05-29 Richard Biener <rguenther@suse.de>
8631 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
8632 access on whether the use is in the BB we currently try to
8634 (vect_bb_vectorization_profitable_p): Pass the BB we currently
8635 vectorize to vect_bb_slp_scalar_cost.
8637 2013-05-29 Richard Biener <rguenther@suse.de>
8639 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
8640 computing scalar cost offsetted by stmts that are kept live
8642 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
8643 for computation of scalar cost.
8645 2013-05-28 Steve Ellcey <sellcey@mips.com>
8647 * config/mips/mips-cpus.def (mips32r2): Change processor type.
8649 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8651 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
8652 array notation built-in reduction functions.
8653 * doc/passes.texi (Passes): Added documentation about changes done
8655 * doc/invoke.texi (C Dialect Options): Added documentation about
8656 the -fcilkplus flag.
8657 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
8658 (BUILTINS_DEF): Depend on cilkplus.def.
8659 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
8660 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
8661 * cilkplus.def: New file.
8663 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
8665 PR rtl-optimization/57439
8666 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
8668 2013-05-28 Easwaran Raman <eraman@google.com>
8670 PR tree-optimization/57337
8671 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
8672 (find_insert_point): Correctly identify the insertion point
8673 when two statements with the same UID is compared.
8675 2013-05-28 Richard Biener <rguenther@suse.de>
8677 PR tree-optimization/56787
8678 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
8679 from the list of data references.
8680 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
8682 (vect_analyze_loop_operations): Likewise.
8683 (vect_transform_loop): Remove clobbers.
8685 2013-05-28 Martin Jambor <mjambor@suse.cz>
8687 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
8688 and REALPART_EXPRs have scalar type.
8690 2013-05-28 Richard Biener <rguenther@suse.de>
8692 PR tree-optimization/57411
8693 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
8695 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
8696 virtual operand propagation.
8698 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8700 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
8701 destination register for bmasksi_vis.
8702 (vector_init_bshuffle): Likewise.
8703 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
8705 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
8707 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
8708 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
8709 mode if the instruction isn't available in the original mode.
8710 * config/sparc/sparc.opt (mfix-ut699): New option.
8711 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
8712 (divdf3): Turn into expander.
8713 (divdf3_nofix): New insn.
8714 (divdf3_fix): Likewise.
8715 (divsf3): Disable if -mfix-ut699.
8716 (sqrtdf2): Turn into expander.
8717 (sqrtdf2_nofix): New insn.
8718 (sqrtdf2_fix): Likewise.
8719 (sqrtsf2): Disable if -mfix-ut699.
8721 2013-05-27 Richard Biener <rguenther@suse.de>
8724 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
8725 block for the new loop.
8727 2013-05-27 Richard Biener <rguenther@suse.de>
8729 PR tree-optimization/57343
8730 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
8731 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
8732 (number_of_iterations_cond): Do not build the folded tree.
8734 2013-05-27 Richard Biener <rguenther@suse.de>
8738 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8739 OEP_CONSTANT_ADDRESS_OF retained.
8741 PR tree-optimization/57417
8742 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
8744 (set_ssa_val_to): Compare addresses using
8745 get_addr_base_and_unit_offset.
8747 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
8749 PR rtl-optimization/56833
8750 * postreload.c (move2add_record_mode): New function.
8751 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
8752 (move2add_use_add2_insn): Use move2add_record_sym_value.
8753 (move2add_use_add3_insn): Likewise.
8754 (reload_cse_move2add): Use move2add_valid_value_p and
8755 move2add_record_mode. Invalidate call-clobbered and REG_INC
8756 affected regs by setting reg_mode to VOIDmode.
8757 (move2add_note_store): Don't pretend the inside of a SUBREG is
8758 the actual destination. Invalidate single/leading registers by
8759 setting reg_mode to VOIDmode.
8760 Use move2add_record_sym_value, move2add_valid_value_p and
8761 move2add_record_mode.
8763 2013-05-27 Richard Biener <rguenther@suse.de>
8765 PR tree-optimization/57396
8766 * tree-affine.c (double_int_constant_multiple_p): Properly
8767 return false for val == 0 and div != 0.
8769 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8771 * config/mips/mips.h: Use #elif in preprocessor conditions.
8773 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8776 * config/mips/constraints.md (kl): New constraint.
8777 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
8778 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
8779 constraint for operand 0. Split after CSE for MIPS16. Emit a move
8781 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
8783 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
8786 * config/mips/mips.c (mips_can_inline_p): New function.
8787 (TARGET_CAN_INLINE_P): Define.
8789 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
8791 * sched-int.h (ds_t, dw_t): Make unsigned int.
8792 Fix documentation that describes how all the ds_t bits are used.
8793 Reserve the last bit for delayed-branch scheduling.
8794 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
8795 (BITS_PER_DEP_WEAK): Fix definition and documentation.
8796 (gen_dep_weak_1): Remove prototype.
8797 * sched-deps.c (get_dep_weak_1): Make static.
8798 * target.def (speculate_insn, needs_block_p, gen_spec_check,
8799 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
8800 * doc/tm.texi: Regenerate.
8801 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
8803 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
8806 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
8808 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
8809 Sandra Loosemore <sandra@codesourcery.com>
8811 * config.gcc (powerpc-*): Allow native for with-cpu.
8813 2013-05-24 Jeff Law <law@redhat.com>
8815 PR tree-optimization/57124
8816 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
8817 conversion feeding a condition if the range has an overflow
8818 if -fstrict-overflow. Add warnings for when we do make the
8821 2013-05-24 Dehao Chen <dehao@google.com>
8823 * tree-cfg.c (locus_discrim_map): Fix the typo.
8824 (locus_discrim_hasher): Likewise.
8825 (locus_discrim_hasher::hash): Likewise.
8826 (locus_discrim_hasher::equal): Likewise.
8828 2013-05-24 Martin Jambor <mjambor@suse.cz>
8830 PR tree-optimization/57294
8831 * cgraph.h (ipa_record_stmt_references): Declare.
8832 * cgraphbuild.c (ipa_record_stmt_references): New function.
8833 (build_cgraph_edges): Use ipa_record_stmt_references.
8834 (rebuild_cgraph_edges): Likewise.
8835 (cgraph_rebuild_references): Likewise.
8836 * ipa-prop.c (ipa_modify_call_arguments): Discard references
8837 associated with the old statement and build references from the
8838 newly built statements.
8839 * ipa-ref.c (ipa_remove_stmt_references): New function.
8840 * ipa-ref.h (ipa_remove_stmt_references): Declare.
8842 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
8844 * lra-constraints.c (emit_spill_move): Use smaller mode for
8846 (check_and_process_move): Consider mem-reg moves for secondary
8848 (curr_insn_transform): Don't lose insns emitted before for
8849 secondary memory moves.
8850 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
8851 reg set up in the current insn.
8853 2013-05-24 Dehao Chen <dehao@google.com>
8855 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
8857 (locus_descrim_hasher::equal): Likewise.
8858 (build_gimple_cfg): New discriminator assignment algorithm.
8859 (make_edges): Likewise.
8860 (next_discriminator_for_locus): Likewise.
8861 (same_line_p): Likewise.
8862 (assign_discriminators): Likewise.
8863 (make_cond_expr_edges): Likewise.
8864 (make_gimple_switch_edges): Likewise.
8865 (make_goto_expr_edges): Likewise.
8866 (make_gimple_asm_edges): Likewise.
8868 2013-05-24 Ian Bolton <ian.bolton@arm.com>
8870 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
8871 X format specifier to only display bottom 16 bits.
8872 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
8873 immediate to match for operand 2, since it will be masked.
8875 2013-05-24 Richard Biener <rguenther@suse.de>
8877 PR tree-optimization/57287
8878 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
8879 all SSA names that occur in abnormal PHIs.
8881 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
8884 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
8885 that index is not negative.
8887 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
8889 PR rtl-optimization/55177
8890 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
8891 (simplify_byte_swapping_operation): New.
8892 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
8893 (simplify_relational_operation_1): Deal with BSWAP.
8895 2013-05-23 Richard Henderson <rth@redhat.com>
8898 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
8899 (ix86_reorg): Call it.
8901 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
8904 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
8905 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
8906 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
8908 2013-05-23 Christian Bruel <christian.bruel@st.com>
8911 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
8913 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8914 Marcus Shawcroft <marcus.shawcroft@arm.com>
8916 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
8917 * config/aarch64/constraints.md (Usa): Remove.
8918 * doc/md.texi (AArch64 Usa): Remove.
8920 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
8921 Marcus Shawcroft <marcus.shawcroft@arm.com>
8923 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
8924 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
8925 * config/aarch64/predicates.md (aarch64_const_address): Remove.
8926 (aarch64_mov_operand): Use aarch64_mov_operand_p.
8928 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
8930 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
8931 instruction (AdvSIMD).
8932 * config/aarch64/aarch64-builtins.c
8933 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
8934 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
8936 2013-05-23 Martin Jambor <mjambor@suse.cz>
8939 * tree.h (contains_bitfld_component_ref_p): Declare.
8940 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
8941 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
8943 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
8944 not access a bit-field. Assert all final offsets are byte-aligned.
8946 2013-05-23 Richard Biener <rguenther@suse.de>
8948 PR tree-optimization/57380
8949 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
8950 least one invariant or re-used load.
8951 * passes.c (init_optimization_passes): Move pass_phiprop before
8954 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
8956 * config/aarch64/aarch64-simd.md
8957 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
8959 2013-05-23 Richard Biener <rguenther@suse.de>
8962 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
8963 OEP_CONSTANT_ADDRESS_OF retained.
8965 2013-05-23 Jakub Jelinek <jakub@redhat.com>
8968 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
8969 don't lower unit. Handle unit not being always BITS_PER_WORD.
8971 2013-05-23 Richard Biener <rguenther@suse.de>
8973 PR rtl-optimization/57341
8974 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
8975 instead of true_dependence.
8977 2013-05-22 David Malcolm <dmalcolm@redhat.com>
8979 * bb-reorder.c (branch_threshold): Make const.
8980 (exec_threshold): Ditto.
8982 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
8983 Pat Haugen <pthaugen@us.ibm.com>
8984 Peter Bergner <bergner@vnet.ibm.com>
8986 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
8987 documentation for the power8 crypto builtins.
8989 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
8991 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
8992 macros for defining power8 builtin functions.
8993 (BU_P8V_AV_2): Likewise.
8994 (BU_P8V_AV_P): Likewise.
8995 (BU_P8V_VSX_1): Likewise.
8996 (BU_P8V_OVERLOAD_1): Likewise.
8997 (BU_P8V_OVERLOAD_2): Likewise.
8998 (BU_CRYPTO_1): Likewise.
8999 (BU_CRYPTO_2): Likewise.
9000 (BU_CRYPTO_3): Likewise.
9001 (BU_CRYPTO_OVERLOAD_1): Likewise.
9002 (BU_CRYPTO_OVERLOAD_2): Likewise.
9003 (XSCVSPDP): Fix typo, point to the correct instruction.
9004 (VCIPHER): Add power8 crypto builtins.
9005 (VCIPHERLAST): Likewise.
9006 (VNCIPHER): Likewise.
9007 (VNCIPHERLAST): Likewise.
9008 (VPMSUMB): Likewise.
9009 (VPMSUMH): Likewise.
9010 (VPMSUMW): Likewise.
9011 (VPERMXOR_V2DI): Likewise.
9012 (VPERMXOR_V4SI: Likewise.
9013 (VPERMXOR_V8HI: Likewise.
9014 (VPERMXOR_V16QI: Likewise.
9015 (VSHASIGMAW): Likewise.
9016 (VSHASIGMAD): Likewise.
9018 (VPERMXOR): Likewise.
9019 (VSHASIGMA): Likewise.
9021 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
9022 __CRYPTO__ if the crypto instructions are available.
9023 (altivec_overloaded_builtins): Add support for overloaded power8
9026 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
9027 support for power8 crypto builtins.
9028 (builtin_function_type): Likewise.
9029 (altivec_init_builtins): Add support for builtins that take vector
9030 long long (V2DI) arguments.
9032 * config/rs6000/crypto.md: New file, define power8 crypto
9035 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
9036 Pat Haugen <pthaugen@us.ibm.com>
9037 Peter Bergner <bergner@vnet.ibm.com>
9039 * doc/invoke.texi (Option Summary): Add power8 options.
9040 (RS/6000 and PowerPC Options): Likewise.
9042 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
9043 constraints.md instead of rs6000.h. Reorder w* constraints. Add
9044 wm, wn, wr documentation.
9046 * gcc/config/rs6000/constraints.md (wm): New constraint for VSX
9047 registers if direct move instructions are enabled.
9048 (wn): New constraint for no registers.
9049 (wq): New constraint for quad word even GPR registers.
9050 (wr): New constraint if 64-bit instructions are enabled.
9051 (wv): New constraint if power8 vector instructions are enabled.
9052 (wQ): New constraint for quad word memory locations.
9054 * gcc/config/rs6000/predicates.md (const_0_to_15_operand): New
9055 constraint for 0..15 for crypto instructions.
9056 (gpc_reg_operand): If VSX allow registers in VSX registers as well
9057 as GPR and floating point registers.
9058 (int_reg_operand): New predicate to match only GPR registers.
9059 (base_reg_operand): New predicate to match base registers.
9060 (quad_int_reg_operand): New predicate to match even GPR registers
9061 for quad memory operations.
9062 (vsx_reg_or_cint_operand): New predicate to allow vector logical
9063 operations in both GPR and VSX registers.
9064 (quad_memory_operand): New predicate for quad memory operations.
9065 (reg_or_indexed_operand): New predicate for direct move support.
9067 * gcc/config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
9068 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
9069 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
9070 (POWERPC_MASKS): Add power8 options.
9071 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
9074 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
9075 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
9077 * gcc/config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
9078 (-mpower8-fusion): New power8 options.
9079 (-mpower8-fusion-sign): Likewise.
9080 (-mpower8-vector): Likewise.
9081 (-mcrypto): Likewise.
9082 (-mdirect-move): Likewise.
9083 (-mquad-memory): Likewise.
9085 * gcc/config/rs6000/rs6000.c (power8_cost): Initial definition for
9087 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
9089 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
9090 (rs6000_debug_vector_unit): Add p8_vector.
9091 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
9092 definitions. Also print fusion state.
9093 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
9094 (rs6000_builtin_mask_calculate): Add power8 builtin support.
9095 (rs6000_option_override_internal): Add support for power8.
9096 (rs6000_common_init_builtins): Add debugging for skipped builtins
9098 (rs6000_adjust_cost): Add power8 support.
9099 (rs6000_issue_rate): Likewise.
9100 (insn_must_be_first_in_group): Likewise.
9101 (insn_must_be_last_in_group): Likewise.
9102 (force_new_group): Likewise.
9103 (rs6000_register_move_cost): Likewise.
9104 (rs6000_opt_masks): Likewise.
9106 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
9107 power8 capable assembler, default to power7 options.
9108 (TARGET_DIRECT_MOVE): Likewise.
9109 (TARGET_CRYPTO): Likewise.
9110 (TARGET_P8_VECTOR): Likewise.
9111 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
9112 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
9113 (VECTOR_MEM_P8_VECTOR_P): Likewise.
9114 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
9115 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
9116 (TARGET_XSCVDPSPN): Likewise.
9117 (TARGET_XSCVSPDPN): Likewsie.
9118 (TARGET_SYNC_HI_QI): Likewise.
9119 (TARGET_SYNC_TI): Likewise.
9120 (MASK_CRYPTO): Likewise.
9121 (MASK_DIRECT_MOVE): Likewise.
9122 (MASK_P8_FUSION): Likewise.
9123 (MASK_P8_VECTOR): Likewise.
9124 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
9125 temporary used by some of the direct move instructions to get two FP
9126 temporary registers does not force creation of a stack frame.
9127 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
9128 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
9129 that any VSX registers are tieable, even if they are also an
9130 Altivec vector mode.
9131 (r6000_reg_class_enum): Add wm, wr, wv constraints.
9132 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
9133 (RS6000_BTM_CRYPTO): Likewise.
9134 (RS6000_BTM_COMMON): Likewise.
9136 * config/rs6000/rs6000.md (cpu attribute): Add power8.
9137 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
9138 (enum rs6000_vector): Add power8 vector support.
9140 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9144 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
9145 (any_sibcall_could_use_r3): this and handle indirect calls.
9146 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
9148 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9150 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
9152 2013-05-22 Richard Biener <rguenther@suse.de>
9155 * profile.c (branch_prob): Do not split blocks that are
9156 abnormally receiving from ECF_RETURNS_TWICE functions.
9158 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9160 * recog.c (offsettable_address_addr_space_p): Fix calculation of
9161 address mode. Move pointer mode initialization to the same place.
9163 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9165 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
9166 while it has any effect.
9168 2013-05-21 Easwaran Raman <eraman@google.com>
9170 PR tree-optimization/57322
9171 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
9172 UID of the statement added to the BB to be 1.
9174 2013-05-21 Jakub Jelinek <jakub@redhat.com>
9176 PR tree-optimization/57331
9177 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
9178 of conversion from pointer type to integral type with integer.
9180 2013-05-21 Martin Jambor <mjambor@suse.cz>
9183 * ipa-prop.c (ipa_read_node_info): Process param_used and
9184 controlled_uses in the same order as when writing.
9186 2013-05-21 Magnus Granberg <baldrick@free.fr>
9189 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
9191 2013-05-21 Richard Biener <rguenther@suse.de>
9193 PR tree-optimization/57318
9194 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
9195 estimate stmts with side-effects as likely eliminated.
9197 2013-05-21 Richard Biener <rguenther@suse.de>
9199 PR tree-optimization/57330
9200 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
9201 preserve the call stmts fntype.
9203 2013-05-21 Richard Biener <rguenther@suse.de>
9205 PR tree-optimization/57303
9206 * tree-ssa-sink.c (statement_sink_location): Improve killing
9207 stmt detection and properly handle self-assignments.
9209 2013-05-21 Christian Bruel <christian.bruel@st.com>
9211 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
9212 spanning registers. LEAF_REG_REMAP is supported only for contiguous
9213 registers. Set register size out of the PARALLEL loop.
9215 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
9218 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
9219 (*fmasf4, *fmasf4_media): New insns.
9221 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9223 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
9224 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
9225 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
9226 (mips_idiv_insns): Update the comments to say that the returned
9227 instruction counts are in units of BASE_INSN_LENGTH.
9228 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
9229 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
9230 using 2 rather than 4 as the length of indirect MIPS16 and
9231 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
9232 length of a NOP. Don't divide MIPS16 lengths by 2.
9233 (mips16_split_long_branches): Assume a branch is long if the
9234 length is greater than 4 rather than 8.
9235 * config/mips/mips.md (length): Give MIPS16 lengths directly,
9236 rather than multiplying them by 2. Multiply instruction counts
9237 by BASE_INSN_LENGTH rather than 4.
9238 (*jump_mips16, tls_get_tp_mips16_<mode>)
9239 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
9241 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9243 * config/mips/mips.md (extended_mips16): Remove branch case.
9244 (length): Remove duplicated extended_mips16 test.
9246 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
9248 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
9250 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
9252 * recog.h (Recog_data): Rename to...
9253 (recog_data_d): ...this.
9254 (recog_data): Update accordingly.
9255 * recog.c (recog_data): Likewise.
9256 * reload.c (save_recog_data): Likewise.
9257 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
9258 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
9260 2013-05-17 Julian Brown <julian@codesourcery.com>
9262 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
9263 found in a REG_EQUAL note, invalidate it.
9265 2013-05-17 Easwaran Raman <eraman@google.com>
9267 * tree-ssa-reassoc.c (find_insert_point): New function.
9268 (insert_stmt_after): Likewise.
9269 (get_def_stmt): Likewise.
9270 (ensure_ops_are_available): Likewise.
9271 (not_dominated_by): Likewise.
9272 (rewrite_expr_tree): Do not move statements beyond what is
9273 necessary. Remove call to swap_ops_for_binary_stmt...
9274 (reassociate_bb): ... and move it here.
9275 (build_and_add_sum): Assign UIDs for new statements.
9276 (linearize_expr): Likewise.
9277 (do_reassoc): Renumber gimple statement UIDs.
9279 2013-05-17 Jan Hubicka <jh@suse.cz>
9281 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
9283 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
9284 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
9286 (output_weakrefs): Update.
9288 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
9289 Martin Jambor <mjambor@suse.cz>
9292 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
9293 value that corresponds to the given aggval is found in values vector.
9295 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
9297 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
9298 sse, sse2, sse3, ssse3 and sse4a flags to options.
9300 2013-05-17 David Malcolm <dmalcolm@redhat.com>
9302 * gengtype-state.c: (s_expr_writer): New class, to handle
9303 prettifying of output layout of s-expressions.
9304 (state_writer): New class, to write out gtype.state.
9305 (state_written_type_count): Move this variable into member data of
9307 (s_expr_writer::s_expr_writer): New code: constructor for new class
9308 (state_writer::state_writer(): ditto
9309 (s_expr_writer::write_new_line): New function
9310 (s_expr_writer::write_any_indent): ditto
9311 (s_expr_writer::begin_s_expr): ditto
9312 (s_expr_writer::end_s_expr): ditto
9313 (write_state_fileloc): convert to method of state_writer...
9314 (state_writer:: write_state_fileloc): ...and use methods of
9315 s_expr_writer to write indentation into the gtype.state output file
9316 to visually represent the hierarchical structure of the list
9318 (write_state_fields): ditto, renaming to...
9319 (state_writer::write_state_fields)
9320 (write_state_a_string): ditto, renaming to...
9321 (state_writer::write_state_a_string)
9322 (write_state_string_option): ditto, renaming to...
9323 (state_writer::write_state_string_option)
9324 (write_state_type_option): ditto, renaming to...
9325 (state_writer::write_state_type_option)
9326 (write_state_nested_option): ditto, renaming to...
9327 (state_writer::write_state_nested_option)
9328 (write_state_option): ditto, renaming to...
9329 (state_writer::write_state_option)
9330 (write_state_options): ditto, renaming to...
9331 (state_writer::write_state_options)
9332 (write_state_lang_bitmap): ditto, renaming to...
9333 (state_writer::write_state_lang_bitmap)
9334 (write_state_version): ditto, renaming to...
9335 (state_writer::write_state_version)
9336 (write_state_scalar_type): ditto, renaming to...
9337 (state_writer::write_state_scalar_type)
9338 (write_state_string_type): ditto, renaming to...
9339 (state_writer::write_state_string_type)
9340 (write_state_undefined_type): ditto, renaming to...
9341 (state_writer::write_state_undefined_type)
9342 (write_state_struct_union_type): ditto, renaming to...
9343 (state_writer::write_state_struct_union_type)
9344 (write_state_struct_type): ditto, renaming to...
9345 (state_writer::write_state_struct_type)
9346 (write_state_user_struct_type): ditto, renaming to...
9347 (state_writer::write_state_user_struct_type)
9348 (write_state_lang_struct_type): ditto, renaming to...
9349 (state_writer::write_state_lang_struct_type)
9350 (write_state_param_struct_type): ditto, renaming to...
9351 (state_writer::write_state_param_struct_type)
9352 (write_state_pointer_type): ditto, renaming to...
9353 (state_writer::write_state_pointer_type)
9354 (write_state_array_type): ditto, renaming to...
9355 (state_writer::write_state_array_type)
9356 (write_state_gc_used): ditto, renaming to...
9357 (state_writer::write_state_gc_used)
9358 (write_state_common_type_content): ditto, renaming to...
9359 (state_writer::write_state_common_type_content)
9360 (write_state_type): ditto, renaming to...
9361 (state_writer::write_state_type)
9362 (write_state_pair_list): ditto, renaming to...
9363 (state_writer::write_state_pair_list)
9364 (write_state_pair): ditto, renaming to...
9365 (state_writer::write_state_pair)
9366 (write_state_typedefs): ditto, renaming to...
9367 (state_writer::write_state_typedefs)
9368 (write_state_structures): ditto, renaming to...
9369 (state_writer::write_state_structures)
9370 (write_state_param_structs): ditto, renaming to...
9371 (state_writer::write_state_param_structs)
9372 (write_state_variables): ditto, renaming to...
9373 (state_writer::write_state_variables)
9374 (write_state_srcdir): ditto, renaming to...
9375 (state_writer::write_state_srcdir)
9376 (write_state_files_list): ditto, renaming to...
9377 (state_writer::write_state_files_list)
9378 (write_state_languages): ditto, renaming to...
9379 (state_writer::write_state_languages)
9380 (write_state): create a state_writer instance and use it when
9381 writing out the state file
9383 2013-05-17 Mike Stump <mikestump@comcast.net>
9385 PR rtl-optimization/57304
9386 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
9387 accessing DF_REF_REAL_LOC.
9389 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9391 PR rtl-optimization/57281
9392 PR rtl-optimization/57300
9393 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
9394 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
9395 what the other splitter did if the registers are dead.
9397 2013-05-17 Richard Biener <rguenther@suse.de>
9399 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
9402 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9404 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
9407 2013-05-17 Marek Polacek <polacek@redhat.com>
9409 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
9410 length when doing non-zero store of storing '\0' to '\0'.
9412 2013-05-17 Jakub Jelinek <jakub@redhat.com>
9414 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
9415 vect_external_def oprnd1 with loop_vinfo, try to emit
9416 optional cast, negation and and stmts on the loop preheader
9417 edge instead of into the pattern def seq.
9419 PR tree-optimization/57051
9420 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
9421 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
9423 2013-05-16 Nick Clifton <nickc@redhat.com>
9425 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
9426 (rl78_is_naked_func): New function.
9427 (rl78_expand_prologue): Skip prologue generation for naked functions.
9428 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
9429 * doc/extend.texi (naked): Add RL78 to the list of processors
9430 that supports this attribute.
9432 2013-05-16 Jeff Law <law@redhat.com>
9434 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
9436 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9438 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
9439 cache parameters using detect_caches_amd also for CYRIX,
9440 NSC and TM2 signatures.
9442 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
9443 Dzianis Kahanovich <mahatma@eu.by>
9447 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9448 VIA/Centaur processors and determine their cache parameters
9449 using detect_caches_amd.
9451 2013-05-16 Teresa Johnson <tejohnson@google.com>
9453 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
9454 (rtl_verify_edges): New function.
9455 (rtl_verify_bb_insns): Ditto.
9456 (rtl_verify_bb_pointers): Ditto.
9457 (rtl_verify_bb_insn_chain): Ditto.
9458 (rtl_verify_fallthru): Ditto.
9459 (rtl_verify_bb_layout): Ditto.
9460 (rtl_verify_flow_info_1): Outline checks into new functions.
9461 (rtl_verify_flow_info): Ditto.
9463 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
9465 * cfghooks.c (copy_bbs): Add update_dominance argument.
9466 * cfghooks.h (copy_bbs): Update prototype.
9467 * tree-cfg.c (gimple_duplicate_sese_region):
9468 Add update_dominance argument.
9469 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
9470 * tree-ssa-loop-ch.c (copy_loop_headers): Update
9471 gimple_duplicate_sese_region call.
9472 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
9473 Update copy_bbs call.
9474 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
9475 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
9477 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9479 * tree-vectorizer.h (NUM_PATTERNS): Increment.
9480 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
9481 vect_recog_rotate_pattern.
9482 (vect_recog_rotate_pattern): New function.
9484 2013-05-16 Jason Merrill <jason@redhat.com>
9486 * Makefile.in (LLINKER): New variable.
9487 (mostlyclean): Remove link mutex.
9488 * configure.ac: Handle --enable-link-mutex.
9489 * lock-and-run.sh: New script.
9491 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9494 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
9497 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9499 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
9501 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
9503 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
9504 * config/arm/arm.c (next_consecutive_mem): New function.
9505 (gen_movmem_ldrd_strd): Likewise.
9506 * config/arm/arm.md (movmemqi): Update condition and code.
9507 (unaligned_loaddi, unaligned_storedi): New patterns.
9509 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9511 * config.gcc: Obsolete *-*-solaris2.9*.
9512 * doc/install.texi (Specific, *-*-solaris2*): Document it.
9514 2013-05-16 Richard Biener <rguenther@suse.de>
9516 * passes.c (init_optimization_passes): Move pass_parallelize_loops
9517 earlier, after GRAPHITE transforms and IV canonicalization.
9519 2013-05-16 Jakub Jelinek <jakub@redhat.com>
9521 * omp-low.c (extract_omp_for_data): For collapsed loops,
9522 if at least one of the loops is known at compile time to
9523 iterate zero times, set count to 0.
9524 (expand_omp_regimplify_p): New function.
9525 (expand_omp_for_generic): For collapsed loops, if at least
9526 one of the loops isn't known to iterate at least once,
9527 add runtime check with setting count to 0.
9528 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
9529 For unsigned types if it isn't known at compile time that
9530 the loop will iterate at least once, add runtime check to bypass
9531 the whole loop if initial condition isn't true.
9533 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
9535 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
9537 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
9540 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
9541 transformations to avoid an infinite loop.
9543 2013-05-16 Marek Polacek <polacek@redhat.com>
9545 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
9547 2013-05-15 Leif Ekblad <leif@rdos.net>
9549 * config/i386/i386.c (ix86_decompose_address): Use
9550 DEFAULT_TLS_SEG_REG to access TLS segment register.
9551 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
9552 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
9553 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
9555 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
9558 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
9559 sibling calls to functions that would normally be lazily bound,
9560 unless $gp is call-clobbered.
9562 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
9564 * config/i386/i386.c (ix86_option_override_internal): Update
9565 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
9566 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
9567 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
9568 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
9570 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
9572 2013-05-15 Andreas Schwab <schwab@suse.de>
9574 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
9575 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
9578 2013-05-15 Teresa Johnson <tejohnson@google.com>
9580 * loop-unroll.c (report_unroll_peel): Check decision before
9581 emitting unroll/peel message.
9583 2013-05-15 Teresa Johnson <tejohnson@google.com>
9585 * function.h (has_bb_partition): New rtl_data flag.
9586 (bb_reorder_complete): Ditto.
9587 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
9588 instead of flag_reorder_blocks_and_partition.
9589 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
9590 with some enhancements.
9591 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
9592 * bb-reorder.c (connect_traces): Check for has_bb_partition
9593 instead of flag_reorder_blocks_and_partition.
9594 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
9595 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
9596 verify_hot_cold_block_grouping.
9597 (partition_hot_cold_basic_blocks): Set has_bb_partition.
9599 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
9602 * config/arm/predicates.md (call_insn_operand): New predicate.
9603 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
9604 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
9605 if insn is not a tail call.
9606 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
9608 * config/arm/arm.h (enum reg_class): New caller save register class.
9609 (REG_CLASS_NAMES): Likewise.
9610 (REG_CLASS_CONTENTS): Likewise.
9611 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
9614 2013-05-15 Richard Biener <rguenther@suse.de>
9616 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
9617 of MSG_OPTIMIZED_LOCATIONS.
9618 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
9619 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
9621 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
9622 of MSG_OPTIMIZED_LOCATIONS.
9623 (execute_vect_slp): Likewise.
9624 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
9625 (vect_create_cond_for_alias_checks): Likewise.
9626 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
9627 (vect_recog_widen_mult_pattern): Likewise.
9628 (vect_recog_widen_sum_pattern): Likewise.
9629 (vect_recog_over_widening_pattern): Likewise.
9630 (vect_recog_widen_shift_pattern): Likewise.
9631 (vect_recog_vector_vector_shift_pattern): Likewise.
9632 (vect_recog_divmod_pattern): Likewise.
9633 (vect_recog_mixed_size_cond_pattern): Likewise.
9634 (vect_recog_bool_pattern): Likewise.
9635 (vect_pattern_recog_1): Likewise.
9637 2013-05-15 Martin Jambor <mjambor@suse.cz>
9639 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
9640 non-functions to builtin_unreachable.
9641 * ipa-inline-transform.c (inline_call): Do not assert estimates were
9642 correct when new direct edges were discovered.
9644 2013-05-15 Martin Jambor <mjambor@suse.cz>
9646 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
9647 header, print symbol order instead of node uid, print more information
9648 about indirect edge targets.
9649 (ipa_make_edge_direct_to_target): Print symbol order instead of node
9651 (ipa_make_edge_direct_to_target): Likewise.
9652 (remove_described_reference): Likewise.
9653 (propagate_controlled_uses): Likewise.
9654 (ipa_print_node_params): Also print symbol order.
9655 (ipcp_transform_function): Print symbol order instead of node uids.
9656 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
9657 (cgraph_get_create_real_symbol_node): Likewise.
9658 * ipa-cp.c (print_lattice): Likewise.
9659 (print_all_lattices): Likewise.
9660 (determine_versionability): Likewise.
9661 (initialize_node_lattices): Likewise.
9662 (estimate_local_effects): Likewise.
9663 (update_profiling_info): Likewise.
9664 (create_specialized_node): Likewise.
9665 (perhaps_add_new_callers): Likewise.
9666 (decide_about_value): Likewise.
9667 (decide_whether_version_node): Likewise.
9668 (identify_dead_nodes): Likewise.
9669 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
9670 (dump_inline_summary): Likewise.
9671 (estimate_node_size_and_time): Likewise.
9672 (inline_analyze_function): Likewise.
9673 * ipa-inline.c (report_inline_failed_reason): Likewise.
9674 (want_early_inline_function_p): Likewise.
9675 (edge_badness): Likewise.
9676 (update_edge_key): Likewise.
9677 (inline_small_functions): Likewise. Add dumping of order to two other
9679 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
9680 instead of node uids.
9681 (propagate_pure_const): Likewise.
9682 (propagate_pure_const): Likewise.
9683 * ipa-utils.c (dump_cgraph_node_set): Likewise.
9684 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
9685 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
9687 * tree-pretty-print.c (dump_function_header): Likewise.
9688 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
9689 Print symbol order instead of node uids.
9691 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9693 * config/s390/s390.c (s390_register_move_cost): Don't impose the
9694 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
9696 2013-05-15 Richard Biener <rguenther@suse.de>
9698 PR tree-optimization/57275
9699 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
9700 return value for fail to do runtime alias checks for gather loads.
9702 2013-05-15 Jan Hubicka <jh@suse.cz>
9706 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
9707 weakrefs are not external.
9708 (lto_symtab_merge_decls): Fix thinko when dealing with
9709 non-lto_symtab decls.
9710 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
9711 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
9712 * varpool.c (dump_varpool_node): Dump more flags.
9714 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9716 * config/i386/i386.c (processor_alias_table): Add instruction
9717 FSGSBASE for AMD bdver3 architecture.
9719 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9721 * tree.c (warn_deprecated_use): Print file:line using locus color.
9722 * diagnostic.c (diagnostic_report_current_module): Print file:line
9723 and file:line:column using locus color.
9725 2013-05-14 Mike Stump <mikestump@comcast.net>
9727 * gdbinit.in: Add __null.
9729 2013-05-14 Mike Stump <mikestump@comcast.net>
9731 * recog.h: Rename struct recog_data to Recog_data.
9732 * recog.c: Likewise.
9733 * reload.c (can_reload_into): Likewise.
9734 * config/picochip/picochip.c: Likewise.
9736 2013-05-14 Mike Stump <mikestump@comcast.net>
9738 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
9740 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
9742 * resource.h (struct resources): Remove unch_memory member.
9743 (CLEAR_RESOURCE): Don't clear unch_memory.
9744 * resource.c (mark_referenced_resources): Don't set it.
9745 (mark_set_resources): Likewise.
9746 (mark_target_live_regs): Don't clear it.
9747 (init_resource_info): Likewise.
9748 * reorg.c (resource_conflicts_p): Don't compare it.
9749 (redundant_insn): Don't set it.
9751 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
9753 * emit-rtl.c (next_label): Remove unused function.
9754 (skip_consecutive_labels, link_cc0_insns): Move to ...
9755 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
9756 only place where these functions are used, and make them static.
9758 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9760 * fold-const.c (fold_negate_expr): Handle vectors.
9761 (fold_truth_not_expr): Make it static.
9762 (fold_invert_truthvalue): New static function.
9763 (invert_truthvalue_loc): Handle vectors. Do not call
9764 fold_truth_not_expr directly.
9765 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
9766 <TRUTH_NOT_EXPR>: Do not cast to boolean.
9767 (fold_comparison): Handle vector constants.
9768 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
9769 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
9770 * tree.h (fold_truth_not_expr): Remove declaration.
9772 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
9774 * config/aarch64/aarch64-simd.md
9775 (aarch64_vcond_internal<mode>): Rename to...
9776 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
9777 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
9778 float modes. Clarify all iterator modes.
9779 (vcond<mode><mode>): Use new name for vcond expanders.
9780 (vcond<v_cmp_result><mode>): Likewise.
9781 (vcondu<mode><mode>: Likewise.
9782 * config/aarch64/iterators.md (VDQF_COND): New.
9784 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
9787 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
9788 variable for the shift amount. Check that we shift by non-negative
9791 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
9794 * config/arm/arm.h (EPILOGUE_USES): Only return true
9795 for LR_REGNUM after epilogue_completed.
9797 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
9799 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
9802 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9805 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
9806 and Solaris 11+/x86 with gld.
9807 * configure: Regenerate.
9809 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9811 * expmed.c (expand_shift_1): Canonicalize rotates by
9812 constant bitsize / 2 to bitsize - 1.
9813 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
9814 case ROTATERT>: Likewise.
9817 2013-05-10 Jakub Jelinek <jakub@redhat.com>
9819 * config/i386/i386.md (rotateinv): New code attr.
9820 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
9821 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
9822 roll $31, %eax, etc.
9824 2013-05-14 Richard Biener <rguenther@suse.de>
9827 * tree-eh.c (sink_clobbers): Give up for successors with
9828 multiple predecessors and no virtual uses.
9830 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
9832 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
9833 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
9835 2013-05-14 Jakub Jelinek <jakub@redhat.com>
9838 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
9839 the case when both op0 and op1 have VOIDmode.
9841 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9843 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
9844 in multiply-accumulate mode.
9846 2013-05-13 Guozhi Wei <carrot@google.com>
9848 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
9850 2013-05-13 Kai Tietz <ktietz@redhat.com>
9853 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
9854 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
9855 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
9856 * config/i386/i386.c (ix86_option_override_internal): Likewise.
9857 (ix86_expand_prologue): Likewise.
9858 (ix86_expand_split_stack_prologue): Likewise.
9859 (legitimate_pic_address_disp_p): Likewise.
9860 (legitimize_pic_address): Likewise.
9861 (legitimize_tls_address): Likewise.
9862 (legitimize_pe_coff_symbol): Likewise.
9863 (output_pic_addr_const): Likewise.
9864 (construct_plt_address): Likewise.
9865 (ix86_expand_call): Likewise.
9866 (x86_output_mi_thunk): Likewise.
9867 (x86_function_profiler): Likewise.
9869 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
9871 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
9872 similar switch cases.
9873 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
9874 (aarch64_simd_mov_to_<mode>low): Delete.
9875 (aarch64_simd_mov_to_<mode>high): Delete.
9876 (move_lo_quad_<mode>): Add w<-r alternative.
9877 (aarch64_simd_move_hi_quad_<mode>): Likewise.
9878 (aarch64_simd_mov_from_*): Update type attribute.
9879 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
9882 2013-05-13 Jan Hubicka <jh@suse.cz>
9884 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
9885 * config/i386/i386.c (ix86_compute_frame_layout,
9886 ix86_expand_epilogue, emit_i387_cw_initialization,
9887 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
9888 ix86_local_alignment): Fix use of size/speed predicates.
9890 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9892 PR tree-optimization/45216
9893 PR tree-optimization/57157
9894 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
9895 the (-Y) & (B - 1) variant if OP is |.
9896 * expmed.c (expand_shift_1): For rotations by const0_rtx just
9897 return shifted. Use (-op1) & (prec - 1) as other_amount
9898 instead of prec - op1.
9900 2013-05-13 Martin Jambor <mjambor@suse.cz>
9903 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
9904 (ipa_constant_data): New type.
9905 (ipa_jump_func): Use ipa_constant_data to hold information about
9906 constant jump functions.
9907 (ipa_get_jf_constant): Adjust to jump function type changes.
9908 (ipa_get_jf_constant_rdesc): New function.
9909 (ipa_param_descriptor): New field controlled_uses.
9910 (ipa_get_controlled_uses): New function.
9911 (ipa_set_controlled_uses): Likewise.
9912 * ipa-ref.h (ipa_find_reference): Declare.
9913 * ipa-prop.c (ipa_cst_ref_desc): New type.
9914 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
9916 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
9917 New parameter cs. Adjust all callers.
9918 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
9919 (remove_described_reference): New function.
9920 (jfunc_rdesc_usable): Likewise.
9921 (try_make_edge_direct_simple_call): Decrement controlled use count,
9922 attempt to remove reference if it hits zero.
9923 (combine_controlled_uses_counters): New function.
9924 (propagate_controlled_uses): Likewise.
9925 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
9926 (ipa_edge_duplication_hook): Duplicate reference descriptions.
9927 (ipa_print_node_params): Print described use counter.
9928 (ipa_write_jump_function): Adjust to jump function type changes.
9929 (ipa_read_jump_function): New parameter CS, pass it to
9930 ipa_set_jf_constant. Adjust caller.
9931 (ipa_write_node_info): Stream controlled use count
9932 (ipa_read_node_info): Likewise.
9933 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
9935 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
9936 count. Remove cloning-added reference if it reaches zero.
9937 * ipa-ref.c (ipa_find_reference): New function.
9939 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
9941 * config/i386/i386.c (processor_target_table): Modified default
9942 alignment values for AMD BD and BT architectures.
9944 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
9946 * tree-vect-generic.c (uniform_vector_p): Move ...
9947 * tree.c (uniform_vector_p): ... here.
9948 * tree.h (uniform_vector_p): Declare it.
9949 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
9952 2013-05-13 Jakub Jelinek <jakub@redhat.com>
9954 PR tree-optimization/57230
9955 * tree-ssa-strlen.c (handle_char_store): Record length for
9956 array store from STRING_CST.
9958 PR tree-optimization/57230
9959 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
9962 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
9964 * config/epiphany/epiphany.c (epiphany_init): Check size of
9965 NUM_MODES_FOR_MODE_SWITCHING.
9966 (epiphany_expand_prologue):
9967 Remove CONFIG_REGNUM initial value handling code.
9968 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
9969 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
9970 (emit_set_fp_mode, epiphany_mode_after): Likewise.
9971 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
9972 Don't return 1 for FP_MODE_NONE.
9973 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
9974 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
9975 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
9976 * config/epiphany/epiphany.md (save_config): New pattern.
9978 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
9980 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
9982 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
9984 * config/i386/i386.md (memory): Handle sseishft1.
9985 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
9986 (*vec_extractv2di_1): Ditto.
9988 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
9990 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
9993 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
9995 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
9996 Add mthumb/march=armv7-a multilib.
9997 Add mthumb/march=armv7-r multilib.
9998 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
10000 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
10002 * config/v850/t-rtems: Add more multilibs.
10004 2013-05-10 Richard Biener <rguenther@suse.de>
10006 PR tree-optimization/57214
10007 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
10008 not propagate from SSA names that occur in abnormal PHI nodes.
10010 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
10012 * stor-layout.c (element_precision): New function.
10013 * machmode.h (element_precision): Declare it.
10014 * tree.c (build_minus_one_cst): New function.
10015 (element_precision): Likewise.
10016 * tree.h (build_minus_one_cst): Declare new function.
10017 (element_precision): Likewise.
10018 * fold-const.c (operand_equal_p): Use element_precision.
10019 (fold_binary_loc): Handle vector types.
10020 * convert.c (convert_to_integer): Use element_precision.
10021 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
10024 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
10026 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
10027 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
10028 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
10029 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10030 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
10031 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
10032 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
10033 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
10034 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
10035 (Uuw8): New constraints.
10036 (Usb4): Move into alphabetical order.
10037 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
10038 (sd8_operand, ub8_operand, uw8_operand): New predicates.
10039 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
10040 previously unnamed patterns.
10041 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
10042 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
10043 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
10044 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
10045 of set_attr_alternative/if_then_else. Use extended_mips16 instead
10046 of specific lengths.
10048 2013-05-10 Jakub Jelinek <jakub@redhat.com>
10050 * config/i386/i386.md (rotateinv): New code attr.
10051 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
10052 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
10053 roll $31, %eax, etc.
10055 PR tree-optimization/45216
10056 PR tree-optimization/57157
10057 * tree-ssa-forwprop.c (simplify_rotate): New function.
10058 (ssa_forward_propagate_and_combine): Call it.
10060 2013-05-10 Richard Biener <rguenther@suse.de>
10062 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
10063 disable peeling when we version for aliasing.
10064 (vector_alignment_reachable_p): Honor explicit user alignment.
10065 (vect_supportable_dr_alignment): Likewise.
10066 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
10067 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
10068 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
10069 then peeling to arrange for the cost-model check to come first.
10071 2013-05-10 Alan Modra <amodra@gmail.com>
10073 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
10074 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
10075 * configure: Regenerate.
10077 2013-05-10 Alan Modra <amodra@gmail.com>
10080 * varasm.c (default_elf_select_section): Move !DECL_P check..
10081 (get_named_section): ..to here before calling get_section_name.
10083 (default_section_type_flags): Add DECL_P check.
10084 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
10085 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
10087 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
10089 * config/epiphany/epiphany.c (epiphany_expand_prologue):
10090 When using gen_stack_adjust_str with a register offset, add a
10091 REG_FRAME_RELATED_EXPR note.
10093 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
10095 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
10096 (*vec_extractv4si_zext_mem): Ditto.
10097 (*vec_extractv2di): Add 0->x and x->x alternatives.
10098 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
10099 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
10101 2013-05-09 Jason Merrill <jason@redhat.com>
10103 N3639 C++1y VLA support
10104 * gimplify.c (gimplify_vla_decl): Don't touch an existing
10107 * tree.c (build_constructor_va): New.
10108 * tree.h: Declare it.
10110 2013-05-09 Martin Jambor <mjambor@suse.cz>
10113 * gimple-fold.c (canonicalize_constructor_val): Call
10114 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
10116 2013-05-09 Jan Hubicka <jh@suse.cz>
10117 Richard Biener <rguenther@suse.de>
10120 * symtab.c (symtab_make_decl_local): Do not add private names.
10122 2013-05-09 Jan Hubicka <jh@suse.cz>
10125 * symtab.c (insert_to_assembler_name_hash): Handle clones.
10126 (unlink_from_assembler_name_hash): Likewise.
10127 (symtab_prevail_in_asm_name_hash, symtab_register_node,
10128 symtab_unregister_node, symtab_initialize_asm_name_hash,
10129 change_decl_assembler_name): Update.
10131 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
10133 * config/aarch64/aarch64.md: New movtf split.
10134 (*movtf_aarch64): Update.
10135 (aarch64_movdi_tilow): Handle TF modes and rename to
10136 aarch64_movdi_<mode>low.
10137 (aarch64_movdi_tihigh): Handle TF modes and rename to
10138 aarch64_movdi_<mode>high
10139 (aarch64_movtihigh_di): Handle TF modes and rename to
10140 aarch64_mov<mode>high_di
10141 (aarch64_movtilow_di): Handle TF modes and rename to
10142 aarch64_mov<mode>low_di
10143 (aarch64_movtilow_tilow): Remove spurious whitespace.
10144 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
10146 (aarch64_print_operand): Update.
10148 2013-05-09 Alan Modra <amodra@gmail.com>
10150 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
10152 * configure: Regenerate.
10154 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10156 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
10157 splitter preparation statements.
10158 * config/i386/sse.md (*vec_extract* splitters): Ditto.
10159 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
10162 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10164 * gimple-ssa-strength-reduction.c (count_candidates): Change
10165 return value to int.
10166 (analyze_candidates_and_replace): Change type of length to int.
10168 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
10170 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
10171 (*vec_extract<mode>): Use VI12_128 mode iterator.
10172 (*vec_extract<mode>_mem): Ditto.
10173 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
10176 2013-05-08 Diego Novillo <dnovillo@google.com>
10181 2012-08-17 Diego Novillo <dnovillo@google.com>
10184 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
10185 * config.in: Regenerate.
10186 * configure: Regenerate.
10187 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
10189 2013-05-08 Jan Hubicka <jh@suse.cz>
10192 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
10193 * cgraph.h (symtab_node_base): Add unique_name.
10194 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10195 input_overwrite_node, input_varpool_node): Stream unique_name.
10196 * cgraphclones.c (cgraph_create_virtual_clone,
10197 cgraph_function_versioning): Set unique_name.
10198 * ipa.c (function_and_variable_visibility): Set unique_name.
10200 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10202 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
10203 (alloc_cand_and_find_basis): Restrict conditional candidate
10204 processing to CAND_MULTs.
10206 2013-05-08 Jan Hubicka <jh@suse.cz>
10209 lto-symtab.c (lto_symtab_symbol_p): New function.
10210 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
10211 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
10212 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
10213 Skip static symbols.
10215 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
10217 PR tree-optimization/57200
10218 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
10219 Only call inform if the preceding warning_at returns true.
10221 2013-05-07 Han Shen <shenhan@google.com>
10223 * cfgexpand.c (record_or_union_type_has_array_p): New function.
10224 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
10225 * common.opt (fstack-protector-strong): New option.
10226 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
10227 * doc/invoke.texi (Optimization Options): Document
10228 "-fstack-protector-strong".
10229 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
10231 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10233 * config/mips/mips.c (mips_machine_reorg2): Return 0.
10235 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
10237 * ira.c (update_equiv_regs): Add insn having equiv memory even if
10238 it is not lhs of the insn.
10239 (setup_reg_equiv): Remove insn having equiv memory which it is not
10241 * lra-constraints.c (process_address): Try to improve generation
10242 code for address base + disp.
10243 (lra_constraints): Make correct the code for checking insn setting
10244 up backward equivalence. Remove insn only if it is in the init
10246 * lra-eliminations.c (update_reg_eliminate): Change return value.
10247 (lra_eliminate): Use the result.
10249 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
10251 * config/i386/sse.md (ssescalarnummask): New mode attribute.
10252 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
10253 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
10254 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
10255 register target operands.
10256 (*vec_extractv8hi_sse2): New pattern.
10257 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
10258 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
10259 (*vec_extract<mode>_mem): New insn and split pattern.
10261 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
10263 * config/arm/arm.c (arm_asan_shadow_offset): New function.
10264 (TARGET_ASAN_SHADOW_OFFSET): Define.
10265 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
10266 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
10268 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10270 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
10271 (incr_vec_index): Return -1 if increment not found.
10272 (create_add_on_incoming_edge): Assert if increment not found.
10273 (record_increment): Limit number of increments recorded.
10274 (all_phi_incrs_profitable): Return false if an increment not found.
10275 (replace_profitable_candidates): Don't process increments that were
10277 (analyze_candidates_and_replace): Limit size of incr_vec.
10279 2013-05-07 Richard Biener <rguenther@suse.de>
10281 * calls.c (special_function_p): setjmp-like functions are leaf.
10282 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
10283 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
10285 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
10287 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
10288 (aarch64_simd_mov<mode>): New expander.
10289 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
10290 (aarch64_simd_mov_to_<mode>high): Likewise.
10291 (aarch64_simd_mov_from_<mode>low): Likewise.
10292 (aarch64_simd_mov_from_<mode>high): Likewise.
10293 (aarch64_dup_lane<mode>): Update.
10294 (aarch64_dup_lanedi): New instruction pattern.
10295 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
10296 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
10298 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10300 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
10301 (replace_mult_candidate): Remove unnecessary argument; remove
10302 unnecessary parameter from call to introduce_cast_before_cand.
10303 (replace_unconditional_candidate): Remove unnecessary parameter
10304 from call to replace_mult_candidate.
10305 (replace_conditional_candidate): Likewise.
10306 (insert_initializers): Use make_temp_ssa_name.
10307 (introduce_cast_before_cand): Remove unnecessary argument; use
10308 make_temp_ssa_name.
10309 (replace_one_candidate): Remove unnecessary argument; remove
10310 unnecessary parameter from calls to introduce_cast_before_cand.
10311 (replace_profitable_candidates): Remove unnecessary parameters
10312 from calls to replace_one_candidate.
10314 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10316 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
10317 phi def as possibly hiding a basis for a CAND_ADD whose operands
10318 have been commuted in the analysis.
10319 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
10321 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
10323 * config/aarch64/aarch64.md
10324 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
10325 shift value between 0-4.
10327 2013-05-07 Richard Biener <rguenther@suse.de>
10329 * double-int.h (rshift): New overload.
10330 * double-int.c (rshift): New function.
10331 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
10332 (create_reference_ops_from_ref): Remove.
10333 (vn_reference_insert): Use shared ops for constructing the
10334 reference and copy it.
10336 2013-05-07 Richard Biener <rguenther@suse.de>
10338 PR middle-end/57190
10339 * tree-eh.c (sink_clobbers): Properly propagate
10340 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
10342 2013-05-07 Jakub Jelinek <jakub@redhat.com>
10344 PR tree-optimization/57149
10345 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
10346 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
10347 collect_phi_def_edges, execute_late_warn_uninitialized): Use
10348 uninit_undefined_value_p instead of ssa_undefined_value_p.
10351 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
10352 for modifier == EXPAND_INITIALIZER.
10354 2013-05-07 Anton Blanchard <anton@samba.org>
10356 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
10357 for powerpc64 little endian.
10358 * configure: Regenerate.
10360 2013-05-06 Graham Stott <grahams@btinternet.com>
10362 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
10363 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
10364 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
10365 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
10367 2013-05-06 Graham Stott <grahams@btinternet.com>
10369 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
10370 codes which allow non-lvalues.
10372 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
10374 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
10375 components are all 1s.
10376 (integer_minus_onep): New function.
10377 * tree.h (integer_minus_onep): Declare it.
10378 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
10379 integer_minus_onep instead of integer_all_onesp.
10381 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10384 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
10385 variations of these patterns.
10387 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10389 * config/i386/i386.md (isa): Add x64_sse4 member.
10390 (enabled): Handle x64_sse4.
10391 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
10392 instruction for 64bit SSE4_1 targets. Update insn attributes.
10393 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
10394 instruction for SSE4_1 targets. Update insn attributes.
10395 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10396 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
10397 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
10399 (*vec_extractv4si): Rename from *sse4_1_pextrd.
10400 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
10401 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
10403 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10406 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
10408 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
10410 * final.c (do_assembler_dialects): Don't handle curly braces and
10411 vertical bar escaped by % as dialect delimiters.
10412 (output_asm_insn): Print curly braces and vertical bar if escaped
10413 by % and ASSEMBLER_DIALECT defined.
10414 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
10415 * doc/tm.texi: Regenerated.
10417 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
10419 * config/mips/mips.c: Include tree-pass.h.
10420 (mips_reorg): Split in pre- and post-dbr_schedule parts.
10421 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
10422 (pass_mips_machine_reorg2): New machine specific pass.
10423 (insert_pass_mips_machine_reorg2): New pass plugin definition.
10424 (mips_option_override): Register the new pass.
10425 * rtl.h (cleanup_barriers): Remove prototype.
10426 (dbr_schedule): Likewise.
10427 * jump.c (cleanup_barriers): Make static.
10428 * reorg.c (dbr_schedule): Likewise.
10430 2013-05-06 Richard Biener <rguenther@suse.de>
10432 PR tree-optimization/57185
10433 * tree-parloops.c (add_field_for_reduction): Handle anonymous
10434 SSA names properly.
10436 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
10439 * config/i386/i386.c (add_parameter_dependencies): Add dependence
10440 between "first_arg" and "insn", not "last" and "insn".
10442 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10444 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
10445 (find_candidates_in_block): Re-enable slsr_process_phi.
10446 (create_phi_basis): Fix double counting of candidate adjustment.
10448 2013-05-06 Richard Biener <rguenther@suse.de>
10450 PR middle-end/57147
10451 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
10452 the edge is also fallthru, preserve it and just clear the
10454 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
10455 also complex, preserve that and just clear the fallthru flag.
10456 * tree-inline.c (update_ssa_across_abnormal_edges): Also
10457 update virtual operands.
10459 2013-05-06 Alan Modra <amodra@gmail.com>
10461 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
10462 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
10463 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
10464 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
10465 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10466 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
10468 2013-05-06 Alan Modra <amodra@gmail.com>
10470 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
10471 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
10472 (DEFAULT_ASM_ENDIAN): Define.
10473 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
10474 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
10475 Update -K PIC clause from sysv4.h.
10476 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
10477 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
10479 2013-05-06 Alan Modra <amodra@gmail.com>
10481 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
10482 twice for little-endian.
10483 (ashrdi3_no_power, ashrdi3): Support little-endian.
10485 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
10488 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
10489 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
10491 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
10492 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
10493 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
10496 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
10497 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
10499 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
10500 * config/pa/pa.opt: Make mbig-switch a no-op.
10501 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
10502 (CASE_VECTOR_MODE): Always return SImode.
10503 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
10504 for the !TARGET_BIG_SWITCH case.
10505 * config/pa/pa-linux.h: Likewise.
10506 * config/pa/pa-openbsd.h: Likewise.
10507 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
10508 * config/pa/pa.md (short_jump): Remove define_insn.
10509 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
10510 (casesi0): Remove define_insn.
10511 (type): Remove btable_branch.
10512 (pa_combine_type): Likewise.
10513 (in_nullified_branch_delay): Likewise.
10514 (in_call_delay): Likewise.
10515 (define_delay): Likewise.
10516 (define_insn_reservation "Z3"): Likewise.
10517 (define_insn_reservation "Z4"): Likewise.
10518 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
10519 (pa_adjust_insn_length): Remove adjustment for btable branches.
10520 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
10523 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
10525 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
10526 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
10527 Add m->r,x alternatives.
10528 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
10529 splitters using SWI48x mode iterator.
10530 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
10531 TARGET_64BIT. Add m->x alternative.
10532 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
10533 Add o->x alternative. Enable for TARGET_SSE.
10534 (sse_storeq): Remove expander.
10535 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
10536 with memory input operand.
10537 (*vec_extractv2di_1 splitter): New.
10538 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
10539 * config/i386/i386.md (ssevecmodelower): New mode attribute.
10541 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10543 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
10544 (INT_LOWPART): Delete.
10545 (extract_MB): Adjust.
10546 (extract_ME): Adjust.
10547 (print_operand): Adjust.
10549 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
10551 * config/rs6000/predicates.md (reg_or_add_cint_operand,
10552 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10553 (reg_or_logical_cint_operand, easy_fp_constant,
10554 logical_const_operand): Delete "CONST_DOUBLE" case.
10555 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
10556 "HOST_BITS_PER_WIDE_INT == 64" test.
10557 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
10558 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
10559 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
10560 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
10562 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
10563 CONST_DOUBLE DImode/VOIDmode case.
10564 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
10565 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10566 CONST_DOUBLE VOIDmode case.
10567 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
10568 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
10569 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
10570 Delete CONST_DOUBLE case.
10571 (splitters for mov FMOVE64 const_double): Delete
10572 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
10573 "HOST_BITS_PER_WIDE_INT >= 64" test.
10574 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
10576 (mov DI const_double): Delete.
10578 2013-05-04 Jakub Jelinek <jakub@redhat.com>
10580 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
10581 on op shows all bits zero in mode of a lowpart subreg, return zero.
10583 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
10586 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
10587 to save TFmode registers and DImode to save TImode registers for
10588 caller save operations.
10589 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
10590 mark being partially clobbered since they only use the first
10593 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
10594 and TDmode only use the upper 64-bits of each VSX register.
10596 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10598 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
10599 (find_candidates_in_block): Disable slsr_process_phi.
10601 2013-05-03 Guozhi Wei <carrot@google.com>
10603 * coverage.c (coverage_obj_init): Move the construction of gcov
10605 (build_init_ctor): ... here.
10607 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10609 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
10610 (slsr_cand_d): Redefine def_phi.
10611 (stride_status, phi_adjust_status, count_phis_status): New enums.
10612 (find_phi_def): New.
10613 (find_basis_for_base_expr): New.
10614 (find_basis_for_candidate): Handle hidden bases.
10615 (alloc_cand_and_find_basis): Handle phi candidates.
10616 (slsr_process_phi): New.
10617 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
10618 (create_mul_imm_cand): Likewise.
10619 (create_add_ssa_cand): Exclude phi base candidates.
10620 (create_add_imm_cand): Likewise.
10621 (slsr_process_cast): Likewise.
10622 (slsr_process_copy): Likewise.
10623 (find_candidates_in_block): Handle phi candidates.
10624 (dump_candidate): Likewise.
10625 (unconditional_cands): Delete.
10626 (unconditional_cands_with_known_stride_p): Delete.
10627 (phi_dependent_cand_p): New.
10628 (cand_increment): Handle phi-dependent candidates.
10629 (replace_dependent): Delete.
10630 (replace_mult_candidate): New.
10631 (replace_unconditional_candidate): New.
10632 (incr_vec_index): Move to avoid forward reference.
10633 (create_add_on_incoming_edge): New.
10634 (create_phi_basis): New.
10635 (replace_dependents): Delete.
10636 (replace_conditional_candidate): New.
10637 (phi_add_costs): New.
10638 (replace_uncond_cands_and_profitable_phis): New.
10639 (record_increment): Handle phi adjustments.
10640 (record_phi_increments): New.
10641 (record_increments): Handle phi adjustments.
10642 (phi_incr_cost): New.
10643 (lowest_cost_path): Handle phis.
10644 (total_savings): Likewise.
10645 (analyze_increments): Likewise.
10646 (ncd_with_phi): New.
10647 (ncd_of_cand_and_phis): New.
10648 (nearest_common_dominator_for_cands): Handle phi increments.
10649 (all_phi_incrs_profitable): New.
10650 (replace_profitable_candidates): Handle phi-dependent candidates.
10651 (analyze_candidates_and_replace): Likewise.
10653 2013-05-03 Teresa Johnson <tejohnson@google.com>
10656 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
10657 do not exceed REG_BR_PROB_BASE.
10659 2013-05-03 Jeff Law <law@redhat.com>
10661 PR tree-optimization/57144
10662 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
10663 operand of the condition will bit into the new type when eliminating
10664 a cast feeding a condition.
10666 2013-05-03 Jakub Jelinek <jakub@redhat.com>
10668 PR rtl-optimization/57130
10669 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
10670 of COMPARE as in_code to the recursive call if needed.
10672 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
10674 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
10675 (enabled): Handle new members.
10676 * config/i386/sse.md (*vec_concatv2si): Merge from
10677 *vec_concatv2si_sse2 and vec_concatv2si_sse.
10678 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
10680 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
10682 PR tree-optimization/57027
10683 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
10684 for fnms opportunity, check we got the prerequisite kind
10685 of tree / gimple before using accessor functions.
10687 2013-05-03 Richard Biener <rguenther@suse.de>
10689 * double-int.h (lshift): New overload without precision
10690 and arith argument.
10691 (operator *=, operator +=, operator -=): Move ...
10692 * double-int.c (operator *=, operator +=, operator -=): ... here
10693 and implement more efficiently.
10694 (mul_double_with_sign): Remove.
10695 (lshift_double): Adjust to take unsinged shift argument, push
10696 dispatching code to callers.
10697 (mul_double_wide_with_sign): Add early out for callers that
10698 are not interested in high parts or overflow.
10699 (lshift): New function.
10700 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
10701 dispatch code here.
10702 (lrotate, rrotate): Use logical shifts.
10703 * expr.c (get_inner_reference): Use lshift.
10704 * fixed-value.c (do_fixed_divide): Likewise.
10705 * tree-dfa.c (get_ref_base_and_extent): Likewise.
10706 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
10707 (indirect_refs_may_alias_p): Likewise.
10708 (stmt_kills_ref_p_1): Likewise.
10710 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10712 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
10714 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
10716 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
10717 scalar form of FABD instruction.
10719 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
10721 * lra-constraints.c (process_alt_operands): Add checking alt
10722 number to choose the best alternative.
10724 2013-05-02 Richard Biener <rguenther@suse.de>
10726 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
10727 bitmap and its handling.
10728 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
10730 2013-05-02 Richard Biener <rguenther@suse.de>
10732 PR middle-end/57140
10733 * tree-inline.c (copy_loops): Properly handle removed loops.
10734 (copy_cfg_body): Mark destination loops for fixup if source
10735 loops needed fixup.
10737 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
10740 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
10741 generating simple_return for naked functions.
10743 2013-05-02 Martin Jambor <mjambor@suse.cz>
10745 PR middle-end/56988
10746 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
10747 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
10749 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
10750 ipa_agg_replacement_value structures.
10751 (known_aggs_to_agg_replacement_list): Likewise.
10752 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
10753 (read_agg_replacement_chain): Likewise.
10754 (ipcp_transform_function): Also check that by_ref flags match.
10756 2013-05-02 Richard Biener <rguenther@suse.de>
10758 * graphds.h (struct graph): Add obstack member.
10759 * graphds.c (new_graph): Initialize obstack and allocate
10761 (add_edge): Allocate edge from the obstack.
10762 (free_graph): Free the obstack instead of all edges and vertices.
10764 2013-05-02 Teresa Johnson <tejohnson@google.com>
10766 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
10768 * cfg.c (update_bb_profile_for_threading): Ditto.
10769 * tree-inline.c (copy_bb): Ditto.
10770 (copy_edges_for_bb): Ditto.
10771 (initialize_cfun): Ditto.
10772 (copy_cfg_body): Ditto.
10773 (expand_call_inline): Ditto.
10774 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
10775 (estimate_node_size_and_time): Ditto.
10776 (inline_merge_summary): Ditto.
10777 * cgraphclones.c (cgraph_clone_edge): Ditto.
10778 (cgraph_clone_node): Ditto.
10779 * sched-rgn.c (compute_dom_prob_ps): Ditto.
10780 (compute_trg_info): Ditto.
10782 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10784 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
10785 S reg when fp attribute set.
10786 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
10788 2013-05-02 Ian Bolton <ian.bolton@arm.com>
10790 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
10792 (*and_one_cmplsi3_compare0_uxtw): Likewise.
10793 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
10794 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
10796 2013-05-02 Richard Biener <rguenther@suse.de>
10798 * tree-scalar-evolution.c (scev_info_hasher): Remove.
10799 (struct instantiate_cache_entry): New type.
10800 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
10801 (struct instantiate_cache_type): New type.
10802 (set_instantiated_value, get_instantiated_value): Remove.
10803 (get_instantiated_value_entry): New function.
10804 (instantiate_scev_name): Use the new cache and adjust.
10805 (instantiate_scev_poly): Adjust.
10806 (instantiate_scev_binary): Likewise.
10807 (instantiate_array_ref): Likewise.
10808 (instantiate_scev_convert): Likewise.
10809 (instantiate_scev_not): Likewise.
10810 (instantiate_scev_3): Likewise.
10811 (instantiate_scev_2): Likewise.
10812 (instantiate_scev_r): Likewise.
10813 (instantiate_scev): Likewise.
10814 (resolve_mixers): Likewise.
10816 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
10819 * lra-constraints.c (best_small_class_operands_num): Remove.
10820 (process_alt_operands): Remove small_class_operands_num. Take
10821 small classes operands into losers and only if the operand is not
10822 matched. Modify debugging output.
10823 (curr_insn_transform): Remove best_small_class_operands_num.
10826 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10828 * config/aarch64/aarch64-builtins.c
10829 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
10830 * config/aarch64/aarch64-simd-builtins.def
10831 (reduc_splus_): Add new modes.
10832 (reduc_uplus_): New.
10833 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
10834 (reduc_uplus_v4sf): Likewise.
10835 (reduc_splus_v4sf): Likewise.
10836 (aarch64_addv<mode>): Likewise.
10837 (reduc_uplus_<mode>): Likewise.
10838 (reduc_splus_<mode>): Likewise.
10839 (aarch64_addvv2di): Likewise.
10840 (reduc_uplus_v2di): Likewise.
10841 (reduc_splus_v2di): Likewise.
10842 (aarch64_addvv2si): Likewise.
10843 (reduc_uplus_v2si): Likewise.
10844 (reduc_splus_v2si): Likewise.
10845 (reduc_<sur>plus_<mode>): New.
10846 (reduc_<sur>plus_v2di): Likewise.
10847 (reduc_<sur>plus_v2si): Likewise.
10848 (reduc_<sur>plus_v4sf): Likewise.
10849 (aarch64_addpv4sf): Likewise.
10850 * config/aarch64/arm_neon.h
10851 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
10852 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
10853 add UNSPEC_SADDV, UNSPEC_UADDV.
10855 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
10857 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10859 * config/aarch64/arm_neon.h
10860 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
10862 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10864 * config/aarch64/aarch64-builtins
10865 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
10867 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10869 * config/aarch64/aarch64-simd-builtins.def
10870 (reduc_smax_): New.
10871 (reduc_smin_): Likewise.
10872 (reduc_umax_): Likewise.
10873 (reduc_umin_): Likewise.
10874 (reduc_smax_nan_): Likewise.
10875 (reduc_smin_nan_): Likewise.
10878 (smax): Update for V2SF, V4SF and V2DF modes.
10881 (smin_nan): Likewise.
10882 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
10883 (<su><maxmin><mode>3): ...This, refactor.
10884 (s<maxmin><mode>3): New.
10885 (<maxmin_uns><mode>3): Likewise.
10886 (reduc_<maxmin_uns>_<mode>): Refactor.
10887 (reduc_<maxmin_uns>_v4sf): Likewise.
10888 (reduc_<maxmin_uns>_v2si): Likewise.
10889 (aarch64_<fmaxmin><mode>: Remove.
10890 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
10892 (vmin<q>_f<32,64>): Likewise.
10893 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10895 (su): Add mappings for smax, smin, umax, umin.
10897 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
10898 (FMAXMIN): Rename as...
10899 (FMAXMIN_UNS): ...This.
10901 (fmaxminv): Likewise.
10902 (fmaxmin): Likewise.
10904 (maxmin_uns_op): Likewise.
10906 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10908 * config/aarch64/arm_neon.h
10909 (vac<ge, gt><sd>_f<32, 64>): Rename to...
10910 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
10911 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
10913 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10915 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
10916 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
10918 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10920 * config/aarch64/aarch64-simd.md
10921 (vcond<mode>_internal): Handle special cases for constant masks.
10922 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
10923 (vcondu<mode><mode>): Likewise.
10924 (vcond<v_cmp_result><mode>): New.
10926 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10928 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
10929 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
10930 * config/aarch64/aarch64-simd-builtins.def
10931 (cmeq): Update to BUILTIN_VALLDI.
10936 * config/aarch64/arm_neon.h
10937 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
10938 to builtins or C as appropriate.
10940 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
10942 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
10944 (cmhi): Rename to...
10946 * config/aarch64/aarch64-simd.md
10947 (simd_mode): Add SF.
10948 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
10949 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
10950 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
10951 (cstore<mode>_neg): ...This.
10952 * config/aarch64/iterators.md
10954 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
10955 (COMPARISONS): New.
10956 (UCOMPARISONS): Likewise.
10957 (optab): Add missing comparisons.
10963 (VCMP_S): Likewise.
10964 (VCMP_U): Likewise.
10965 (V_cmp_result): Add DF, SF modes.
10966 (v_cmp_result): Likewise.
10968 (vmtype): Likewise.
10969 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
10971 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
10973 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
10974 define_insn to define_insn_and_split.
10975 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
10976 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
10977 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
10978 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
10979 (thumb2_negscc): Likewise.
10981 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10983 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
10985 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
10987 * config/arm/thumb2.md: Remove trailing whitespaces.
10989 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
10991 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
10992 Use gen_int_mode rather than GEN_INT.
10994 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
10996 * value-prof.c (stream_in_histogram_value): Remove the strayed
10999 2013-04-30 Richard Biener <rguenther@suse.de>
11001 PR middle-end/57122
11002 * cfghooks.c (split_edge): Properly check for the loop latch edge.
11004 2013-04-30 Richard Biener <rguenther@suse.de>
11006 PR middle-end/57107
11007 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
11009 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
11011 PR rtl-optimization/56957
11012 PR rtl-optimization/57105
11013 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
11014 variable. Use just INSN_UID for determining whether an insn
11015 should be only disconnected from the insn stream.
11016 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
11018 2013-04-30 Jakub Jelinek <jakub@redhat.com>
11020 PR tree-optimization/57104
11021 * tsan.c (instrument_expr): Don't instrument accesses to
11022 DECL_HARD_REGISTER VAR_DECLs.
11024 2013-04-30 Richard Biener <rguenther@suse.de>
11026 * function.h (loops_for_fn): New inline function.
11027 (set_loops_for_fn): Likewise.
11028 * cfgloop.h (place_new_loop): Add struct function parameter.
11029 (get_loop): Likewise.
11030 (get_loops): Likewise.
11031 (number_of_loops): Likewise.
11032 (fel_next): Adjust.
11033 (fel_init): Likewise.
11034 * cfg.c (get_loop_copy): Adjust.
11035 * cfgloop.c (flow_loops_dump): Likewise.
11036 (record_loop_exits): Likewise.
11037 (verify_loop_structure): Likewise.
11038 * cfgloopanal.c (mark_irreducible_loops): Likewise.
11039 (estimate_reg_pressure_cost): Likewise.
11040 (mark_loop_exit_edges): Likewise.
11041 * cfgloopmanip.c (place_new_loop): Likewise.
11042 (add_loop): Likewise.
11043 (duplicate_loop): Likewise.
11044 * graph.c (draw_cfg_nodes): Likewise.
11045 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
11046 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
11047 (extract_affine_chrec): Likewise.
11048 (build_scop_iteration_domain): Likewise.
11049 * graphite.c (graphite_initialize): Likewise.
11050 * ira-build.c (create_loop_tree_nodes): Likewise.
11051 (more_one_region_p): Likewise.
11052 (rebuild_regno_allocno_maps): Likewise.
11053 (mark_loops_for_removal): Likewise.
11054 (mark_all_loops_for_removal): Likewise.
11055 (remove_unnecessary_regions): Likewise.
11056 (ira_build): Likewise.
11057 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
11058 * loop-init.c (fix_loop_structure): Likewise.
11059 (gate_rtl_move_loop_invariants): Likewise.
11060 (gate_rtl_unswitch): Likewise.
11061 (gate_rtl_unroll_and_peel_loops): Likewise.
11062 (rtl_doloop): Likewise.
11063 * lto-streamer-in.c (input_cfg): Likewise.
11064 * lto-streamer-out.c (output_cfg): Likewise.
11065 * modulo-sched.c (sms_schedule): Likewise.
11066 * predict.c (tree_estimate_probability): Likewise.
11067 (tree_estimate_probability_driver): Likewise.
11068 (estimate_loops): Likewise.
11069 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
11070 (move_sese_region_to_fn): Likewise.
11071 (debug_loop_num): Likewise.
11072 * tree-chrec.c (chrec_evaluate): Likewise.
11073 (hide_evolution_in_other_loops_than_loop): Likewise.
11074 (chrec_component_in_loop_num): Likewise.
11075 (reset_evolution_in_loop): Likewise.
11076 (evolution_function_is_invariant_rec_p): Likewise.
11077 * tree-if-conv.c (main_tree_if_conversion): Likewise.
11078 * tree-inline.c (copy_loops): Likewise.
11079 (copy_cfg_body): Likewise.
11080 (tree_function_versioning): Likewise.
11081 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
11082 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
11084 (add_to_evolution_1): Likewise.
11085 (scev_const_prop): Likewise.
11086 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
11087 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
11088 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
11089 (tree_ssa_lim_initialize): Likewise.
11090 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
11091 (verify_loop_closed_ssa): Likewise.
11092 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
11093 (tree_ssa_loop_im): Likewise.
11094 (tree_ssa_loop_unswitch): Likewise.
11095 (tree_vectorize): Likewise.
11096 (check_data_deps): Likewise.
11097 (tree_ssa_loop_ivcanon): Likewise.
11098 (tree_ssa_loop_bounds): Likewise.
11099 (tree_complete_unroll): Likewise.
11100 (tree_complete_unroll_inner): Likewise.
11101 (tree_parallelize_loops): Likewise.
11102 (tree_ssa_loop_prefetch): Likewise.
11103 (tree_ssa_loop_ivopts): Likewise.
11104 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
11105 * tree-vectorizer.c (vectorize_loops): Likewise.
11107 2013-04-29 Mike Frysinger <vapier@gentoo.org>
11109 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
11110 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
11111 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
11112 with EABI_LINK_SPEC.
11114 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11117 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
11120 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
11123 * lra-constraints.c (process_alt_operands): Discourage a bit more
11124 using memory for pseudos. Print cost dump for alternatives.
11125 Modify cost values for conflicts with early clobbers.
11126 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
11128 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11131 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
11133 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11135 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
11136 from/to S register.
11137 (movdi_aarch64): Support LDR/STR from/to D register.
11139 2013-04-29 Ian Bolton <ian.bolton@arm.com>
11141 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
11142 or higher by default.
11144 2013-04-29 Richard Biener <rguenther@suse.de>
11146 PR middle-end/57075
11147 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
11148 even if not adding abnormal edges for calls that can make
11151 2013-04-29 Richard Biener <rguenther@suse.de>
11153 PR middle-end/57103
11154 * tree-cfg.c (move_stmt_op): Fix condition under which to update
11156 (move_stmt_r): Remove redundant checking.
11158 2013-04-29 Teresa Johnson <tejohnson@google.com>
11161 * basic-block.h (apply_scale): New function.
11162 (apply_probability): Use apply_scale.
11163 * gimple-streamer-in.c (input_bb): Ditto.
11164 * lto-streamer-in.c (input_cfg): Ditto.
11165 * lto-cgraph.c (merge_profile_summaries): Ditto.
11166 * tree-optimize.c (execute_fixup_cfg): Ditto.
11167 * tree-inline.c (copy_bb): Update comment to use apply_scale.
11168 (copy_edges_for_bb): Ditto.
11169 (copy_cfg_body): Ditto.
11171 2013-04-29 Tom de Vries <tom@codesourcery.com>
11173 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
11174 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
11175 (tail_merge_optimize): Handle current_loops == NULL.
11177 2013-04-26 Jeff Law <law@redhat.com>
11179 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
11180 (simplify_cond_using_ranges): Generalize code to simplify
11181 COND_EXPRs where one argument is a constant and the other
11182 is an SSA_NAME created by an integral type conversion.
11184 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11186 * config/arm/arm.md (store_minmaxsi): Use only when
11187 optimize_insn_for_size_p.
11189 2013-04-29 Christian Bruel <christian.bruel@st.com>
11192 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
11194 2013-04-29 Richard Biener <rguenther@suse.de>
11196 PR middle-end/57089
11197 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
11198 loop tree make sure to schedule a fixup for the child as well.
11199 (expand_omp_for_generic): Properly add loops.
11200 (expand_omp_for_static_nochunk): Likewise.
11201 (expand_omp_for_static_chunk): Likewise.
11202 (expand_omp_for): For the degenerate case fixup loops.
11203 (expand_omp_sections): Fix default bb placement in loops.
11204 (expand_omp_atomic_pipeline): Properly add loops.
11206 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11208 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
11210 2013-04-29 Tom de Vries <tom@codesourcery.com>
11212 * tree-ssa-tail-merge.c: Update header comment.
11214 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11216 * config/aarch64/arm_neon.h
11217 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
11218 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
11219 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
11220 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
11221 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
11222 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
11223 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
11224 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
11226 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11228 * config/aarch64/aarch64-simd.md
11229 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
11230 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
11231 fix_trunc, fixuns_trunc.
11232 (ftrunc<VDQF:mode>2): New.
11233 * config/aarch64/iterators.md (optab): Add fix, fixuns.
11234 (fix_trunc_optab): New.
11236 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11238 * config/aarch64/aarch64-builtins.c
11239 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
11240 iceilf, lround, iroundf.
11242 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
11245 * config/i386/i386.h (enum ix86_tune_indices)
11246 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
11247 New, split from X86_TUNE_INTER_UNIT_MOVES.
11248 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
11249 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
11250 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
11251 (TARGET_INTER_UNIT_MOVES): Remove.
11252 * config/i386/i386.c (initial_ix86_tune_features): Update.
11253 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
11254 (ix86_expand_convert_uns_didf_sse): Use
11255 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11256 (ix86_expand_vector_init_one_nonzero): Ditto.
11257 (ix86_expand_vector_init_interleave): Ditto.
11258 (inline_secondary_memory_needed): Return true for moves from SSE class
11259 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
11260 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
11261 * config/i386/constraints.md (Yi, Ym): Depend on
11262 TARGET_INTER_UNIT_MOVES_TO_VEC.
11263 (Yj, Yn): New constraints.
11264 * config/i386/i386.md (*movdi_internal): Change constraints of
11265 operand 1 from Yi to Yj and from Ym to Yn.
11266 (*movsi_internal): Ditto.
11267 (*movdf_internal): Ditto.
11268 (*movsf_internal): Ditto.
11269 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
11270 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
11271 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
11272 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
11273 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
11274 * config/i386/sse.md (movdi_to_sse): Ditto.
11275 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
11276 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
11277 TARGET_INTER_UNIT_MOVES.
11278 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
11279 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
11280 instead of TARGET_INTER_UNIT_MOVES.
11281 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
11282 operand 1 from Yi to Yj and from Ym to Yn.
11284 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11286 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
11287 (float_truncate_hi_): Likewise.
11288 (float_extend_lo_): Likewise.
11289 (float_truncate_lo_): Likewise.
11290 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
11291 (aarch64_float_extend_lo_v2df): Likewise.
11292 (vec_unpacks_hi_v4sf): Likewise.
11293 (aarch64_float_truncate_lo_v2sf): Likewise.
11294 (aarch64_float_truncate_hi_v4sf): Likewise.
11295 (vec_pack_trunc_v2df): Likewise.
11296 (vec_pack_trunc_df): Likewise.
11298 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11300 * config/aarch64/aarch64-builtins.c
11301 (aarch64_fold_builtin): Fold float conversions.
11302 * config/aarch64/aarch64-simd-builtins.def
11303 (floatv2si, floatv4si, floatv2di): New.
11304 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
11305 * config/aarch64/aarch64-simd.md
11306 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
11307 * config/aarch64/iterators.md (FLOATUORS): New.
11308 (optab): Add float, floatuns.
11309 (su_optab): Likewise.
11311 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11313 * config/aarch64/aarch64-builtins.c
11314 (aarch64_builtin_vectorized_function): Use new names for
11316 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
11317 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
11318 (fcvtzu): Split as...
11319 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
11320 (fcvtas): Split as...
11321 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
11322 (fcvtau): Split as...
11323 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
11324 (fcvtps): Split as...
11325 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
11326 (fcvtpu): Split as...
11327 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
11328 (fcvtms): Split as...
11329 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
11330 (fcvtmu): Split as...
11331 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
11332 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
11333 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
11334 (lfrintnusf, lfrintnudf): Likewise.
11335 * config/aarch64/aarch64-simd.md
11336 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
11338 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
11339 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
11340 (fcvt_pattern): Likewise.
11342 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11344 * config/aarch64/aarch64-simd.md
11345 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
11346 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
11348 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11350 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
11351 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
11352 (vrnd<a,m,n,p>_f32): Implement using builtins.
11353 (vrnd<i,x><q>_f<32, 64>): New.
11355 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
11357 * config/aarch64/aarch64-builtins.c
11358 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
11359 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
11360 (frintz): Rename to...
11362 (frintp): Rename to...
11364 (frintm): Rename to...
11366 (frinti): Rename to...
11367 (nearbyint): ...this.
11368 (frintx): Rename to...
11370 (frinta): Rename to...
11372 * config/aarch64/aarch64-simd.md
11373 (aarch64_frint<frint_suffix><mode>): Delete.
11374 (<frint_pattern><mode>2): Convert to insn.
11375 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
11376 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
11377 (frint_pattern): Likewise.
11378 (frint_suffix): Likewise.
11380 2013-04-29 Richard Biener <rguenther@suse.de>
11382 PR tree-optimization/57081
11383 * loop-init.c: Include tree-flow.h.
11384 (loop_optimizer_finalize): Free number of iteration estimates.
11385 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
11387 2013-04-29 Jakub Jelinek <jakub@redhat.com>
11389 PR tree-optimization/57083
11390 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
11391 non-singleton shift count range, zero extend low_bound for uns case.
11393 * config/i386/predicates.md (general_vector_operand): New predicate.
11394 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
11395 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
11396 if they aren't nonimmediate operands. If their original values
11397 satisfy const_vector_equal_evenodd_p, don't shift them.
11398 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
11399 predicates. For the SSE4.1 case force operands[{1,2}] into registers
11400 if not nonimmediate_operand.
11401 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
11402 instead of register_operand.
11403 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
11405 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
11407 * stor-layout.c (finalize_size_functions): Allocate a structure and
11408 reset cfun before dumping the functions.
11410 2013-04-27 Jakub Jelinek <jakub@redhat.com>
11412 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
11415 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
11416 use xop_pmacsdqh if uns_p.
11417 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
11418 the immediate rotate count.
11420 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
11422 * rtl.h (struct rtx_def): Add comment for field jump.
11423 (LRA_SUBREG_P): New macro.
11424 * recog.c (register_operand): Check LRA_SUBREG_P.
11425 * lra.c (lra): Add note at the end of RTL code. Align non-empty
11427 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
11428 (lra_final_code_change): Skip subreg change for operators.
11429 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
11430 if there are no operand changes.
11431 * lra-constraints.c (curr_insn_set): New.
11432 (match_reload): Set LRA_SUBREG_P.
11433 (emit_spill_move): Ditto.
11434 (check_and_process_move): Use curr_insn_set. Process only single
11435 set insns. Don't initialize sec_mem_p and change_p.
11436 (simplify_operand_subreg): Use LRA_SUBREG_P.
11437 (reg_in_class_p): New function.
11438 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
11439 of #ifdef. Add code to remove cycling.
11440 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
11441 non-null disp. Reload inner instead of disp when base and index
11442 are null. Try to put lo_sum into register.
11443 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
11444 (check_and_process_move): Move code for move cost check to
11445 simple_move_p. Remove equiv_substitution.
11446 (simple_move_p): New function.
11447 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
11448 curr_insn_set. Call check_and_process_move only for single set
11449 insns. Use the new function. Move call of check_and_process_move
11450 after operand equiv substitution and address process.
11452 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11455 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
11456 with nonlocal goto receivers or returns twice calls, ignore
11457 unininitialized values from abnormal edges to nl goto receiver
11458 or returns twice call.
11460 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11462 PR tree-optimization/57051
11463 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
11464 and VEC_RSHIFT_EXPR if shift count is a multiple of element
11467 2013-04-26 Richard Biener <rguenther@suse.de>
11469 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
11470 (expand_omp_taskreg): Likewise. Mark loops for fixup.
11471 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
11472 (fixup_loop_arrays_after_move): New function.
11473 (move_sese_region_to_fn): Properly outline the loop tree parts
11474 of the SESE region.
11476 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
11478 * config/i386/i386.md (type, unit): Fix long lines.
11480 2013-04-26 Richard Biener <rguenther@suse.de>
11482 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
11483 (lto-streamer-out.o): Likewise.
11484 * cfgloop.c (init_loops_structure): Export, add struct function
11485 argument and adjust.
11486 (flow_loops_find): Adjust.
11487 * cfgloop.h (enum loop_estimation): Add EST_LAST.
11488 (init_loops_structure): Declare.
11489 * lto-streamer-in.c: Include cfgloop.h.
11490 (input_cfg): Input the loop tree.
11491 * lto-streamer-out.c: Include cfgloop.h.
11492 (output_cfg): Output the loop tree.
11493 (output_struct_function_base): Do not drop PROP_loops.
11495 2013-03-26 Richard Biener <rguenther@suse.de>
11497 * tree-cfg.c (execute_build_cfg): Build the loop tree.
11498 (pass_build_cfg): Provide PROP_loops.
11499 (move_sese_region_to_fn): Remove loops that are outlined into fn
11501 * tree-inline.c: Include cfgloop.h.
11502 (initialize_cfun): Do not drop PROP_loops.
11503 (copy_loops): New function.
11504 (copy_cfg_body): Copy loop structure.
11505 (tree_function_versioning): Initialize destination loop tree.
11506 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
11507 (pass_parallelize_loops): Do IL verification.
11508 * loop-init.c (loop_optimizer_init): Fixup loops if required.
11509 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
11510 the CFG make sure we fixup loops as well.
11511 * tree-ssa-tail-merge.c: Include cfgloop.h.
11512 (replace_block_by): When merging loop latches mark loops for fixup.
11513 * lto-streamer-out.c (output_struct_function_base): Drop
11514 PROP_loops for now.
11515 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
11516 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
11517 * ipa-split.c: Include cfgloop.h.
11518 (split_function): Add the new return block to the loop tree root.
11519 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
11520 whether we have removed the forwarder block.
11521 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
11522 * cfgloop.h (place_new_loop): Declare.
11523 * cfgloopmanip.c (place_new_loop): Export.
11524 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
11525 (tree-switch-conversion.o): Likewise.
11526 (tree-complex.o): Likewise.
11527 (tree-inline.o): Likewise.
11528 (tree-ssa-tailmerge.o): Likewise.
11529 (ipa-split.o): Likewise.
11530 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
11531 (tree-ssa-copy.o): Likewise.
11532 * tree-switch-conversion.c: Include cfgloop.h
11533 (process_switch): If we emit a bit-test cascade, schedule loops
11535 * tree-complex.c: Include cfgloop.h.
11536 (expand_complex_div_wide): Properly add new basic-blocks to loops.
11537 * asan.c: Include cfgloop.h.
11538 (create_cond_insert_point): Properly add new basic-blocks to
11539 loops, schedule loop fixup.
11540 * cfgloop.c (verify_loop_structure): Check that looks are not
11542 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
11544 (expand_omp_for_generic): Likewise.
11545 (expand_omp_sections): Likewise.
11546 (expand_omp_atomic_pipeline): Schedule loops for fixup.
11547 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
11548 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
11549 is initialized, not when loops are present.
11550 * tree-parloops.c (parallelize_loops): Remove checking here.
11551 * passes.c (init_optimization_passes): Schedule a copy-propagation
11552 pass before complete unrolling of inner loops.
11554 2013-04-26 Jakub Jelinek <jakub@redhat.com>
11556 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
11557 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
11558 (colorize_init): Add argument to _WIN32 version.
11559 * toplev.c: Include diagnostic-color.h.
11560 (process_options): Default to -fdiagnostics-color=auto if
11561 GCC_COLORS env var is in the environment.
11562 * common.opt (fdiagnostics-color=): Add Var and Init.
11563 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
11564 env var is in the environment, the default is auto rather than never.
11566 * diagnostic.h (file_name_as_prefix): Add context argument.
11567 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
11568 the string as locus.
11569 * langhooks.c (lhd_print_error_function): Adjust caller.
11571 2013-04-25 Lawrence Crowl <crowl@google.com>
11573 * var-tracking.c (shared_hash_def::htab):
11574 Change type to hash_table. Update dependent calls and types.
11576 2013-04-25 Lawrence Crowl <crowl@google.com>
11578 * Makefile.in: Update as needed below.
11580 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
11581 Move declaration to after the type's method definitons.
11583 * attribs.c (htab_t scoped_attributes::attribute_hash):
11584 Change type to hash_table. Update dependent calls and types.
11586 * bitmap.c (htab_t bitmap_desc_hash):
11587 Change type to hash_table. Update dependent calls and types.
11589 * cselib.c (htab_t cselib_hash_table):
11590 Change type to hash_table. Update dependent calls and types.
11592 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
11593 (hash_string_slot_node): Move implementation into lto-streamer.h
11594 struct string_slot_hasher.
11595 (eq_string_slot_node): Likewise.
11597 * data-streamer-out.c: Update output_block::string_hash_table
11598 dependent calls and types.
11600 * dwarf2cfi.c (htab_t trace_index):
11601 Change type to hash_table. Update dependent calls and types.
11603 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
11604 Change type to hash_table. Update dependent calls and types.
11605 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
11606 (htab_t optimize_external_refs::map): Likewise.
11607 (htab_t output_comp_unit::extern_map): Likewise.
11608 (htab_t output_comdat_type_unit::extern_map): Likewise.
11609 (htab_t output_macinfo::macinfo_htab): Likewise.
11610 (htab_t optimize_location_lists::htab): Likewise.
11611 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
11613 * except.c (htab_t ehspec_hash_type):
11614 Change type to hash_table. Update dependent calls and types.
11615 (assign_filter_values::ttypes): Likewise.
11616 (assign_filter_values::ehspec): Likewise.
11617 (sjlj_assign_call_site_values::ar_hash): Likewise.
11618 (convert_to_eh_region_ranges::ar_hash): Likewise.
11620 * gcse.c (htab_t pre_ldst_table):
11621 Change type to hash_table. Update dependent calls and types.
11623 * ggc-common.c (htab_t saving_htab):
11624 Change type to hash_table. Update dependent calls and types.
11625 (htab_t loc_hash): Likewise.
11626 (htab_t ptr_hash): Likewise.
11627 (call_count): Rename ggc_call_count.
11628 (call_alloc): Rename ggc_call_alloc.
11629 (loc_descriptor): Rename make_loc_descriptor.
11630 (add_statistics): Rename ggc_add_statistics.
11632 * ggc-common.c (saving_htab):
11633 Change type to hash_table. Update dependent calls and types.
11635 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
11636 (push_gimplify_context): Likewise.
11637 (pop_gimplify_context): Likewise.
11638 (struct gimple_temp_hash_elt): Added.
11639 (struct gimplify_hasher): Likewise.
11640 (struct gimplify_ctx.temp_htab):
11641 Change type to hash_table. Update dependent calls and types.
11643 * gimple-fold.c: Include gimplify-ctx.h.
11645 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
11646 Change type to hash_table. Update dependent calls and types.
11647 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
11648 avoid potential global name collision.
11650 * gimplify.c: Include gimplify-ctx.h.
11651 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
11652 (htab_t gimplify_ctx::temp_htab):
11653 Update dependent calls and types for new type hash_table.
11654 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
11655 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
11657 * gimplify-ctx.h: New.
11658 (struct gimple_temp_hash_elt): Move from gimplify.c.
11659 (class gimplify_hasher): New.
11660 (struct gimplify_ctx): Move from gimple.h.
11661 (htab_t gimplify_ctx::temp_htab):
11662 Change type to hash_table. Update dependent calls and types.
11664 * graphite-clast-to-gimple.c: Include graphite-htab.h.
11665 (htab_t ivs_params::newivs_index):
11666 Change type to hash_table. Update dependent calls and types.
11667 (htab_t ivs_params::params_index): Likewise.
11668 (htab_t print_generated_program::params_index): Likewise.
11669 (htab_t gloog::newivs_index): Likewise.
11670 (htab_t gloog::params_index): Likewise.
11672 * graphite.c: Include graphite-htab.h.
11673 4htab_t graphite_transform_loops::bb_pbb_mapping):
11674 Change type to hash_table. Update dependent calls and types.
11676 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
11677 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
11678 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
11680 * graphite-dependences.c: Include graphite-htab.h.
11681 (loop_is_parallel_p): Change hash table type of parameter.
11683 * graphite-htab.h: New.
11684 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
11685 (extern find_pbb_via_hash): Move from graphite-poly.h.
11686 (extern loop_is_parallel_p): Move from graphite-poly.h.
11687 (extern get_loop_body_pbbs): Move from graphite-poly.h.
11689 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
11690 (extern loop_is_parallel_p): Move to graphite-htab.h.
11691 (extern get_loop_body_pbbs): Move to graphite-htab.h.
11693 * haifa-sched.c (htab_t delay_htab):
11694 Change type to hash_table. Update dependent calls and types.
11695 (htab_t delay_htab_i2): Likewise.
11697 * ira-color.c (htab_t allocno_hard_regs_htab):
11698 Change type to hash_table. Update dependent calls and types.
11700 * ira-costs.c (htab_t cost_classes_htab):
11701 Change type to hash_table. Update dependent calls and types.
11703 * loop-invariant.c (htab_t merge_identical_invariants::eq):
11704 Change type to hash_table. Update dependent calls and types.
11706 * loop-iv.c (htab_t bivs):
11707 Change type to hash_table. Update dependent calls and types.
11709 * loop-unroll.c (htab_t opt_info::insns_to_split):
11710 Change type to hash_table. Update dependent calls and types.
11711 (htab_t opt_info::insns_with_var_to_expand): Likewise.
11713 * lto-streamer.h (struct string_slot): Move from data-streamer.h
11714 (struct string_slot_hasher): New.
11715 (htab_t output_block::string_hash_table):
11716 Change type to hash_table. Update dependent calls and types.
11718 * lto-streamer-in.c (freeing_string_slot_hasher): New.
11719 (htab_t file_name_hash_table):
11720 Change type to hash_table. Update dependent calls and types.
11722 * lto-streamer-out.c: Update output_block::string_hash_table dependent
11725 * lto-streamer.c (htab_t tree_htab):
11726 Change type to hash_table. Update dependent calls and types.
11728 * omp-low.c: Include gimplify-ctx.h.
11730 * passes.c (htab_t name_to_pass_map):
11731 Change type to hash_table. Update dependent calls and types.
11732 (pass_traverse): Rename to passes_pass_traverse.
11734 * plugin.c (htab_t event_tab):
11735 Change type to hash_table. Update dependent calls and types.
11737 * postreload-gcse.c (htab_t expr_table):
11738 Change type to hash_table. Update dependent calls and types.
11739 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
11741 * sese.c (debug_rename_map_1): Make extern.
11742 (htab_t copy_bb_and_scalar_dependences::rename_map):
11743 Change type to hash_table. Update dependent calls and types.
11745 * sese.h (extern debug_rename_map): Move to .c file.
11747 * store-motion.c (htab_t store_motion_mems_table):
11748 Change type to hash_table. Update dependent calls and types.
11750 * trans-mem.c (htab_t tm_new_mem_hash):
11751 Change type to hash_table. Update dependent calls and types.
11753 * tree-browser.c (htab_t TB_up_ht):
11754 Change type to hash_table. Update dependent calls and types.
11756 * tree-cfg.c (htab_t discriminator_per_locus):
11757 Change type to hash_table. Update dependent calls and types.
11759 * tree-complex.c: Include tree-hasher.h
11760 (htab_t complex_variable_components):
11761 Change type to hash_table. Update dependent calls and types.
11763 * tree-eh.c (htab_t finally_tree):
11764 Change type to hash_table. Update dependent calls and types.
11766 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
11767 struct int_tree_hasher.
11768 (extern int_tree_map_eq): Likewise.
11769 (uid_decl_map_hash): Removed.
11770 (extern decl_tree_map_eq): Likewise.
11772 * tree-hasher.h: New.
11773 (struct int_tree_hasher): New.
11774 (typedef int_tree_htab_type): New.
11776 * tree-inline.c: Include gimplify-ctx.h.
11778 * tree-mudflap.c: Include gimplify-ctx.h.
11780 * tree-parloops.c: Include tree-hasher.h.
11781 (htab_t eliminate_local_variables_stmt::decl_address):
11782 Change type to hash_table. Update dependent calls and types.
11783 (htab_t separate_decls_in_region::decl_copies): Likewise.
11785 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
11786 Change type to hash_table. Update dependent calls and types.
11788 * tree-sra.c (candidates):
11789 Change type to hash_table. Update dependent calls and types.
11791 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
11793 (int_tree_map_hash): Likewise.
11795 * tree-ssa-dom.c (htab_t avail_exprs):
11796 Change type to hash_table. Update dependent calls and types.
11798 * tree-ssa-live.c (var_map_base_init::tree_to_index):
11799 Change type to hash_table. Update dependent calls and types.
11801 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
11802 Change type to hash_table. Update dependent calls and types.
11804 * tree-ssa-phiopt.c (seen_ssa_names):
11805 Change type to hash_table. Update dependent calls and types.
11807 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
11808 Change type to hash_table. Update dependent calls and types.
11810 * tree-ssa-uncprop.c (equiv):
11811 Change type to hash_table. Update dependent calls and types.
11813 2013-04-25 Jakub Jelinek <jakub@redhat.com>
11815 PR rtl-optimization/57003
11816 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
11817 call note_stores with kill_clobbered_value callback again after
11818 killing regs_invalidated_by_call.
11820 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11822 * config/aarch64/aarch64-simd.md
11823 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
11824 (aarch64_simd_bsl<mode>): Likewise.
11825 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
11827 2013-04-25 Marek Polacek <polacek@redhat.com>
11829 PR tree-optimization/57066
11830 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
11832 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
11834 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
11836 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11838 * config/aarch64/aarch64-builtins.c
11839 (aarch64_fold_builtin): New.
11840 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
11841 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
11842 * config/aarch64/aarch64-simd-builtins.def (abs): New.
11843 * config/aarch64/arm_neon.h
11844 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
11846 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11847 Tejas Belagod <tejas.belagod@arm.com>
11849 * config/aarch64/aarch64-builtins.c
11850 (aarch64_gimple_fold_builtin): New.
11851 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
11852 * config/aarch64/aarch64-simd-builtins.def (addv): New.
11853 * config/aarch64/aarch64-simd.md (addpv4sf): New.
11854 (addvv4sf): Update.
11855 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
11857 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11859 * config/aarch64/aarch64.md
11860 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
11862 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
11864 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
11865 (*ngcsi_uxtw): New pattern.
11867 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11868 Julian Brown <julian@codesourcery.com>
11870 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
11871 (TB_DREG): Add T_V4HF.
11872 (v4hf_UP): New macro.
11873 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11874 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11875 Handle initialisation of V4HF. Adjust initialisation of reinterpret
11877 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
11878 (arm_vector_mode_supported_p): Handle V4HF.
11879 (arm_mangle_map): Handle V4HFmode.
11880 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
11881 * config/arm/arm_neon_builtins.def: Add entries for
11882 vcvtv4hfv4sf, vcvtv4sfv4hf.
11883 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
11884 (neon_vcvtv4hfv4sf): Likewise.
11885 * config/arm/neon-gen.ml: Handle half-precision floating point
11887 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
11888 * config/arm/arm_neon.h: Regenerate.
11889 * config/arm/neon.ml (type elts): Add F16.
11890 (type vectype): Add T_float16x4, T_floatHF.
11891 (type vecmode): Add V4HF.
11892 (type features): Add Requires_FP_bit feature.
11893 (elt_width): Handle F16.
11894 (elt_class): Likewise.
11895 (elt_of_class_width): Likewise.
11896 (mode_of_elt): Refactor.
11897 (type_for_elt): Handle F16, fix error messages.
11898 (vectype_size): Handle T_float16x4.
11899 (vcvt_sh): New function.
11900 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
11901 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
11902 (string_of_mode): Handle V4HF.
11903 * doc/arm-neon-intrinsics.texi: Regenerate.
11905 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
11907 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
11908 format specifier in 'X' case.
11910 2013-04-25 Alan Modra <amodra@gmail.com>
11913 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
11914 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
11915 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
11916 Repeat for many other rotate/shift and mask patterns using subregs.
11917 Name lshiftrt insns.
11918 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
11919 on WORDS_BIG_ENDIAN.
11921 2013-04-25 Alan Modra <amodra@gmail.com>
11923 * config.gcc: Support little-endian powerpc-linux targets.
11924 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
11925 (LINK_OS_LINUX_SPEC): Define.
11926 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
11927 Preserve MASK_LITTLE_ENDIAN.
11928 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
11929 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
11930 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
11931 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
11932 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
11933 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
11934 Correct fp word order for little-endian. Don't shift toc entries
11935 smaller than a word for little-endian.
11936 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
11937 (bswapdi2 splits): Correct low-part subreg for little-endian.
11938 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
11939 low/high where such is correct only for be.
11940 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
11941 little-endian for -mcall-aixdesc.
11943 2013-04-25 Alan Modra <amodra@gmail.com>
11945 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
11946 replace_equiv_address_nv.
11948 2013-04-25 Alan Modra <amodra@gmail.com>
11950 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
11952 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11955 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11956 * rtl.h (struct rtx_def): ...
11958 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11960 PR rtl-optimizations/57046
11961 * lra-constraints (split_reg): Set up lra_risky_transformations_p
11962 for multi-reg splits.
11964 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
11966 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
11968 2013-04-24 Sterling Augustine <saugustine@google.com>
11970 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
11971 (comp_dir_string, debug_str_dwo_section): New.
11972 (DEBUG_STR_DWO_SECTION): Rename to ...
11973 (DEBUG_DWO_STR_SECTION): ... this.
11974 (DEBUG_NORM_STR_SECTION): Delete.
11975 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
11976 (DEBUG_STR_DWO_SECTION_FLAGS): New.
11977 (find_AT_string): Move most logic to ...
11978 (find_AT_string_in_table): ... here. New.
11979 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
11980 add_skeleton_AT_string. Delete logic.
11981 (output_skeleton_debug_sections): Remove call to
11982 add_top_level_skeleton_die_attrs.
11983 (add_comp_dir_attribute): Move logic to comp_dir_string.
11984 (dwarf2out_init): Initialize debug_str_dwo_section.
11985 (output_indirect_string): Call find_string_form.
11986 (output_indirect_strings): Rewrite.
11987 (prune_unused_types): Empty skeleton_debug_str_hash.
11988 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
11989 (dwarf2out_finish): Call output_indirect_strings.
11991 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
11993 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
11995 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
11997 * rtl.h (struct rtx_def): Add comment for field jump.
11998 (LRA_SUBREG_P): New macro.
11999 * recog.c (register_operand): Check LRA_SUBREG_P.
12000 * lra.c (lra): Add note at the end of RTL code. Align non-empty
12002 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
12003 (lra_final_code_change): Skip subreg change for operators.
12004 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
12005 if there are no operand changes.
12006 * lra-constraints.c (curr_insn_set): New.
12007 (match_reload): Set LRA_SUBREG_P.
12008 (emit_spill_move): Ditto.
12009 (check_and_process_move): Use curr_insn_set. Process only single
12010 set insns. Don't initialize sec_mem_p and change_p.
12011 (simplify_operand_subreg): Use LRA_SUBREG_P.
12012 (reg_in_class_p): New function.
12013 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
12014 of #ifdef. Add code to remove cycling.
12015 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
12016 non-null disp. Reload inner instead of disp when base and index
12017 are null. Try to put lo_sum into register.
12018 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
12019 (check_and_process_move): Move code for move cost check to
12020 simple_move_p. Remove equiv_substitution.
12021 (simple_move_p): New function.
12022 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
12023 curr_insn_set. Call check_and_process_move only for single set
12024 insns. Use the new function. Move call of check_and_process_move
12025 after operand equiv substitution and address process.
12027 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
12029 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
12030 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
12031 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
12033 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
12035 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
12037 2013-04-24 Marek Polacek <polacek@redhat.com>
12039 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
12040 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
12041 (select_loops_exit_conditions): Likewise.
12042 (number_of_iterations_for_all_loops): Likewise.
12043 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
12044 (scev_analysis): Likewise.
12046 2013-04-02 Catherine Moore <clm@codesourcery.com>
12047 Chao-ying Fu <fu@mips.com>
12049 * config/mips/micromips.md (jraddiusp): New pattern.
12050 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
12051 instruction if possible.
12053 2013-04-24 Alan Modra <amodra@gmail.com>
12055 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
12057 2013-04-24 Julian Brown <julian@codesourcery.com>
12058 Chung-Lin Tang <cltang@codesourcery.com>
12060 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
12061 dependency behavior in enumeration type DIE generation. Add TODO note
12062 to comments about future DW_FORM_sdata/udata re-work of related code.
12064 2013-04-23 Lawrence Crowl <crowl@google.com>
12066 * Makefile.in: Update as needed below.
12068 * hash-table.h (class hash_table):
12069 Correct many methods with parameter types compare_type to the correct
12070 value_type. (Correct code was unlikely to notice the change.)
12071 (hash_table::elements_with_deleted) New.
12072 (class hashtable::iterator): New.
12073 (hashtable::begin()): New.
12074 (hashtable::end()): New.
12075 (FOR_EACH_HASH_TABLE_ELEMENT): New.
12077 * statistics.c (statistics_hashes):
12078 Change type to hash_table. Update dependent calls and types.
12080 * tree-into-ssa.c (var_infos):
12081 Change type to hash_table. Update dependent calls and types.
12083 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
12084 Change type to hash_table. Update dependent calls and types.
12086 * tree-ssa-loop-im.c (struct mem_ref.refs):
12087 Change type to hash_table. Update dependent calls and types.
12089 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
12090 Change type to hash_table. Update dependent calls and types.
12092 * tree-ssa-sccvn.c (vn_tables_s::nary):
12093 Change type to hash_table. Update dependent calls and types.
12094 (vn_tables_s::phis): Likewise.
12095 (vn_tables_s::references): Likewise.
12097 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
12098 (vn_reference_eq): Update parameter and return types.
12100 * tree-ssa-structalias.c (pointer_equiv_class_table):
12101 Change type to hash_table. Update dependent calls and types.
12102 (location_equiv_class_table): Likewise.
12104 * tree-vect-data-refs.c: Consequential changes for making
12105 peeling a hash_table.
12107 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
12108 (destroy_loop_vec_info): Dependent hash_table update.
12110 * tree-vectorizer.h (peeling_htab):
12111 Change type to hash_table. Update dependent calls and types.
12113 2013-04-23 Shiva Chen <shiva0217@gmail.com>
12115 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
12116 to check the register content is equal or not.
12117 * lra-constraints.c (match_reload): Use lra_assign_reg_val
12118 to assign register content record.
12119 * lra-eliminations.c (update_reg_eliminate): Use
12120 lra_update_reg_val_offset to update register content offset.
12121 * lra-int.h (struct lra_reg): Add offset member.
12122 (lra_reg_val_equal_p): New static inline function.
12123 (lra_update_reg_val_offset): New static inline function.
12124 (lra_assign_reg_val): New static inline function.
12125 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
12126 to assign register content record.
12127 (initialize_lra_reg_info_element): Initial offset to zero.
12129 2013-04-23 Catherine Moore <clm@codesourcery.com>
12131 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
12132 operands. Record compression.
12134 2013-04-23 Xinliang David Li <davidxl@google.com>
12136 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
12138 2013-04-23 Richard Biener <rguenther@suse.de>
12140 PR middle-end/57036
12141 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
12142 parameter, only add abnormal goto edges from the copied body
12143 if the call could perform abnormal gotos.
12144 (copy_cfg_body): Adjust.
12146 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
12148 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
12150 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
12152 * coretypes.h (gimple_stmt_iterator): Add struct to make
12155 2013-04-23 Richard Biener <rguenther@suse.de>
12157 PR tree-optimization/57026
12158 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
12159 from SSA names occuring in abnormal PHI nodes.
12161 2013-04-22 Andi Kleen <ak@linux.intel.com>
12163 * lto/lto.c (print_lto_report_1): Fix LTO report names.
12165 2013-04-22 Andi Kleen <ak@linux.intel.com>
12167 * lto/lto.c (print_lto_report_1): Declare early.
12168 (read_cgraph_and_symbols): Call print_lto_report_1 early.
12170 2013-04-22 Andi Kleen <ak@linux.intel.com>
12172 * common.opt (-flto-report-wpa): Add.
12173 * doc/invoke.texi (-flto-report-wpa): Add.
12174 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
12177 2013-04-22 Xinliang David Li <davidxl@google.com>
12179 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
12180 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
12181 * Makefile.in: New dependency
12183 David Daney <ddaney.cavm@gmail.com>
12185 * configure.ac (gcc_cv_as_micromips_support): Use the
12186 --fatal-warnings option.
12187 * configure: Regenerate.
12189 2013-04-22 Marek Polacek <polacek@redhat.com>
12192 * tsan.c (instrument_expr): Don't instrument expression
12193 in case its size is zero.
12195 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
12199 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
12201 * config/alpha/alpha.c (TARGET_LRA_P): New define.
12203 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12205 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
12206 (gimple_stmt_iterator): New typedef.
12207 * gimple.h (gimple_stmt_iterator): Rename to...
12208 (gimple_stmt_iterator_d): ... This.
12209 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
12210 trees be valid for GIMPLE and GENERIC.
12211 (TARGET_GIMPLE_FOLD_BUILTIN): New.
12212 * gimple-fold.c (gimple_fold_call): Call target hook
12213 gimple_fold_builtin.
12214 * hooks.c (hook_bool_gsiptr_false): New.
12215 * hooks.h (hook_bool_gsiptr_false): New.
12216 * target.def (fold_stmt): New.
12217 * doc/tm.texi: Regenerate.
12219 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
12222 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
12223 a set sp if no stack realignment.
12225 2013-04-22 Nick Clifton <nickc@redhat.com>
12227 * config.gcc (tilegx-linux): Extend extra_objs rather than
12229 (tilepro-linux): Likewise.
12231 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12233 * config/aarch64/aarch64-builtins.c
12235 (CF0, CF1, CF2, CF3, CF4, CF10): New.
12236 (VAR<1-12>): Add MAP parameter.
12237 (BUILTIN_*): Likewise.
12238 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
12239 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
12240 (aarch64_ushl_n<mode>): Likewise.
12241 (aarch64_sshr_n<mode>): Likewise.
12242 (aarch64_ushr_n<mode>): Likewise.
12243 (aarch64_<maxmin><mode>): Likewise.
12244 (aarch64_sqrt<mode>): Likewise.
12245 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
12246 (vshr<q>_n_*): Likewise.
12248 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
12250 * config/aarch64/aarch64-builtins.c
12251 (aarch64_simd_builtin_type_mode): Handle SF types.
12253 (BUILTIN_GPF): Define.
12254 (aarch64_init_simd_builtins): Handle SF types.
12255 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
12256 (frecps): Likewise.
12257 (frecpx): Likewise.
12258 * config/aarch64/aarch64-simd.md
12259 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
12260 (aarch64_frecpe<mode>): New.
12261 (aarch64_frecps<mode>): Likewise.
12262 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
12263 (v8type): Add frecp<esx>.
12264 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
12265 (aarch64_frecps<mode>): Likewise.
12266 * config/aarch64/iterators.md (FRECP): New.
12267 (frecp_suffix): Likewise.
12268 * config/aarch64/arm_neon.h
12269 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
12271 2013-04-22 Christian Bruel <christian.bruel@st.com>
12274 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
12275 (REG_CLASS_NAMES): Idem.
12276 (REG_CLASS_CONTENTS): Idem.
12277 (REGCLASS_HAS_FP_REG): Idem.
12278 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
12279 (sh_conditional_register_usage): Idem.
12281 2013-04-21 Jeff Law <law@redhat.com>
12283 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
12284 (ssa_forward_propagate_and_combine): Use it.
12286 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12288 * lra.c: Update the flow chart diagram.
12290 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
12292 PR rtl-optimization/56847
12293 * lra-constraints.c (process_alt_operands): Discourage alternative
12294 with non-matche doffsettable memory constraint fro memory with
12297 2013-04-19 Richard Biener <rguenther@suse.de>
12299 PR tree-optimization/56982
12300 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
12302 * gimplify.c (gimplify_call_expr): Notice special calls.
12303 (gimplify_modify_expr): Likewise.
12304 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
12305 abnormal control flow receivers.
12306 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
12307 in the same way as cfun->has_nonlocal_labels.
12308 (gimple_purge_dead_abnormal_call_edges): Likewise.
12309 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
12310 receivers start a basic-block.
12312 2013-04-19 Richard Biener <rguenther@suse.de>
12314 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
12316 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
12317 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
12318 (SLP_TREE_LOAD_PERMUTATION): Add.
12319 (vect_transform_slp_perm_load): Adjust prototype.
12320 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
12321 (vect_free_slp_instance): Likewise.
12322 (vect_create_new_slp_node): Likewise.
12323 (vect_supported_slp_permutation_p): Remove.
12324 (vect_slp_rearrange_stmts): Adjust.
12325 (vect_supported_load_permutation_p): Likewise. Inline
12326 vect_supported_slp_permutation_p here.
12327 (vect_analyze_slp_instance): Compute load permutations per
12328 slp node instead of per instance.
12329 (vect_get_slp_defs): Adjust.
12330 (vect_transform_slp_perm_load): Likewise.
12331 (vect_schedule_slp_instance): Remove redundant code.
12332 (vect_schedule_slp): Remove hack for PR56270, add it ...
12333 * tree-vect-stmts.c (vectorizable_load): ... here, do not
12334 CSE loads for SLP. Adjust.
12336 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12338 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
12339 spelling in two comments.
12341 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
12344 * config/arm/arm.c (load_multiple_sequence): Require SP
12345 as base register for loads if SP is in the register list.
12347 2013-04-19 Martin Jambor <mjambor@suse.cz>
12349 PR tree-optimization/56718
12350 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
12351 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
12352 and made public. Adjusted all callers.
12353 (ipa_intraprocedural_devirtualization): New function.
12354 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
12355 (ipa_intraprocedural_devirtualization): Likewise.
12356 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
12358 2013-04-19 Richard Biener <rguenther@suse.de>
12360 PR tree-optimization/57000
12361 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
12363 2013-04-19 Terry Guo <terry.guo@arm.com>
12365 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
12367 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
12368 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
12369 (cortex_m4_fmacs): Use new reservations.
12370 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
12372 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
12374 PR rtl-optimization/56999
12375 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
12377 (lra_coalesce): Remove split_origin_bitmap and related code.
12378 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
12379 ranges if necessary.
12381 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
12383 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
12385 (ix86_expand_call): Remove clobbered_registers array and use
12386 x86_64_ms_sysv_extra_clobbered_registers instead.
12387 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
12389 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
12391 * config/i386/i386.md (*call_rex64_ms_sysv): Use
12392 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
12393 (*call_value_rex64_ms_sysv): Ditto.
12395 2013-04-18 Cary Coutant <ccoutant@google.com>
12397 * dwarf2out.c (output_pubnames): Check die_perennial_p of
12398 parent instead of die_mark.
12400 2013-04-18 Diego Novillo <dnovillo@google.com>
12402 * gimple.c (create_gimple_tmp): New.
12403 (get_expr_type): New.
12404 (build_assign): New.
12405 (build_type_cast): New.
12406 * gimple.h (enum ssa_mode): Define.
12407 (gimple_seq_set_location): New.
12408 * asan.c (build_check_stmt): Change some gimple_build_* calls
12409 to use build_assign and build_type_cast.
12411 2013-04-18 Richard Biener <rguenther@suse.de>
12413 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
12414 handle negative step. Remove redundant checks.
12415 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
12416 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
12417 for negative step and grouped loads fail to vectorize.
12419 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12421 * emit-rtl.c (reset_insn_used_flags): New function.
12422 (reset_all_used_flags): Use it.
12423 (verify_insn_sharing): New function.
12424 (verify_rtl_sharing): Fix verification for SEQUENCEs.
12426 2013-04-18 Jakub Jelinek <jakub@redhat.com>
12428 PR tree-optimization/56984
12429 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
12430 and (x >> M) >= N don't register any assertion if N << M is the
12433 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
12435 * lower-subreg.c (resolve_simple_move): If called self-recursive,
12436 do not delete_insn insns that have not yet been emitted, only
12437 unlink them with remove_insn.
12438 * df-scan.c (df_insn_delete): Revert r197492.
12440 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
12442 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
12443 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
12445 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12447 * config/arm/arm.md (movsicc_insn): Convert define_insn into
12448 define_insn_and_split.
12449 (and_scc,ior_scc,negscc): Likewise.
12450 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
12452 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
12454 * config/arm/arm.c (use_return_insn): Return 0 for targets that
12455 can benefit from using a sequence of LDRD instructions in epilogue
12456 instead of a single LDM instruction.
12458 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
12461 * doc/extend.texi: Fix typo.
12463 2013-04-17 Richard Biener <rguenther@suse.de>
12465 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
12466 (vect_build_slp_tree): ... here.
12467 (vect_build_slp_tree_1): Compute which stmts of the SLP group
12468 match. Remove special-casing of mismatched complex loads.
12469 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
12470 re-try the match with swapped commutative operands.
12471 (vect_supported_load_permutation_p): Remove special-casing of
12472 mismatched complex loads.
12473 (vect_analyze_slp_instance): Adjust.
12475 2013-04-17 Richard Biener <rguenther@suse.de>
12477 PR rtl-optimization/56921
12478 * cfgloop.h (struct loop): Add simple_loop_desc member.
12479 (struct niter_desc): Mark with GTY(()).
12480 (simple_loop_desc): Do not use aux field but simple_loop_desc.
12481 * loop-iv.c (get_simple_loop_desc): Likewise.
12482 (free_simple_loop_desc): Likewise.
12485 2013-04-16 Richard Biener <rguenther@suse.de>
12487 PR rtl-optimization/56921
12488 * loop-init.c (pass_rtl_move_loop_invariants): Add
12489 TODO_do_not_ggc_collect to todo_flags_finish.
12490 (pass_rtl_unswitch): Same.
12491 (pass_rtl_unroll_and_peel_loops): Same.
12492 (pass_rtl_doloop): Same.
12494 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
12496 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
12497 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
12498 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
12500 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
12501 * tree-streamer.c (record_common_node): Adjust reference in comment.
12503 2013-04-17 Terry Guo <terry.guo@arm.com>
12505 * config/arm/cortex-m4.md: Add a new bypass.
12507 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12509 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
12511 (*subs_<optab><mode>_multp2): New pattern.
12512 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12513 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
12515 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12517 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
12518 (*subs_mul_imm_<mode>): New pattern.
12520 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
12523 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
12524 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
12525 (vsx_movti_32bit): Change j->wa to O->wa.
12527 2013-04-16 Richard Biener <rguenther@suse.de>
12529 PR rtl-optimization/56921
12530 * loop-init.c (pass_rtl_move_loop_invariants): Add
12531 TODO_do_not_ggc_collect to todo_flags_finish.
12532 (pass_rtl_unswitch): Same.
12533 (pass_rtl_unroll_and_peel_loops): Same.
12534 (pass_rtl_doloop): Same.
12536 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
12538 * config/arm/arm.c (emit_multi_reg_push): New declaration
12539 for an existing function.
12540 (arm_emit_strd_push): New function.
12541 (arm_expand_prologue): Used here.
12542 (arm_emit_ldrd_pop): New function.
12543 (arm_expand_epilogue): Used here.
12544 (arm_get_frame_offsets): Update condition.
12545 (arm_emit_multi_reg_pop): Add a special case for load of a single
12546 register with writeback.
12548 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
12550 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
12553 2013-04-16 Richard Biener <rguenther@suse.de>
12555 PR tree-optimization/56756
12556 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
12557 (first_mem_ref_loc): New.
12558 (execute_sm): Place the load temporarily before a previous
12559 access instead of in the latch edge to ensure its SSA dependencies
12560 are defined at points dominating the load.
12562 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
12564 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
12565 correct fix by moving header and footer insn to the footer of
12566 the merged basic block. Clear BB_END of the merged-away block.
12568 PR middle-end/43631
12569 * emit-rtl.c (make_note_raw): New function.
12570 (link_insn_into_chain): New static inline function.
12571 (add_insn): Use it.
12572 (add_insn_before, add_insn_after): Factor insn chain linking code...
12573 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
12574 using link_insn_into_chain.
12575 (note_outside_basic_block_p): New helper function for emit_note_after
12576 and emit_note_before.
12577 (emit_note_after): Use nobb variant of add_insn_after if the note
12578 should not be contained in a basic block.
12579 (emit_note_before): Use nobb variant of add_insn_before if the note
12580 should not be contained in a basic block.
12581 (emit_note_copy): Use make_note_raw.
12582 (emit_note): Likewise.
12583 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
12584 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
12585 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
12586 the moved barrier the tail of the basic block it follows.
12587 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
12589 2013-04-15 Jakub Jelinek <jakub@redhat.com>
12591 PR tree-optimization/56962
12592 * gimple-ssa-strength-reduction.c (record_increment): Only set
12593 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
12594 either rhs1 or rhs2 is equal to c->base_expr.
12596 2013-04-15 Richard Biener <rguenther@suse.de>
12598 PR tree-optimization/56933
12599 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
12601 (GROUP_READ_WRITE_DEPENDENCE): Remove.
12602 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
12603 * tree-vect-data-refs.c (vect_analyze_group_access): Move
12604 dependence check ...
12605 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
12607 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
12608 GROUP_READ_WRITE_DEPENDENCE.
12610 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12612 * emit-rtl.c (reset_all_used_flags): New function.
12613 (verify_rtl_sharing): Call reset_all_used_flags before and after
12614 performing the checks.
12616 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12618 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
12619 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
12620 * config/arm/constraints.md (De): New constraint.
12621 * config/arm/neon.md (anddi3_neon): Delete.
12622 (neon_vand<mode>): Expand to standard anddi3 pattern.
12623 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
12624 Move earlier in the file.
12625 (neon_inv_logic_op2): Likewise.
12626 (arm_anddi_operand_neon): New predicate.
12628 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12630 * configure.ac (gcc_cv_ld_as_needed): Set
12631 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
12632 Use -z ignore, -z record on *-*-solaris2*.
12633 (HAVE_LD_AS_NEEDED): Update comment.
12634 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
12635 * configure: Regenerate.
12636 * config.in: Regenerate.
12637 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
12638 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
12639 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
12640 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
12641 equivalents. Fix markup.
12642 * doc/tm.texi: Regenerate.
12644 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
12646 * config/i386/i386.opt: New option mstack-protector-guard=.
12647 * config/i386/i386-opts.h: Add enum stack_protector_guard.
12648 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
12649 TARGET_SSP_TLS_GUARD.
12650 * config/i386/i386.c (ix86_option_override_internal): Set
12651 ix86_stack_protector_guard.
12652 * config/i386/i386.md (stack_protect_set): Enable for
12653 TARGET_SSP_TLS_GUARD only.
12654 (stack_protect_set_<mode>): Ditto.
12655 (stack_protect_test): Ditto.
12656 (stack_protect_test_<mode>): Ditto.
12657 * doc/invoke.texi (i386 Option): Document.
12659 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
12662 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
12663 (S_MODES): Set H_MODE bit.
12664 (SF_MODES): Set only S_MODE and SF_MODE bits.
12665 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
12666 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
12667 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
12668 <MODE_FLOAT>: Likewise.
12670 2013-04-15 Joey Ye <joey.ye@arm.com>
12672 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
12674 2013-04-15 Joey Ye <joey.ye@arm.com>
12676 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
12678 (thumb_far_jump_used_p): Count instruction size and set
12681 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
12683 * reorg.c (fill_simple_delay_slots): Reindent block of code.
12684 * resource.c (mark_target_live_regs): Reformat conditional block.
12686 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12688 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
12689 notes, they are emitted only just before final.
12690 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
12692 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
12694 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
12695 * cfgrtl.c (delete_insn): Call it here instead.
12696 * lra-spills.c (lra_final_code_change): Use delete_insn.
12697 * haifa-sched.c (sched_remove_insn): Likewise.
12698 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
12699 returning to the nop pool.
12700 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
12701 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
12703 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12705 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
12706 * doc/tm.texi: Regenerated.
12708 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
12710 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
12713 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
12715 * df-core.c (df_find_def): Compare register numbers.
12716 (df_find_use): Likewise.
12718 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
12721 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
12722 lra_in_progress for return.
12724 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12726 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
12727 define_insn into define_insn_and_split and emit movsicc patterns.
12729 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
12731 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
12733 2013-04-12 Richard Biener <rguenther@suse.de>
12735 * tree-pass.h (TODO_do_not_ggc_collect): New.
12736 * passes.c (execute_one_ipa_transform_pass): Honor
12737 TODO_do_not_ggc_collect.
12738 (execute_one_pass): Likewise.
12741 2013-04-10 Richard Biener <rguenther@suse.de>
12743 * passes.c (init_optimization_passes): Remove reload pass.
12744 * ira.c (do_reload): Merge into ...
12746 (rest_of_handle_reload): Remove.
12747 (pass_reload): Likewise.
12748 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12749 of reload for vzeroupper pass placement.
12751 2013-04-12 Jakub Jelinek <jakub@redhat.com>
12753 PR tree-optimization/56918
12754 PR tree-optimization/56920
12755 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
12756 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
12757 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
12758 use wide_mul_with_sign method.
12760 2013-04-12 Richard Biener <rguenther@suse.de>
12762 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
12763 not be considered a gimple constant.
12765 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
12767 * fold-const.c (const_binop): Handle vector shifts by a scalar.
12768 (fold_binary_loc): Call const_binop also for mixed vector-scalar
12771 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
12772 Jakub Jelinek <jakub@redhat.com>
12774 * opts.c: Include diagnostic-color.h.
12775 (common_handle_option): Handle OPT_fdiagnostics_color_.
12776 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
12777 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
12778 (diagnostic-color.o): New.
12779 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
12780 (diagnostic_color_rule): New enum.
12781 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
12782 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
12783 the location string.
12784 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
12785 either NULL, or color kind.
12786 * diagnostic-color.c: New file.
12787 * diagnostic-color.h: New file.
12788 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
12790 * doc/invoke.texi (-fdiagnostics-color): Document.
12791 * pretty-print.h (pp_show_color): Define.
12792 (struct pretty_print_info): Add show_color field.
12793 * diagnostic.c: Include diagnostic-color.h.
12794 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
12795 macros. Colorize error:, warning: etc. strings and also the location
12797 (diagnostic_show_locus): Colorize the caret line.
12798 * pretty-print.c: Include diagnostic-color.h.
12799 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
12800 inside of %< %> quotes or quoted through q format modifier.
12802 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12804 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
12806 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12808 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
12809 code in CC_NZ mode.
12810 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
12813 2013-04-11 Marek Polacek <polacek@redhat.com>
12815 PR tree-optimization/48184
12816 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
12818 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
12820 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
12821 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
12822 (skip_simple_arithmetic): Tidy up.
12823 * tree.h (skip_simple_constant_arithmetic): Declare.
12825 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12827 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
12829 2013-04-11 Richard Biener <rguenther@suse.de>
12831 * tree-vect-loop.c (get_initial_def_for_induction): Properly
12832 generate vector constants.
12834 2013-04-11 Richard Biener <rguenther@suse.de>
12836 PR tree-optimization/56878
12837 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
12838 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
12840 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
12841 Prefer to align the DR with the most invariant base address.
12843 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12845 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
12848 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
12850 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
12851 floating-point vector comparisons against 0.
12853 2013-04-11 Jakub Jelinek <jakub@redhat.com>
12855 PR tree-optimization/56899
12856 * fold-const.c (extract_muldiv_1): Apply distributive law
12857 only if TYPE_OVERFLOW_WRAPS (ctype).
12859 2013-04-11 Bin Cheng <bin.cheng@arm.com>
12862 * ira-costs.c (scan_one_insn): Check whether the source rtx of
12863 loading has side effect.
12865 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
12867 * config/sparc/sparc.c: Include tree-pass.h.
12868 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
12869 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
12870 head of file. Change return type. Split off gate function.
12871 (sparc_gate_work_around_errata): New function.
12872 (pass_work_around_errata): New pass definition.
12873 (insert_pass_work_around_errata) New pass insert definition to
12874 insert pass_work_around_errata just after delayed-branch scheduling.
12875 (sparc_option_override): Insert the pass.
12876 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
12878 2013-04-10 David S. Miller <davem@davemloft.net>
12880 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
12881 or -mcpu=hypersparc.
12883 * target.def (cstore_mode): New hook.
12884 * target.h: Include insn-codes.h
12885 * targhooks.c: Likewise.
12886 (default_cstore_mode): New function.
12887 * targhooks.h: Declare it.
12888 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
12889 * doc/tm.texi: Rebuild.
12890 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
12891 target hook, rather than inspecting the insn_data.
12892 * config/sparc/sparc.c (sparc_cstore_mode): New function.
12893 (TARGET_CSTORE_MODE): Redefine.
12894 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
12896 * config/sparc/predicates.md (cstore_result_operand): New special
12898 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
12899 Use it for operand 0.
12900 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
12901 (*snesi_special): Likewise.
12902 (*snesi_zero): Likewise.
12903 (*seqsi_zero): Likewise.
12904 (*sltu_insn): Likewise.
12905 (*sgeu_insn): Likewise.
12906 (*seqdi_special): Make operand 0 and comparison operation be of
12908 (*snedi_special): Likewise.
12909 (*snedi_special_vis3): Likewise.
12910 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
12911 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
12912 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
12913 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
12914 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
12915 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
12916 (*sltu_extend_sp64): Likewise.
12917 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
12918 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
12919 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
12920 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
12921 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
12923 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
12925 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
12926 (aarch64_start_file): Use the new function.
12928 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
12929 Jason Merrill <jason@redhat.com>
12931 * common.opt: Add -gdwarf.
12932 * opts.c (common_handle_option): Handle it.
12933 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
12935 2013-04-10 Richard Biener <rguenther@suse.de>
12937 * passes.c (execute_todo): Do not call ggc_collect conditional here.
12938 (execute_one_ipa_transform_pass): But unconditionally here.
12939 (execute_one_pass): And here.
12940 (init_optimization_passes): Remove reload pass.
12941 * tree-pass.h (TODO_ggc_collect): Remove.
12942 (pass_reload): Likewise.
12943 * ira.c (do_reload): Merge into ...
12945 (rest_of_handle_reload): Remove.
12946 (pass_reload): Likewise.
12947 * config/i386/i386.c (ix86_option_override): Refer to ira instead
12948 of reload for vzeroupper pass placement.
12949 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
12950 and todo_flags_finish of all passes.
12952 2013-04-10 Richard Biener <rguenther@suse.de>
12954 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
12955 first_const_oprnd field, rename first_def_type to first_op_type.
12956 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
12957 (vect_get_and_check_slp_defs): Always use the type of the
12958 operand. Allow mixed vect_external_def, vect_constant_def types.
12959 (vect_get_constant_vectors): Handle mixed vect_external_def,
12960 vect_constant_def types.
12962 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
12964 PR tree-optimization/55524
12965 * tree-ssa-math-opts.c
12966 (convert_mult_to_fma): Don't use an fms construct
12967 when we don't have an fms operation, but fnma, and it looks
12968 likely that we'll be able to use the latter.
12970 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12972 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
12974 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
12975 inline fail caused by overwritable functions.
12977 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
12979 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
12980 unnecessary bits in the constant power of two case.
12982 2013-04-10 Richard Biener <rguenther@suse.de>
12984 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
12985 broken code swapping operands.
12986 (vect_build_slp_tree): Do not compute load permutations here.
12987 (vect_analyze_slp_instance): Compute load permutations here,
12988 after building the SLP tree.
12990 2013-04-09 Christian Bruel <christian.bruel@st.com>
12992 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
12993 of next/prev_real_insn.
12995 2013-04-09 Jan Hubicka <jh@suse.cz>
12997 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
12998 Drop aliased parameter.
12999 (function_and_variable_visibility): Do not handle alias pairs.
13000 * cgraph.c (varpool_externally_visible_p): Update prototype.
13001 * varpool.c (varpool_add_new_variable): Update.
13003 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13005 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
13007 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
13009 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
13011 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
13013 2013-04-09 Marek Polacek <polacek@redhat.com>
13015 PR tree-optimization/48762
13016 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
13018 2013-04-09 Richard Biener <rguenther@suse.de>
13020 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
13022 (vect_build_slp_tree): Likewise.
13023 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
13024 calculating the cost of a SLP instance.
13025 (vect_analyze_slp_instance): Use it from here, after building
13028 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13030 PR middle-end/56883
13031 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
13032 expand_omp_for_static_chunk): Use simple_p = true in
13033 force_gimple_operand_gsi calls when assigning to addressable decls.
13035 2013-04-09 Jeff Law <law@redhat.com>
13037 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
13038 when the boolean was created by converting a wider object which
13039 had a boolean range.
13041 2013-04-09 Richard Biener <rguenther@suse.de>
13043 * tree-vectorizer.h (slp_void_p): Remove.
13044 (slp_tree): Typedef before _slp_tree declaration.
13045 (struct _slp_tree): Use a vector of slp_tree as children.
13046 (vect_get_place_in_interleaving_chain): Remove.
13047 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
13049 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
13051 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
13052 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
13053 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
13054 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
13055 Use slp_node instead of slp_void_p and adjust.
13057 2013-04-09 Richard Biener <rguenther@suse.de>
13059 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
13060 work that is not necessary.
13062 2013-04-09 Jakub Jelinek <jakub@redhat.com>
13064 PR tree-optimization/56854
13065 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
13066 forward into clobber stmts if it would change MEM_REF lhs into
13069 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
13071 * tree.c (type_hash_lookup, type_hash_add): Make static.
13072 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
13074 2013-04-09 Richard Biener <rguenther@suse.de>
13076 * tree.h (unsave_expr_now): Remove.
13077 * tree-inline.c (mark_local_for_remap_r): Remove.
13078 (unsave_expr_1): Likewise.
13079 (unsave_r): Likewise.
13080 (unsave_expr_now): Likewise.
13081 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
13082 (propagate_tree_value): Likewise.
13084 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
13086 * doc/rtl.texi (sequence): Rewrite documentation to match the
13087 current use of SEQUENCE rtl objects.
13088 * rtl.def (SEQUENCE): Likewise.
13090 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
13091 Update documentation.
13092 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
13093 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
13095 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
13097 2013-04-08 Teresa Johnson <tejohnson@google.com>
13099 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
13100 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
13102 (estimate_edge_size_and_time): Add comment to suggest using rounding
13104 (estimate_node_size_and_time): Ditto.
13105 (remap_edge_change_prob): Use helper rounding divide methods.
13106 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
13107 (gimple_mod_pow2_value_transform): Ditto.
13108 (gimple_mod_subtract_transform): Ditto.
13109 (gimple_ic_transform): Ditto.
13110 (gimple_stringops_transform): Ditto.
13111 * stmt.c (conditional_probability): Ditto.
13112 (emit_case_dispatch_table): Ditto.
13113 * lto-cgraph.c (merge_profile_summaries): Ditto.
13114 * tree-optimize.c (execute_fixup_cfg): Ditto.
13115 * cfgcleanup.c (try_forward_edges): Ditto.
13116 * cfgloopmanip.c (scale_loop_profile): Ditto.
13118 (duplicate_loop_to_header_edge): Ditto.
13119 (lv_adjust_loop_entry_edge): Ditto.
13120 * tree-vect-loop.c (vect_transform_loop): Ditto.
13121 * profile.c (compute_branch_probabilities): Ditto.
13122 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
13123 * lto-streamer-in.c (input_cfg): Ditto.
13124 * gimple-streamer-in.c (input_bb): Ditto.
13125 * ipa-cp.c (update_profiling_info): Ditto.
13126 (update_specialized_profile): Ditto.
13127 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
13128 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
13130 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13131 (compute_trg_info): Ditto.
13132 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
13133 (purge_dead_edges): Ditto.
13134 * loop-unswitch.c (unswitch_loop): Ditto.
13135 * cgraphclones.c (cgraph_clone_edge): Ditto.
13136 (cgraph_clone_node): Ditto.
13137 * tree-inline.c (copy_bb): Ditto.
13138 (copy_edges_for_bb): Ditto.
13139 (initialize_cfun): Ditto.
13140 (copy_cfg_body): Ditto.
13141 (expand_call_inline): Ditto.
13143 2013-04-08 Kai Tietz <ktietz@redhat.com>
13145 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
13146 TARGET_CYGWIN64 by TARGET_64BIT.
13148 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13150 * config/epiphany/epiphany.md (GPR_1): New constant.
13151 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
13152 * config/epiphany/epiphany.c (gen_compare_reg):
13153 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
13154 is already in place.
13155 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
13156 Don't require being called during rtl expansion; If y operlaps r0,
13158 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
13159 (epiphany_expand_epilogue): Likewise.
13161 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
13162 Don't use CC_FPmode for ORDERED / UNORDERED.
13163 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
13165 * config/epiphany/constraints.md (CnL): New constraint.
13166 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
13167 * config/epiphany/predicates.md (add_operand): Allow 1024.
13169 * config/epiphany/epiphany.md (logical_op): New code iterator.
13170 (op_mnc): New code attribute.
13171 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
13172 (mov_f+1, mov_f+2): New peephole2 patterns.
13174 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
13175 (cstoresi4): Also allow re-use of zero result when doing a NE
13176 comparison to a non-zero operand.
13177 Use (clobber (scratch)) for first insn if the gpr output is not needed.
13179 * config/epiphany/epiphany.md (<insn_opname>v2si3):
13180 Use gen_addsi3_i / gen_subsi3_i.
13182 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13186 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
13187 contain anything but clobbers, at most one __builtin_stack_restore,
13188 optionally debug stmts and final resx, and if it has at least one
13189 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
13190 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
13191 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
13192 which isn't defaut definition, remove them.
13193 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
13194 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
13195 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
13196 with MEM_REF LHS with SSA_NAME address.
13198 2013-04-08 Jeff Law <law@redhat.com>
13200 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
13202 2013-04-08 Richard Biener <rguenther@suse.de>
13204 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
13206 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
13207 determined vector type.
13208 (vect_analyze_data_refs): Likewise.
13209 (vect_get_new_vect_var): Adjust.
13210 (vect_create_destination_var): Preserve SSA name versions.
13211 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
13212 not dump anything here.
13214 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
13216 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
13217 Add member lr_slot_known.
13218 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
13220 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
13221 Remove code that sets lr_slot_offset according to what a previous
13222 version of epiphany_emit_save_restore used to do.
13223 (epiphany_emit_save_restore): When doing an lr save or restore,
13224 set/verify lr_slot_known and lr_slot_offset.
13226 2013-04-08 Xinyu Qi <xyqi@marvell.com>
13229 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
13232 2013-04-08 Richard Biener <rguenther@suse.de>
13234 * alias.c (find_base_term): Fix thinko in previous change.
13236 2013-04-08 Jakub Jelinek <jakub@redhat.com>
13238 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
13239 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
13240 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
13241 if possible to compute val.
13242 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
13243 For QImode integers don't require anything about precision. Use
13244 const_with_all_bytes_same to find out if the constant doesn't have
13245 repeated bytes in it.
13247 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13249 * config/s390/s390.c (s390_expand_insv): Only accept insertions
13252 2013-04-08 Marek Polacek <polacek@redhat.com>
13254 PR rtl-optimization/48182
13255 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
13258 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
13261 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
13262 nuses, make sure we have a label.
13264 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13267 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
13268 (rs6000_emit_swdiv_low_precision): Remove.
13269 (rs6000_emit_swdiv): Rewrite to handle between one and four
13270 iterations of Newton-Raphson generally; modify required number of
13271 iterations for some cases.
13272 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
13274 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
13276 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
13277 set-but-unused variable.
13279 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
13280 basic blocks of released function bodies garbage-collectable.
13282 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
13283 (struct rtl_opt_pass): Add TODO_df_finish.
13285 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
13287 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13289 * config/arm/constraints.md (q): New constraint.
13290 * config/arm/ldrdstrd.md: New file.
13291 * config/arm/arm.md (ldrdstrd.md) New include.
13292 (arm_movdi): Use "q" instead of "r" constraint
13293 for double-word memory access.
13294 (movdf_soft_insn): Likewise.
13295 * config/arm/vfp.md (movdi_vfp): Likewise.
13296 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
13297 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
13298 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
13299 (mem_ok_for_ldrd_strd): Likewise.
13300 (output_move_double): Update assertion.
13302 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13304 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
13306 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13308 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
13309 define_insn_and_split.
13310 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
13312 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13314 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
13315 define_insn_and_split.
13316 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
13317 (shiftsi3_compare): New pattern.
13318 (rrx): New pattern.
13319 * config/arm/unspecs.md (UNSPEC_RRX): New.
13321 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13323 * config/arm/arm.md (negdi_extendsidi): New pattern.
13324 (negdi_zero_extendsidi): Likewise.
13326 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13328 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
13329 define_insn_and_split.
13330 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
13331 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
13333 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13335 * config/arm/arm.md (arm_subdi3): Convert define_insn into
13336 define_insn_and_split.
13337 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
13338 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
13340 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13342 * config/arm/arm.md (subsi3_carryin): New pattern.
13343 (subsi3_carryin_const): Likewise.
13344 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
13345 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
13347 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13349 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
13351 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
13353 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
13354 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
13356 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13358 * config/arm/arm.c (arm_expand_builtin): Change fcode
13359 type to unsigned int.
13361 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13363 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
13365 2013-04-04 Ian Lance Taylor <iant@google.com>
13367 * doc/standards.texi (Standards): The Go frontend supports the Go 1
13370 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
13372 PR middle-end/56729
13373 * df-scan.c (df_insn_delete): Disable failing assert.
13375 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13377 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
13378 New function prototype.
13379 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
13380 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
13381 (arm_builtin_vectorized_function): New function.
13383 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13385 * config/arm/arm_neon_builtins.def: New file.
13386 * config/arm/arm.c (neon_builtin_data): Move contents to
13387 arm_neon_builtins.def.
13388 (enum arm_builtins): Include neon builtin definitions.
13389 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
13390 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
13392 2013-04-04 Marek Polacek <polacek@redhat.com>
13394 PR tree-optimization/48186
13395 * predict.c (maybe_hot_frequency_p): Return false if
13396 HOT_BB_FREQUENCY_FRACTION is 0.
13397 (cgraph_maybe_hot_edge_p): Likewise.
13399 2013-04-04 Richard Biener <rguenther@suse.de>
13401 PR tree-optimization/56826
13402 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
13405 2013-04-04 Richard Biener <rguenther@suse.de>
13407 PR tree-optimization/56213
13408 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
13409 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
13411 2013-04-04 Richard Biener <rguenther@suse.de>
13413 PR tree-optimization/56837
13414 * tree-loop-distribution.c (classify_partition): For non-zero
13415 values require that the value has the same precision as its
13416 mode to be useful as memset value.
13418 2013-04-03 Nick Clifton <nickc@redhat.com>
13420 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
13421 (fmssf4): Use fmsf.s on E3V5 architectures.
13422 (fnmasf4): Use fnmaf.s on E3V5 architectures.
13423 (fnmssf4): Use fnmsf.s on E3V5 architectures.
13425 2013-04-03 Jeff Law <law@redhat.com>
13427 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
13428 (lra-eliminations.o): Likewise.
13430 2013-04-03 Teresa Johnson <tejohnson@google.com>
13432 * gcov-io.c (compute_working_sets): Moved most of body of old
13433 compute_working_sets here from profile.c.
13434 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
13435 (gcov_working_set_t): Moved typedef here from basic-block.h
13436 (compute_working_set): Declare.
13437 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
13438 (get_working_sets): Renamed from compute_working_set,
13439 replace most of body with call to new compute_working_sets.
13440 (get_exec_counts): Replace call to compute_working_sets
13441 to get_working_sets.
13442 * profile.h (get_working_sets): Renamed from compute_working_set.
13443 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
13444 to get_working_sets.
13445 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
13446 * gcov-dump.c (dump_working_sets): New function.
13448 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
13450 * hwint.c (sext_hwi, zext_hwi): New functions.
13451 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
13452 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
13453 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
13454 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
13455 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
13456 (sext_hwi, zext_hwi): New functions.
13458 2013-04-03 Jeff Law <law@redhat.com>
13460 PR tree-optimization/56799
13461 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
13462 back test for widening conversion erroneously dropped in prior change.
13464 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13467 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
13468 instead of next_real_insn.
13470 2013-04-03 Marek Polacek <polacek@redhat.com>
13473 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
13475 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13478 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
13480 (thumb1_output_casesi): Likewise.
13481 (thumb2_output_casesi): Likewise.
13483 2013-04-03 Richard Biener <rguenther@suse.de>
13485 PR tree-optimization/56817
13486 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
13488 (tree_unroll_loops_completely_1): ... new function to manually
13489 walk the loop tree, properly defering outer loops of unrolled
13490 loops to later iterations.
13492 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13494 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
13495 (vectorizable_load): Likewise.
13496 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
13497 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
13499 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13501 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
13504 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13506 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
13508 2013-04-03 Bin Cheng <bin.cheng@arm.com>
13510 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
13512 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13514 PR tree-optimization/56790
13515 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
13518 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
13520 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
13522 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
13523 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
13526 2013-04-03 Jakub Jelinek <jakub@redhat.com>
13529 * tree.h (force_folding_builtin_constant_p): New decl.
13530 * builtins.c (force_folding_builtin_constant_p): New variable.
13531 (fold_builtin_constant_p): Fold immediately also if
13532 force_folding_builtin_constant_p.
13534 2013-04-03 Richard Biener <rguenther@suse.de>
13536 PR tree-optimization/56812
13537 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
13538 DRs of the same interleaving chain are independent.
13540 2013-04-02 Jason Merrill <jason@redhat.com>
13542 * gdbinit.in (pbb): Use debug fn.
13544 2013-04-02 Lawrence Crowl <crowl@google.com>
13546 * sese.h (struct ivtype_map_elt_s): Remove unused.
13547 (extern debug_ivtype_map): Remove unused.
13548 (extern eq_ivtype_map_elts): Remove unused.
13549 * sese.c (debug_ivtype_map): Removed unused.
13550 (debug_ivtype_map_1): Removed unused.
13551 (debug_ivtype_elt): Remove unused.
13552 (eq_ivtype_map_elts): Remove unused.
13555 2013-04-02 Kai Tietz <ktietz@redhat.com>
13558 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
13559 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
13560 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
13562 (legitimize_pe_coff_symbol): Likewise.
13563 (is_imported_p): New helper-function.
13564 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
13565 for Windows x64 targets.
13566 (ix86_expand_prologue): Optimize for pe-coff targets.
13567 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
13568 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
13569 medium/large code-model.
13570 (legitimize_pic_address): Likewise.
13571 (legitimize_tls_address): Likewise.
13572 (ix86_expand_call): Likewise.
13573 (x86_output_mi_thunk): Likewise.
13574 (get_dllimport_decl): Add new beimport argument.
13575 (construct_plt_address): Don't assert for x64 pe-coff targets.
13576 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
13578 (SYMBOL_FLAG_STUBVAR): New macro.
13579 (SYMBOL_REF_STUBVAR_P): Likewise.
13580 * config/i386/winnt.c (stub_list): New structure.
13581 (stub_head): New local variable.
13582 (i386_pe_record_stub): New function.
13583 (i386_pe_file_end): Emit refptr-stubs.
13585 2013-04-02 Jakub Jelinek <jakub@redhat.com>
13587 PR rtl-optimization/56745
13588 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
13589 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
13592 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
13593 and both of them are MEM_REFs, just compare first argument for
13594 equality and attempt to deal even with differing offsets.
13597 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
13598 of gimple_clobber_p to be MEM_REF.
13599 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
13600 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
13601 after gimplification.
13602 * asan.c (get_mem_ref_of_assignment): Don't instrument
13603 gimple_clobber_p stmts.
13604 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
13605 gimple_clobber_p stmt if they have MEM_REF lhs and
13606 are dead because of another gimple_clobber_p stmt.
13607 * tree-ssa-live.c (clear_unused_block_pointer): Treat
13608 gimple_clobber_p stmts like debug stmts.
13609 (remove_unused_locals): Remove clobbers with MEM_REF lhs
13610 that refer to unused VAR_DECLs or uninitialized values.
13611 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
13612 gimple_clobber_p stmts if they refer to removed parameters.
13613 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
13616 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
13618 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
13619 using SWI48 mode attribute.
13621 2013-04-02 Wei Mi <wmi@google.com>
13623 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
13624 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
13625 *<rotate_insn><mode>3_mask in i386.md.
13627 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
13629 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
13631 2013-04-02 Richard Biener <rguenther@suse.de>
13633 PR tree-optimization/56778
13634 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
13635 Runtime alias tests are not supported for gather loads.
13636 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
13637 stmts referenced from SSA operands before updating SSA form.
13639 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
13640 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
13642 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
13643 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
13644 * config/arm/cortex-a53.md: New file.
13645 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
13646 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
13647 * config/arm/arm.c (arm_issue_rate): Likewise.
13648 * config/arm/arm-tune.md: Regenerate
13649 * config/arm/arm-tables.opt: Regenerate.
13650 * config/arm/arm-cores.def: Add cortex-a53.
13652 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
13654 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
13657 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13659 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
13660 scalar load/store operations using B/H registers.
13661 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
13663 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
13665 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
13667 * config/aarch64/aarch64.c
13668 (aarch64_simd_scalar_immediate_valid_for_move): New.
13669 * config/aarch64/aarch64-protos.h
13670 (aarch64_simd_scalar_immediate_valid_for_move): New.
13671 * config/aarch64/constraints.md (Dh, Dq): New.
13672 * config/aarch64/iterators.md (hq): New.
13674 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
13676 * reorg.c (get_branch_condition): Deal with conditional returns.
13677 (fill_simple_delay_slots): Remove dead code dealing with jumps.
13679 2013-04-01 Wei Mi <wmi@google.com>
13681 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
13682 Truncate operand 2 using %b asm operand modifier.
13683 (*<shift_insn><mode>3_mask): Ditto.
13684 (*<rotate_insn><mode>3_mask): Ditto.
13686 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
13688 PR middle-end/56798
13689 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
13691 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
13693 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
13695 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
13697 2013-03-30 Lawrence Crowl <crowl@google.com>
13699 * dse.c (clear_alias_sets): Remove never set.
13700 (disqualified_clear_alias_sets): Remove never set.
13701 (clear_alias_mode_pool): Remove never set.
13702 (dse_step0): Remove condition that is never true.
13703 (canon_address): Remove condition that is never true.
13704 (dse_step7): Remove condition that is never true.
13705 (rest_of_handle_dse): Remove condition that is never true.
13706 (rest_of_handle_dse::did_global): Remove never read from above.
13707 (dse_step2_spill): Remove never called from above.
13708 (dse_step5_spill): Remove never called from above.
13710 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
13712 * doc/md.texi (Standard Names) <casesi>: Update documentation for
13713 JUMP_TABLE_DATA changes.
13714 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
13715 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
13716 (Insns) <jump_table_data>: New entry.
13717 * doc/tm.texi: Regenerate.
13719 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
13721 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
13722 for table jump at the end of a basic block using tablejump_p.
13723 * targhooks.c (default_invalid_within_doloop): Likewise.
13724 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
13725 target hook implementation that is identical to the default hook.
13726 (rs6000_invalid_within_doloop): Remove.
13728 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
13729 unused variable from tablejump_p call.
13731 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
13732 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
13733 (INSN_DELETED_P): Likewise.
13734 (emit_jump_table_data): New prototype.
13735 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
13736 after 4th as unused.
13737 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
13738 * sched-vis.c (print_insn): Likewise.
13739 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
13740 insn for compatibility with back ends that use next_active_insn to
13741 identify jump table data.
13742 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
13743 (remove_insn): Likewise.
13744 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
13746 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
13747 (emit_jump_table_data): New function.
13749 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
13750 basic block, a JUMP_TABLE_DATA never is.
13751 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
13752 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
13753 off from code handling real insns.
13754 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
13755 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
13756 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
13757 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
13758 is not a NONDEBUG_INSN_P.
13759 * ira-costs.c (scan_one_insn): Likewise.
13760 * jump.c (mark_all_labels): Likewise.
13761 (mark_jump_label_1): Likewise.
13762 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13763 * lra.c (get_insn_freq): Expect all insns reaching here to be in
13765 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
13766 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
13767 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
13768 JUMP_TABLE_DATA_P insns.
13769 (calculate_elim_costs_all_insns): Likewise.
13770 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
13771 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
13772 (delete_output_reload): Code style fixups.
13773 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
13774 insn flags on this non-insn.
13775 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
13776 as scheduling barriers, for pre-change compatibility.
13777 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
13778 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
13780 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
13781 redundant JUMP_TABLE_DATA_P test.
13782 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
13783 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13784 (frv_for_each_packet): Likewise.
13785 * config/i386/i386.c (min_insn_size): Likewise.
13786 (ix86_avoid_jump_mispredicts): Likewise.
13787 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13788 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
13789 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13790 (mips16_insn_length): Robustify.
13791 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
13792 (mips16_split_long_branches): Likewise.
13793 * config/pa/pa.c (pa_combine_instructions): Likewise.
13794 * config/rs6000/rs6000.c (get_next_active_insn): Treat
13795 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
13796 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
13797 as contributing to pool range lengths.
13798 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
13799 Remove redundant JUMP_TABLE_DATA_P test.
13800 (sh_loop_align): Likewise.
13801 (split_branches): Likewise.
13802 (sh_insn_length_adjustment): Likewise.
13803 * config/spu/spu.c (get_branch_target): Likewise.
13805 2013-03-29 Jan Hubicka <jh@suse.cz>
13807 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
13808 gcov streaming; stream hot bb threshold to ltrans.
13809 * predict.c (get_hot_bb_threshold): Break out from ....
13810 (maybe_hot_count_p): ... here.
13811 (set_hot_bb_threshold): New function.
13812 * lto-section-in.c (lto_section_name): Add profile.
13813 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
13814 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
13815 and data-streamer.h
13816 (histogram_entry): New structure.
13817 (histogram, histogram_pool): New global vars.
13818 (histogram_hash): New structure.
13819 (histogram_hash::hash): New method.
13820 (histogram_hash::equal): Likewise.
13821 (account_time_size): New function.
13822 (cmp_counts): New function.
13823 (dump_histogram): New function.
13824 (ipa_profile_generate_summary): New function.
13825 (ipa_profile_write_summary): New function.
13826 (ipa_profile_read_summary): New function.
13827 (ipa_profile): Decide on threshold.
13828 (pass_ipa_profile): Add ipa_profile_write_summary and
13829 ipa_profile_read_summary.
13830 * Makefile.in (ipa.o): Update dependencies.
13831 * lto-streamer.h (LTO_section_ipa_profile): New section.
13833 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
13835 * tree.h (VAR_P): New.
13837 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
13840 * doc/invoke.texi ([-fwhole-program]): Fix typo.
13842 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
13844 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
13845 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13846 (control_flow_insn_p): Likewise.
13847 * cfgrtl.c (duplicate_insn_chain): Likewise.
13848 * final.c (get_attr_length_1): Likewise.
13849 (shorten_branches): Likewise.
13850 (final_scan_insn): Likewise.
13851 * function.c (instantiate_virtual_regs): Likewise.
13852 * gcse.c (insert_insn_end_basic_block): Likewise.
13853 * ira-costs.c (scan_one_insn): Likewise.
13854 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
13855 * lra.c (check_rtl): Likewise.
13856 * reload1.c (elimination_costs_in_insn): Likewise.
13857 * reorg.c (follow_jumps): Likewise.
13859 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
13860 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
13861 (thumb_far_jump_used_p): Likewise.
13862 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
13863 (workaround_speculation): Likewise.
13864 (add_sched_insns_for_speculation): Likewise.
13865 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
13866 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
13867 (frv_for_each_packet): Likewise.
13868 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
13869 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
13870 (final_emit_insn_group_barriers): Likewise.
13871 * config/m32r/m32r.c (m32r_is_insn): Likewise.
13872 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
13873 (mips16_insn_length): Likewise.
13874 * config/pa/pa.c (pa_reorg): Likewise.
13875 (pa_combine_instructions): Likewise.
13876 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
13877 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
13878 (sh_reorg): Likewise.
13879 (split_branches): Likewise.
13880 * config/spu/spu.c (get_branch_target): Likewise.
13882 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
13885 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
13887 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
13888 Fix declaration name.
13890 2013-03-28 Lawrence Crowl <crowl@google.com>
13892 * graphds.h (struct graph.indicies): Remove unused.
13893 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
13894 (SCOP_ORIGINAL_PDDRS): Remove unused.
13895 * sese.h (extern insert_loop_close_phis): Removed unused.
13896 (extern insert_guard_phis): Removed unused.
13897 (extern ivtype_map_elt_info): Removed unused.
13898 (new_ivtype_map_elt): Removed unused.
13899 * sese.c (ivtype_map_elt_info): Removed unused.
13901 2013-03-28 Lawrence Crowl <crowl@google.com>
13903 * Makefile.in: Add several missing include dependences.
13905 (test-dump.o): New. This object is not added to any executable,
13906 but is present for ad-hoc testing.
13908 (debug (const bitmap_head_def &)): New.
13909 (debug (const bitmap_head_def *)): New.
13911 (extern debug (const bitmap_head_def &)): New.
13912 (extern debug (const bitmap_head_def *)): New.
13914 (debug (edge_def &)): New.
13915 (debug (edge_def *)): New.
13917 (debug (basic_block_def &)): New.
13918 (debug (basic_block_def *)): New.
13920 (dump_node (const_tree, int, FILE *)): Correct source file.
13922 (debug (die_struct &)): New.
13923 (debug (die_struct *)): New.
13925 (extern debug (die_struct &)): New.
13926 (extern debug (die_struct *)): New.
13927 * gimple-pretty-print.c
13928 (debug (gimple_statement_d &)): New.
13929 (debug (gimple_statement_d *)): New.
13930 * gimple-pretty-print.h
13931 (extern debug (gimple_statement_d &)): New.
13932 (extern debug (gimple_statement_d *)): New.
13934 (debug (ira_allocno_copy &)): New.
13935 (debug (ira_allocno_copy *)): New.
13936 (debug (ira_allocno &)): New.
13937 (debug (ira_allocno *)): New.
13939 (extern debug (ira_allocno_copy &)): New.
13940 (extern debug (ira_allocno_copy *)): New.
13941 (extern debug (ira_allocno &)): New.
13942 (extern debug (ira_allocno *)): New.
13944 (debug (live_range &)): New.
13945 (debug (live_range *)): New.
13947 (debug (lra_live_range &)): New.
13948 (debug (lra_live_range *)): New.
13950 (debug (lra_live_range &)): New.
13951 (debug (lra_live_range *)): New.
13953 (debug (omega_pb_d &)): New.
13954 (debug (omega_pb_d *)): New.
13956 (extern debug (omega_pb_d &)): New.
13957 (extern debug (omega_pb_d *)): New.
13959 (debug (const rtx_def &)): New.
13960 (debug (const rtx_def *)): New.
13962 (debug_tree (tree): Move within file.
13963 (debug_raw (const tree_node &)): New.
13964 (debug_raw (const tree_node *)): New.
13965 (dump_tree_via_hooks (const tree_node *, int)): New.
13966 (debug (const tree_node &)): New.
13967 (debug (const tree_node *)): New.
13968 (debug_verbose (const tree_node &)): New.
13969 (debug_verbose (const tree_node *)): New.
13970 (debug_head (const tree_node &)): New.
13971 (debug_head (const tree_node *)): New.
13972 (debug_body (const tree_node &)): New.
13973 (debug_body (const tree_node *)): New.
13974 (debug_vec_tree (tree): Move and reimplement in terms of dump.
13975 (debug (vec<tree, va_gc> &)): New.
13976 (debug (vec<tree, va_gc> *)): New.
13978 (extern debug (const rtx_def &)): New.
13979 (extern debug (const rtx_def *)): New.
13981 (debug_raw (simple_bitmap_def &)): New.
13982 (debug_raw (simple_bitmap_def *)): New.
13983 (debug (simple_bitmap_def &)): New.
13984 (debug (simple_bitmap_def *)): New.
13986 (extern debug (simple_bitmap_def &)): New.
13987 (extern debug (simple_bitmap_def *)): New.
13988 (extern debug_raw (simple_bitmap_def &)): New.
13989 (extern debug_raw (simple_bitmap_def *)): New.
13991 (debug (vinsn_def &)): New.
13992 (debug (vinsn_def *)): New.
13993 (debug_verbose (vinsn_def &)): New.
13994 (debug_verbose (vinsn_def *)): New.
13995 (debug (expr_def &)): New.
13996 (debug (expr_def *)): New.
13997 (debug_verbose (expr_def &)): New.
13998 (debug_verbose (expr_def *)): New.
13999 (debug (vec<rtx> &)): New.
14000 (debug (vec<rtx> *)): New.
14002 (extern debug (vinsn_def &)): New.
14003 (extern debug (vinsn_def *)): New.
14004 (extern debug_verbose (vinsn_def &)): New.
14005 (extern debug_verbose (vinsn_def *)): New.
14006 (extern debug (expr_def &)): New.
14007 (extern debug (expr_def *)): New.
14008 (extern debug_verbose (expr_def &)): New.
14009 (extern debug_verbose (expr_def *)): New.
14010 (extern debug (vec<rtx> &)): New.
14011 (extern debug (vec<rtx> *)): New.
14013 (_list_iter_cond_expr): Make inline instead of static.
14015 (debug (sreal &)): New.
14016 (debug (sreal *)): New.
14018 (extern debug (sreal &)): New.
14019 (extern debug (sreal *)): New.
14021 (extern debug_raw (const tree_node &)): New.
14022 (extern debug_raw (const tree_node *)): New.
14023 (extern debug (const tree_node &)): New.
14024 (extern debug (const tree_node *)): New.
14025 (extern debug_verbose (const tree_node &)): New.
14026 (extern debug_verbose (const tree_node *)): New.
14027 (extern debug_head (const tree_node &)): New.
14028 (extern debug_head (const tree_node *)): New.
14029 (extern debug_body (const tree_node &)): New.
14030 (extern debug_body (const tree_node *)): New.
14031 (extern debug (vec<tree, va_gc> &)): New.
14032 (extern debug (vec<tree, va_gc> *)): New.
14034 (debug (struct loop &)): New.
14035 (debug (struct loop *)): New.
14036 (debug_verbose (struct loop &)): New.
14037 (debug_verbose (struct loop *)): New.
14038 * tree-dump.c: Add header dependence.
14040 (extern debug (struct loop &)): New.
14041 (extern debug (struct loop *)): New.
14042 (extern debug_verbose (struct loop &)): New.
14043 (extern debug_verbose (struct loop *)): New.
14045 (debug (data_reference &)): New.
14046 (debug (data_reference *)): New.
14047 (debug (vec<data_reference_p> &)): New.
14048 (debug (vec<data_reference_p> *)): New.
14049 (debug (vec<ddr_p> &)): New.
14050 (debug (vec<ddr_p> *)): New.
14052 (extern debug (data_reference &)): New.
14053 (extern debug (data_reference *)): New.
14054 (extern debug (vec<data_reference_p> &)): New.
14055 (extern debug (vec<data_reference_p> *)): New.
14056 (extern debug (vec<ddr_p> &)): New.
14057 (extern debug (vec<ddr_p> *)): New.
14059 (debug (pt_solution &)): New.
14060 (debug (pt_solution *)): New.
14062 (extern debug (pt_solution &)): New.
14063 (extern debug (pt_solution *)): New.
14065 (debug (_var_map &)): New.
14066 (debug (_var_map *)): New.
14067 (debug (tree_live_info_d &)): New.
14068 (debug (tree_live_info_d *)): New.
14070 (extern debug (_var_map &)): New.
14071 (extern debug (_var_map *)): New.
14072 (extern debug (tree_live_info_d &)): New.
14073 (extern debug (tree_live_info_d *)): New.
14075 2013-03-28 Jan Hubicka <jh@suse.cz>
14077 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
14079 2013-03-28 Ian Bolton <ian.bolton@arm.com>
14081 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
14082 record only when desired or required.
14084 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
14086 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
14087 *vec_extractv2di_1_rex64. Use x64 isa attribute.
14089 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14091 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
14092 (*andsi3_compare0_uxtw): New pattern.
14093 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
14094 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
14096 2013-03-28 Jan Hubicka <jh@suse.cz>
14098 * data-streamer-in.c (streamer_read_gcov_count): New function.
14099 * gimple-streamer-out.c: Include value-prof.h.
14100 (output_gimple_stmt): Output histogram.
14101 (output_bb): Use streamer_write_gcov_count.
14102 * value-prof.c: Include data-streamer.h
14103 (dump_histogram_value): Add HIST_TYPE_MAX.
14104 (stream_out_histogram_value): New function.
14105 (stream_in_histogram_value): New function.
14106 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
14107 (stream_out_histogram_value, stream_in_histogram_value): Declare.
14108 * data-streamer-out.c (streamer_write_gcov_count): New function.
14109 (streamer_write_gcov_count_stream): New function.
14110 * lto-cgraph.c (lto_output_edge): Update counter streaming.
14111 (lto_output_node): Likewise.
14112 (input_node, input_edge): Likewise.
14113 * lto-streamer-out.c (output_cfg): Update streaming.
14114 * lto-streamer-in.c (input_cfg): Likewise.
14115 * data-streamer.h (streamer_write_gcov_count,
14116 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
14117 * gimple-streamer-in.c: Include value-prof.h
14118 (input_gimple_stmt): Input histograms.
14119 (input_bb): Update profile streaming.
14121 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14123 * genmodes.c (emit_max_int): New function.
14124 (emit_insn_modes_h): Added call to emit_max_function.
14125 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
14127 * machmode.def: Fixed comment.
14129 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
14131 * combine.c (try_combine): Removed useless assert.
14132 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
14134 2013-03-28 Marek Polacek <polacek@redhat.com>
14135 Richard Biener <rguenther@suse.de>
14137 PR tree-optimization/56695
14138 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
14139 build signed result of a vector comparison.
14140 * tree-cfg.c (verify_gimple_comparison): Check that a result
14141 of a vector comparison has signed type.
14143 2013-03-28 Richard Biener <rguenther@suse.de>
14145 PR tree-optimization/37021
14146 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
14147 do not restrict gaps between groups.
14148 * tree-vect-stmts.c (vectorizable_load): Properly account for
14149 a gap between groups.
14151 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
14153 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
14154 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
14157 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
14159 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
14160 * doc/extend.texi (Named Address Spaces): Ditto.
14161 (Variable Attributes): Ditto.
14163 2013-03-27 Kai Tietz <ktietz@redhat.com>
14165 * config.build: Add support for cygwin x64 target.
14166 * config.gcc: Likewise.
14167 * config.host: Likewise.
14168 * configure.ac: Likewise
14169 * configure: Regenerated.
14171 2013-03-27 Kai Tietz <ktietz@redhat.com>
14173 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
14174 * config/i386/t-cygwin-w64: New file.
14175 * config/i386/cygwin-w64.h: New file.
14176 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
14177 and add support for x64-cygwin target.
14178 (CPP_SPEC): Likewise.
14179 (CXX_WRAP_SPEC_LIST): Undefine before define.
14180 (LIBGCJ_SONAME): Use 15 as version.
14182 2013-03-27 Richard Biener <rguenther@suse.de>
14184 PR tree-optimization/56716
14185 * tree-ssa-structalias.c (perform_var_substitution): Adjust
14186 dumping for ref nodes.
14188 2013-03-27 Martin Jambor <mjambor@suse.cz>
14190 PR tree-optimization/55334
14191 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
14192 restricted pointers to arrays.
14194 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
14196 * Makefile.in (.SUFFIXES): Add .cc.
14197 (.c.o): Apply same recipe for implicit rule .cc.o.
14199 2013-03-27 Richard Biener <rguenther@suse.de>
14201 PR tree-optimization/37021
14202 * tree-vect-data-refs.c (vect_check_strided_load): Allow
14203 REALPART/IMAGPART_EXPRs around the supported refs.
14204 * tree-ssa-structalias.c (find_func_aliases): Assume that
14205 floating-point values are not used to transfer pointers.
14207 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
14209 * gcc/target.def (TARGET_HAS_IFUNC_P): New target hook.
14210 * gcc/doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
14211 * gcc/doc/tm.texi: Regenerate.
14212 * gcc/targhooks.h (default_has_ifunc_p): New.
14213 * gcc/targhooks.c (default_has_ifunc_p): Ditto.
14214 * gcc/config/linux-protos.h: New file.
14215 * gcc/config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of
14216 this hook for linux which disables support of indirect functions in
14218 * gcc/config/linux-android.c: New file.
14219 * gcc/config/t-linux-android.c: Ditto.
14220 * gcc/config.gcc: Added new object file linux-android.o.
14221 * gcc/config/i386/i386.c (ix86_get_function_versions_dispatcher):
14222 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
14223 * gcc/varasm.c (do_assemble_alias): Likewise.
14224 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
14225 doesn't support indirect functions.
14226 * configure: Regenerate.
14228 2013-03-27 Bin Cheng <bin.cheng@arm.com>
14231 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
14232 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
14235 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14237 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
14239 2013-03-27 Terry Guo <terry.guo@arm.com>
14241 * config/arm/arm-cores.def: Added core cortex-r7.
14242 * config/arm/arm-tune.md: Regenerated.
14243 * config/arm/arm-tables.opt: Regenerated.
14244 * doc/invoke.texi: Added entry for core cortex-r7.
14246 2013-03-27 Walter Lee <walt@tilera.com>
14248 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
14249 double-decrement of next_scratch_regno.
14251 2013-03-27 Walter Lee <walt@tilera.com>
14253 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
14255 (insn_v1mulus): Ditto.
14256 (insn_v2muls): Ditto.
14258 2013-03-27 Walter Lee <walt@tilera.com>
14260 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
14261 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
14263 2013-03-27 Walter Lee <walt@tilera.com>
14265 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
14266 (*sibcall_value): Ditto.
14268 2013-03-27 Walter Lee <walt@tilera.com>
14270 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
14271 (insn_mnz_v8qi): ... this ...
14272 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
14274 (insn_v<n>mnz): Replaced by ...
14275 (insn_v1mnz): ... this ...
14276 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
14278 (insn_mz_<mode>): Replaced by ...
14279 (insn_mz_v8qi): ... this ...
14280 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
14282 (insn_v<n>mz): Replaced by ...
14283 (insn_v1mz): ... this ...
14284 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
14287 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
14289 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
14291 2013-03-26 Roland McGrath <mcgrathr@google.com>
14293 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
14294 than fprintf with a non-constant, non-format string.
14296 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14298 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
14299 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
14300 operand 0 predicate.
14301 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
14302 attribute. Use general_x64nomem_operand as operand 1 predicate.
14303 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
14304 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14305 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
14306 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
14307 (mov<mode>_insv_1): Remove expander. Merge insn with
14308 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
14309 Use general_x64nomem_operand as operand 1 predicate.
14310 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
14311 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
14312 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14313 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
14314 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
14315 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
14316 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
14317 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
14318 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
14319 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
14320 (general_x64nomem_operand): Ditto.
14322 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
14324 * config/rtems.opt: Add -pthread option.
14326 2013-03-26 Richard Biener <rguenther@suse.de>
14328 * alias.c (find_base_term): Avoid redundant and not used recursion.
14329 (base_alias_check): Get the initial base term from the caller.
14330 (true_dependence_1): Compute and pass base terms to base_alias_check.
14331 (write_dependence_p): Likewise.
14332 (may_alias_p): Likewise.
14334 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
14336 * config/aarch64/aarch64.c (aarch64_classify_address): Support
14337 PC-relative load in SI modes and above only.
14339 2013-03-26 Xinyu Qi <xyqi@marvell.com>
14341 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
14342 * config/arm/iwmmxt.md (WCGR0): Update.
14343 (WCGR1, WCGR2, WCGR3): Likewise.
14345 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
14347 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
14348 Use x64 and nox64 isa attributes.
14350 2013-03-26 Richard Biener <rguenther@suse.de>
14352 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
14353 alignment computations and rely on get_object_alignment_1
14354 for the !TYPE_P case.
14355 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
14357 2013-03-26 Walter Lee <walt@tilera.com>
14359 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
14360 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
14362 2013-03-25 Jeff Law <law@redhat.com>
14364 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
14365 check for INTEGRAL_TYPE_P that was missing due to checking in
14366 wrong version of prior patch.
14368 2013-03-25 Walter Lee <walt@tilera.com>
14370 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
14371 TILEGX_INSN_SHUFFLEBYTES1.
14372 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
14374 (tilegx_builtins): Ditto.
14375 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
14377 2013-03-25 Walter Lee <walt@tilera.com>
14379 * config/tilegx/tilegx.md (floatsisf2): New pattern.
14380 (floatunssisf2): New pattern.
14381 (floatsidf2): New pattern.
14382 (floatunssidf2): New pattern.
14384 2013-03-25 Walter Lee <walt@tilera.com>
14386 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
14387 tests for constraint J, K, N, P.
14389 2013-03-25 Walter Lee <walt@tilera.com>
14391 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
14392 Use indirect/pcrel encoding.
14393 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
14396 2013-03-25 Steve Ellcey <sellcey@mips.com>
14398 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
14399 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
14400 * config/mips/mips.c (mips_option_override): Set IMADD default.
14401 * config/mips/mips.h (PTF_AVOID_IMADD): New.
14402 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
14403 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
14404 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
14405 * doc/invoke.texi (-mimadd/-mno-imadd): New.
14407 2013-03-25 Jeff Law <law@redhat.com>
14409 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
14410 slightly to avoid creating and folding useless trees. Simplify
14411 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
14413 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
14415 * config/i386/i386.md (*zero_extendsidi2): Merge with
14416 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
14417 * config/i386/predicates.md (x86_64_zext_operand): Rename from
14418 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
14419 targets. Clarify comment.
14421 2013-03-25 Martin Jambor <mjambor@suse.cz>
14423 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
14424 pass-through jump functions differently.
14425 (ipa_read_jump_function): Likewise. Also use setter functions to set
14428 2013-03-25 Martin Jambor <mjambor@suse.cz>
14430 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
14431 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
14433 (ipa_get_indirect_edge_target): New function.
14434 (devirtualization_time_bonus): New parameter known_aggs, pass it to
14435 ipa_get_indirect_edge_target. Update all callers.
14436 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
14437 ipa_get_indirect_edge_target_1 instead of calling
14438 ipa_get_indirect_edge_target.
14439 (create_specialized_node): Pass aggvlas to
14440 ipcp_discover_new_direct_edges.
14442 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14444 * config/arm/arm.md (f_sels, f_seld): New types.
14445 (*cmov<mode>): New pattern.
14446 * config/arm/predicates.md (arm_vsel_comparison_operator): New
14449 2013-03-25 Kai Tietz <ktietz@redhat.com>
14451 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
14452 POSIX-printf for mingw-hosted builds.
14454 2013-03-25 Richard Biener <rguenther@suse.de>
14456 PR middle-end/56694
14457 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
14458 must-not-throw stmt location.
14460 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14462 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
14463 Emit load-acquire versions when acq is true.
14464 (arm_emit_store_exclusive): Add rel parameter.
14465 Emit store-release versions when rel is true.
14466 (arm_split_compare_and_swap): Use acquire-release instructions
14468 of barriers when appropriate.
14469 (arm_split_atomic_op): Likewise.
14470 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
14471 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
14472 (VUNSPEC_SLX): Likewise.
14473 (VUNSPEC_LDA): Likewise.
14474 (VUNSPEC_STL): Likewise.
14475 * config/arm/sync.md (atomic_load<mode>): New pattern.
14476 (atomic_store<mode>): Likewise.
14477 (arm_load_acquire_exclusive<mode>): Likewise.
14478 (arm_load_acquire_exclusivesi): Likewise.
14479 (arm_load_acquire_exclusivedi): Likewise.
14480 (arm_store_release_exclusive<mode>): Likewise.
14482 2013-03-25 Catherine Moore <clm@codesourcery.com>
14484 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
14485 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
14486 * config/mip/predicates.md (lwsp_swsp_operand,
14487 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
14488 sb16_operand, db4_operand, db7_operand, ib3_operand,
14489 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
14490 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
14491 andi16_operand): New predicates.
14492 * config/mips/mips.md (compression): New attribute.
14493 (enabled): New attribute.
14494 (length): Consider compression in computing length.
14495 (shift_compression): New code attribute.
14496 (*add<mode>3): New operands. Record compression.
14497 (sub<mode>3): Likewise.
14498 (one_cmpl<mode>2): Likewise.
14499 (*and<mode>3): Likewise.
14500 (*ior<mode>3): Likewise.
14501 (unnamed pattern for xor): Likewise.
14502 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
14503 (*<optab><mode>3): Likewise.
14504 (*mov<mode>_internal: Likewise.
14505 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
14506 (mips_unsigned_immediate_p): New.
14507 (umips_lwsp_swsp_address_p): New.
14508 (m16_based_address_p): New.
14509 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
14510 (mips_unsigned_immediate_p): New prototype.
14511 (lwsp_swsp_address_p): New prototype.
14512 (m16_based_address_p): New prototype.
14513 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
14514 (mips_signed_immediate_p): New function.
14515 (m16_based_address_p): New function.
14516 (lwsp_swsp_address_p): New function.
14517 (mips_print_operand_punctuation): Recognize short delay slot insns
14518 for microMIPS.add<mode>3"
14520 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14523 * config/arm/iterators.md (v_cmp_result): New mode attribute.
14524 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
14526 2013-03-25 Richard Biener <rguenther@suse.de>
14528 PR tree-optimization/56689
14529 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
14532 2013-03-25 Richard Biener <rguenther@suse.de>
14534 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
14536 (memory_references): Likewise.
14537 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
14538 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
14539 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
14540 (gather_mem_refs_in_loops): Fold into ...
14541 (analyze_memory_references): ... this. Move initialization
14542 to tree_ssa_lim_initialize.
14543 (fill_always_executed_in): Rename to ...
14544 (fill_always_executed_in_1): ... this.
14545 (fill_always_executed_in): Move contains_call computation to
14546 this new function from ...
14547 (tree_ssa_lim_initialize): ... here.
14548 (tree_ssa_lim): Call fill_always_executed_in.
14550 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14552 * postreload.c (reload_combine): Fix code detecting returns.
14554 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
14556 * function.c (emit_use_return_register_into_block): On cc0 targets,
14557 do not emit the sequence between cc0 setter and user.
14559 2013-03-25 Kai Tietz <ktietz@redhat.com>
14561 * config/i386/predicates.md (local_symbolic_operand): Interpret
14562 dll-imported symbols as none-local.
14564 2013-03-25 Richard Biener <rguenther@suse.de>
14566 * tree-ssa-loop-im.c (struct depend): Remove.
14567 (struct lim_aux_data): Make depends a vec of gimples.
14568 (free_lim_aux_data): Adjust.
14569 (add_dependency): Likewise.
14570 (set_level): Likewise.
14572 2013-03-25 Richard Biener <rguenther@suse.de>
14574 PR middle-end/56434
14575 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
14576 the pointer returned by calls with ECF_MALLOC set.
14578 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14580 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
14582 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
14584 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
14585 using MMXMODE mode iterator.
14586 (*move<mode>_internal): Merge with *movv2sf_internal and
14587 *movv2sf_internal_rex64 using MMXMODE mode iterator.
14589 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14591 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
14592 (record_last_mem_set_info): Likewise.
14594 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
14595 of XNEWVEC followed by memset.
14596 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
14598 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
14600 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
14601 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
14602 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
14603 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
14604 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
14605 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
14606 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
14607 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
14608 BARRIER_P instead of GET_CODE.
14610 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14612 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
14613 inaccuracy in the probing code.
14615 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
14616 (ctrapdi4): Likewise.
14618 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14620 * calls.c (expand_call): Add missing guard to code handling return
14621 of non-BLKmode structures in MSB.
14622 * function.c (expand_function_end): Likewise.
14624 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
14626 * combine.c (try_combine): Adjust comment. Do not add the set of
14627 insn #0 if the destination indirectly is set or dies in insn #2.
14628 Tidy up code to distribute a new note.
14630 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14632 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
14633 also for alternatives 16 and 17.
14635 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
14637 * config/i386/sse.md (*mov<mode>_internal): Merge with
14638 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
14639 Emit insn template depending on type attribute. Use
14640 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14641 movd instead of movq mnemonic for interunit moves. Rewrite mode
14642 attribute calculation. Remove unit attribute calculation.
14643 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
14644 Set prefix_data16 attribute for DImode ssemov types.
14645 Use Ym instead of y for SSE-MMX conversion alternatives.
14646 Reorder operand constraints.
14648 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
14650 * df.h (df_insn_delete): Adjust prototype.
14651 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
14652 and let it decide whether mark the basic block dirty.
14653 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
14654 * df-scan.c (df_insn_info_delete): New helper function, split
14655 off from df_insn_delete.
14656 (df_scan_free_bb_info): Use it.
14657 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
14659 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
14660 that the insn is actually an insn and it has a non-NULL basic block.
14661 Do not mark basic block dirty if only deleting a DEBUG_INSN.
14663 2013-03-22 Richard Biener <rguenther@suse.de>
14665 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
14667 (mem_ref_alloc): Do not allocate them.
14668 (refs_independent_p): Do not query or maintain a cache.
14670 2013-03-22 Richard Biener <rguenther@suse.de>
14672 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
14673 (gather_mem_refs_in_loops): Do not compute it.
14674 (analyze_memory_references): Do not allocate it.
14675 (tree_ssa_lim_finalize): Do not free it.
14676 (for_all_locs_in_loop): Do not query all_refs_in_loop.
14678 2013-03-22 Richard Biener <rguenther@suse.de>
14680 * is-a.h (as_a): Use gcc_checking_assert.
14682 2013-03-22 Ian Bolton <ian.bolton@arm.com>
14684 * config/aarch64/aarch64.c (aarch64_print_operand): New
14685 format specifier for printing a constant in hex.
14686 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
14687 format specifier for printing second operand.
14689 2013-03-22 Richard Biener <rguenther@suse.de>
14691 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
14693 (gather_mem_refs_in_loops): Perform store accumulation here.
14694 (create_vop_ref_mapping_loop): Remove.
14695 (create_vop_ref_mapping): Likewise.
14696 (analyze_memory_references): Initialize refs_stored_in_loop.
14697 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
14698 (record_indep_loop): Remove.
14699 (record_dep_loop): New function.
14700 (ref_indep_loop_p_1): Adjust to only walk over references
14701 in the loop, not its subloops.
14702 (ref_indep_loop_p): Rename to ...
14703 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
14704 maintaining a more fine-grained cache.
14705 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
14706 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
14708 2013-03-22 Richard Biener <rguenther@suse.de>
14710 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
14711 (struct mem_ref): Make accesses_in_loop a vec of a vec of
14712 aggregate mem_ref_loc.
14713 (free_mem_ref_locs): Inline into ...
14714 (memref_free): ... this and adjust.
14715 (mem_ref_alloc): Adjust.
14716 (mem_ref_locs_alloc): Remove.
14717 (record_mem_ref_loc): Adjust.
14718 (get_all_locs_in_loop): Rewrite into ...
14719 (for_all_locs_in_loop): ... this iterator.
14720 (rewrite_mem_ref_loc): New functor.
14721 (rewrite_mem_refs): Use for_all_locs_in_loop.
14722 (sm_set_flag_if_changed): New functor.
14723 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
14724 (ref_always_accessed): New functor.
14725 (ref_always_accessed_p): Use for_all_locs_in_loop.
14727 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14729 * tree-pass.h (PROP_gimple_lvec): New.
14730 * passes.c (dump_properties): Handle PROP_gimple_lvec.
14731 (init_optimization_passes): Move pass_lower_vector.
14732 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
14734 (pass_lower_vector): Provide PROP_gimple_lvec.
14735 (pass_lower_vector_ssa): Likewise.
14736 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
14738 2013-03-21 Mark Wielaard <mjw@redhat.com>
14740 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
14742 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14744 * config/i386/i386.md (*movdi_internal): Disparage slightly
14745 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
14746 conversion alternatives.
14748 2013-03-21 Jakub Jelinek <jakub@redhat.com>
14750 PR middle-end/48087
14751 * diagnostic.def (DK_WERROR): New kind.
14752 * diagnostic.h (werrorcount): Define.
14753 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
14754 promoted to DK_ERROR, increment DK_WERROR counter instead of
14756 * toplev.c (toplev_main): Call print_ignored_options even if
14757 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
14758 even if just werrorcount is non-zero.
14761 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
14763 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
14764 (string_cst_pool_decl): New function.
14765 (optimize_one_addr_into_implicit_ptr): New function.
14766 (resolve_addr_in_expr): Optimize DWARF location expression
14767 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
14768 which doesn't live in memory, but has DW_AT_location or
14769 DW_AT_const_value, or refers to a string literal, into
14770 DW_OP_GNU_implicit_pointer.
14771 (optimize_location_into_implicit_ptr): New function.
14772 (resolve_addr): If removing DW_AT_location of a variable because
14773 it was DW_OP_addr of address of the variable, but the variable doesn't
14774 live in memory, try to emit const value attribute for the initializer.
14776 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14778 * tree.h (VECTOR_TYPE_P): New macro.
14779 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
14780 TYPE_MODE): Use it.
14781 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
14782 VEC_COND_EXPR cannot be lvalues.
14783 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
14785 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
14787 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
14788 Restrict the transformation to equal modes.
14790 2013-03-21 Richard Biener <rguenther@suse.de>
14792 PR tree-optimization/39326
14793 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
14794 (MEM_ANALYZABLE): Adjust.
14795 (record_mem_ref_loc): Move bitmap ops ...
14796 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
14797 unanalyzable refs, do not record locations for it.
14798 (analyze_memory_references): Allocate ref zero as shared
14800 (refs_independent_p): Do not test for unanalyzed mems here.
14801 (ref_indep_loop_p_1): Special-case disambiguation against
14802 the unanalyzed ref.
14803 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
14805 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
14807 * config/arm/arm-protos.h (tune_params): Add
14808 prefer_neon_for_64bits field.
14809 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
14810 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
14811 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
14812 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
14813 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
14814 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
14815 (arm_option_override): Handle -mneon-for-64bits new option.
14816 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
14817 (prefer_neon_for_64bits): Declare new variable.
14818 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
14819 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
14820 (arch_enabled): Handle new arch types. Remove support for onlya8
14822 (one_cmpldi2): Use new arch names.
14823 (zero_extend<mode>di2, extend<mode>di2): Ditto.
14824 * config/arm/arm.opt (mneon-for-64bits): Add option.
14825 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
14826 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
14827 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
14829 * doc/invoke.texi (-mneon-for-64bits): Document.
14831 2013-03-21 Richard Biener <rguenther@suse.de>
14833 PR tree-optimization/39326
14834 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
14835 (sort_bbs_in_loop_postorder_cmp): New function.
14836 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
14838 2013-03-21 Richard Biener <rguenther@suse.de>
14840 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
14841 (vect_insert_into_interleaving_chain): Likewise.
14842 (vect_drs_dependent_in_basic_block): Inline ...
14843 (vect_slp_analyze_data_ref_dependence): ... here. New function,
14845 (vect_analyze_data_ref_dependence): ... here. Simplify.
14846 (vect_check_interleaving): Simplify.
14847 (vect_analyze_data_ref_dependences): Likewise. Split out ...
14848 (vect_slp_analyze_data_ref_dependences): ... this new function.
14849 (dr_group_sort_cmp): New function.
14850 (vect_analyze_data_ref_accesses): Compute data-reference groups
14851 here instead of in vect_analyze_data_ref_dependence. Use
14852 a more efficient algorithm.
14853 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
14854 vect_slp_analyze_data_ref_dependences. Call
14855 vect_analyze_data_ref_accesses earlier.
14856 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
14857 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
14858 (vect_slp_analyze_data_ref_dependences): New prototype.
14860 2013-03-21 Richard Biener <rguenther@suse.de>
14862 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
14863 ref is stored in the loop.
14864 (find_refs_for_sm): Walk only over all stores.
14865 (store_motion_loop): Allocate from lim_bitmap_obstack.
14866 (store_motion): Likewise.
14868 2013-03-21 Richard Biener <rguenther@suse.de>
14870 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
14871 Update virtual SSA form.
14873 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14875 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
14876 * configure: Regenerate.
14877 * config.in: Regenerate.
14878 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
14879 if !HAVE_LD_EH_FRAME_CIEV3.
14881 2013-03-21 Richard Biener <rguenther@suse.de>
14883 * tree-cfg.c (verify_expr_no_block): New function.
14884 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
14885 nor DECL_VALUE_EXPR have locations with associated blocks.
14886 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
14887 (clear_unused_block_pointer): Remove code dealing with
14888 blocks in DECL_DEBUG_EXPR locations.
14890 2013-03-21 Richard Biener <rguenther@suse.de>
14892 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
14893 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
14894 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
14895 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
14896 instead of DECL_DEBUG_EXPR_IS_FROM.
14897 * gimplify.c (gimplify_modify_expr): Likewise.
14898 * tree-cfg.c (verify_expr_location_1): Likewise.
14899 * tree-complex.c (create_one_component_var): Likewise.
14900 * tree-sra.c (create_access_replacement): Likewise.
14901 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
14902 (clear_unused_block_pointer): Likewise.
14903 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
14904 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
14905 * var-tracking.c (var_debug_decl): Likewise.
14906 (track_expr_p): Likewise.
14907 * tree-inline.c (add_local_variables): Likewise. Set
14908 DECL_HAS_DEBUG_EXPR_P after copying it.
14909 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
14910 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
14912 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
14915 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
14916 * configure: Regenerate.
14917 * config.in: Regenerate.
14918 * config/i386/i386.md (*movdf_internal): Use
14919 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
14920 movd instead of movq mnemonic for interunit moves.
14921 (*movdi_internal): Ditto.
14923 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14925 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
14926 (abd<mode>_3): New pattern.
14927 (aba<mode>_3): New pattern.
14928 (fabd<mode>_3): New pattern.
14930 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14932 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
14933 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
14934 occurrence of REGISTER_PREFIX as its empty string.
14936 2013-03-20 Jeff Law <law@redhat.com>
14938 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
14939 addititional equivalences for equality comparisons between an SSA_NAME
14940 and a constant where the SSA_NAME was set from a widening conversion.
14942 2013-03-20 Walter Lee <walt@tilera.com>
14944 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
14946 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14948 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
14949 depending on type attribute.
14950 (*movti_internal): Ditto.
14951 (*movtf_internal): Ditto.
14952 (*movxf_internal): Ditto.
14953 (*movdf_internal): Ditto.
14954 (*movsf_internal): Ditto.
14956 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
14958 * config/i386/i386.md (*movti_internal): Set prefix attribute to
14959 maybe_vex for sselog1 and ssemov types.
14960 (*movdi_internal): Reorder operand constraints.
14961 (*movsi_internal): Ditto. Set prefix attribute to
14962 maybe_vex for sselog1 and ssemov types.
14963 (*movtf_internal): Set prefix attribute to maybe_vex
14964 for sselog1 and ssemov types.
14965 (*movdf_internal): Ditto. Set prefix_data16 attribute for
14966 DImode ssemov types. Reorder operand constraints.
14967 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
14968 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
14969 attribute for SImode ssemov types. Reorder operand constraints.
14971 2013-03-20 Martin Jambor <mjambor@suse.cz>
14973 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
14974 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
14976 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
14978 * config/rs6000/predicates.md (indexed_address, update_address_mem
14979 update_indexed_address_mem): New predicates.
14980 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
14981 attribute for load/store instructions.
14982 * config/rs6000/dfp.md (movsd_store): Likewise.
14983 (movsd_load): Likewise.
14984 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
14985 (unnamed HI->DI extend define_insn): Likewise.
14986 (unnamed SI->DI extend define_insn): Likewise.
14987 (unnamed QI->SI extend define_insn): Likewise.
14988 (unnamed QI->HI extend define_insn): Likewise.
14989 (unnamed HI->SI extend define_insn): Likewise.
14990 (unnamed HI->SI extend define_insn): Likewise.
14991 (extendsfdf2_fpr): Likewise.
14992 (movsi_internal1): Likewise.
14993 (movsi_internal1_single): Likewise.
14994 (movhi_internal): Likewise.
14995 (movqi_internal): Likewise.
14996 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
14997 attribute for load/store instructions.
14998 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
15000 (mov<mode>_softfloat): Likewise.
15001 (mov<mode>_hardfloat32): Likewise.
15002 (mov<mode>_hardfloat64): Likewise.
15003 (mov<mode>_softfloat64): Likewise.
15004 (movdi_internal32): Likewise.
15005 (movdi_internal64): Likewise.
15006 (probe_stack_<mode>): Likewise.
15008 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
15010 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
15011 floating point, and decimal floating point to reload iterator.
15013 * config/rs6000/constraints.md (wl constraint): New constraints to
15014 return FLOAT_REGS if certain options are used to reduce the number
15015 of separate patterns that exist in the file.
15016 (wx constraint): Likewise.
15017 (wz constraint): Likewise.
15019 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15020 -mdebug=reg, print wg, wl, wx, and wz constraints.
15021 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15022 Initialize the reload functions for 64-bit binary/decimal floating
15024 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15025 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15026 create the buffer on the stack to overcome not having a 32-bit
15028 (rs6000_emit_move): Likewise.
15029 (rs6000_secondary_memory_needed_rtx): Likewise.
15030 (rs6000_alloc_sdmode_stack_slot): Likewise.
15031 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15032 via xxlxor, just like DFmode 0.0.
15034 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
15035 define as 1 if we are running on a power7 or newer.
15036 (enum r6000_reg_class_enum): Add new constraints.
15038 * config/rs6000/dfp.md (movsd): Delete, combine with binary
15039 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15040 with other moves by using conditional constraits (wg). Use LFIWZX
15041 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
15042 (movsd splitter): Likewise.
15043 (movsd_hardfloat): Likewise.
15044 (movsd_softfloat): Likewise.
15046 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
15047 binary and decimal floating point moves.
15048 (fmove_ok): New attributes to combine binary and decimal floating
15049 point moves, and to combine power6x (mfpgpr) moves along normal
15051 (real_value_to_target): Likewise.
15052 (f32_lr): Likewise.
15053 (f32_lm): Likewise.
15054 (f32_li): Likewise.
15055 (f32_sr): Likewise.
15056 (f32_sm): Likewise.
15057 (f32_si): Likewise.
15058 (movsf): Combine binary and decimal floating point moves. Combine
15059 power6x (mfpgpr) moves with other moves by using conditional
15060 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
15061 (mov<mode> for SFmode/SDmode); Likewise.
15062 (SFmode/SDmode splitters): Likewise.
15063 (movsf_hardfloat): Likewise.
15064 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
15065 (movsf_softfloat): Likewise.
15066 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
15068 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
15069 wx and wz constraints.
15071 * config/rs6000/constraints.md (wg constraint): New constraint to
15072 return FLOAT_REGS if -mmfpgpr (power6x) was used.
15074 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
15077 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
15078 -mdebug=reg, print wg, wl, wx, and wz constraints.
15079 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
15080 Initialize the reload functions for 64-bit binary/decimal floating
15082 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
15083 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
15084 create the buffer on the stack to overcome not having a 32-bit
15086 (rs6000_emit_move): Likewise.
15087 (rs6000_secondary_memory_needed_rtx): Likewise.
15088 (rs6000_alloc_sdmode_stack_slot): Likewise.
15089 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
15090 via xxlxor, just like DFmode 0.0.
15092 * config/rs6000/dfp.md (movdd): Delete, combine with binary
15093 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
15094 with other moves by using conditional constraits (wg). Use LFIWZX
15095 and STFIWX for loading SDmode on power7.
15096 (movdd splitters): Likewise.
15097 (movdd_hardfloat32): Likewise.
15098 (movdd_softfloat32): Likewise.
15099 (movdd_hardfloat64_mfpgpr): Likewise.
15100 (movdd_hardfloat64): Likewise.
15101 (movdd_softfloat64): Likewise.
15103 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
15104 64-bit binary and decimal floating point moves.
15105 (FMOVE64X): Likewise.
15106 (movdf): Combine 64-bit binary and decimal floating point moves.
15107 Combine power6x (mfpgpr) moves with other moves by using
15108 conditional constraits (wg).
15109 (mov<mode> for DFmode/DDmode): Likewise.
15110 (DFmode/DDmode splitters): Likewise.
15111 (movdf_hardfloat32): Likewise.
15112 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
15113 (movdf_softfloat32): Likewise.
15114 (movdf_hardfloat64_mfpgpr): Likewise.
15115 (movdf_hardfloat64): Likewise.
15116 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
15117 (movdf_softfloat64): Likewise.
15118 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
15119 (reload_<mode>_load): Move to later in the file so they aren't in
15120 the middle of the floating point move insns.
15121 (reload_<mode>_store): Likewise.
15123 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
15126 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
15127 constraint if -mdebug=reg.
15128 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
15129 Enable using dd reload support if needed.
15131 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
15132 binary and decimal floating point moves in rs6000.md.
15133 (movtd_internal): Likewise.
15135 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
15136 decimal floating point moves.
15138 (movtf_internal): Likewise.
15139 (mov<mode>_internal, TDmode/TFmode): Likewise.
15140 (movtf_softfloat): Likewise.
15141 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
15143 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
15144 movdi_internal64, using wg constraint for move direct operations.
15145 (movdi_internal64): Likewise.
15147 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
15148 MODES_TIEABLE_P for selected modes. Print the numerical value of
15149 the various virtual registers. Use GPR/FPR first/last values,
15150 instead of hard coding the register numbers. Print which modes
15151 have reload functions registered.
15152 (rs6000_option_override_internal): If -mdebug=reg, trace the options
15153 settings before/after setting cpu, target and subtarget settings.
15154 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
15155 and for secondary reload failures in rs6000_secondary_reload_inner.
15156 (rs6000_secondary_reload_fail): Likewise.
15157 (rs6000_secondary_reload_inner): Likewise.
15159 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
15160 macros for first/last GPR and FPR registers.
15161 (LAST_GPR_REGNO): Likewise.
15162 (FIRST_FPR_REGNO): Likewise.
15163 (LAST_FPR_REGNO): Likewise.
15165 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
15166 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
15167 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
15168 (vcond<mode><mode>): Likewise.
15169 (vcondu<mode><mode>): Likewise.
15170 (vector_gtu<mode>): Likewise.
15171 (vector_gte<mode>): Likewise.
15172 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
15173 to prevent the compiler from converting DImode operations to TImode.
15174 (ior<mode>3): Likewise.
15175 (and<mode>3): Likewise.
15176 (one_cmpl<mode>2): Likewise.
15177 (nor<mode>3): Likewise.
15178 (andc<mode>3): Likewise.
15180 * config/rs6000/constraints.md (wt constraint): New constraint
15181 that returns VSX_REGS if TImode is allowed in VSX registers.
15183 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
15184 constant under VSX.
15186 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
15187 similar to TImode, but it is restricted to being in the GPRs.
15189 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
15190 TImode to occupy a single VSX register.
15192 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
15193 -mvsx-timode for power7/power8.
15194 (power7 cpu): Likewise.
15195 (power8 cpu): Likewise.
15197 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
15198 sure that TFmode/TDmode take up two registers if they are ever
15199 allowed in the upper VSX registers.
15200 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
15202 (rs6000_init_hard_regno_mode_ok): Likewise.
15203 (rs6000_debug_reg_global): Add debugging for PTImode and wt
15204 constraint. Print if LRA is turned on.
15205 (rs6000_option_override_internal): Give an error if -mvsx-timode
15206 and VSX is not enabled.
15207 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
15208 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
15209 to reg+offset addressing. Use PTImode when checking offset
15210 addresses for validity.
15211 (reg_offset_addressing_ok_p): Likewise.
15212 (rs6000_legitimate_offset_address_p): Likewise.
15213 (rs6000_legitimize_address): Likewise.
15214 (rs6000_legitimize_reload_address): Likewise.
15215 (rs6000_legitimate_address_p): Likewise.
15216 (rs6000_eliminate_indexed_memrefs): Likewise.
15217 (rs6000_emit_move): Likewise.
15218 (rs6000_secondary_reload): Likewise.
15219 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
15220 reloads to fpr registers to continue to use reg+offset addressing,
15221 but 64-bit reloads to altivec registers need reg+reg addressing.
15222 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
15223 it. Treat LO_SUM like a PLUS operation.
15224 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
15225 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
15226 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
15227 registers to share a register with a smaller sized type, since VSX
15228 puts scalars in the upper 64-bits.
15229 (print_operand): Add support for PTImode.
15230 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
15231 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
15232 registers, but don't have arithmetic support.
15233 (rs6000_memory_move_cost): Add test for VSX.
15234 (rs6000_opt_masks): Add -mvsx-timode.
15236 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
15239 (VSr): Use wt constraint for TImode.
15240 (VSv): Drop TImode support.
15241 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
15242 (vsx_movti_64bit): Likewise.
15243 (vsx_movti_32bit): Likewise.
15244 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
15245 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
15246 one '?' on the appropriate output constraint. Do not allow TImode
15247 logical operations on 32-bit systems.
15248 (vsx_ior<mode>3): Likewise.
15249 (vsx_xor<mode>3): Likewise.
15250 (vsx_one_cmpl<mode>2): Likewise.
15251 (vsx_nor<mode>3): Likewise.
15252 (vsx_andc<mode>3): Likewise.
15253 (vsx_concat_<mode>): Likewise.
15254 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
15256 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
15257 OPTION_MASK_VSX_TIMODE.
15258 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
15259 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
15261 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
15262 (TI2 iterator): New iterator for TImode, PTImode.
15263 (wd mode attribute): Add values for vector types.
15264 (movti_string): Replace TI move operations with operations for TImode
15265 and PTImode. Add support for TImode being allowed in VSX registers.
15266 (mov<mode>_string, TImode/PTImode): Likewise.
15267 (movti_ppc64): Likewise.
15268 (mov<mode>_ppc64, TImode/PTImode): Likewise.
15269 (TI mode splitters): Likewise.
15271 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
15274 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
15276 PR tree-optimization/56355
15277 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
15278 Also handle integers with undefined overflow.
15280 2013-03-20 Catherine Moore <clm@codesourcery.com>
15281 Maciej W. Rozycki <macro@codesourcery.com>
15282 Tom de Vries <tom@codesourcery.com>
15283 Nathan Sidwell <nathan@codesourcery.com>
15284 Iain Sandoe <iain@codesourcery.com>
15285 Nathan Froyd <froydnj@codesourcery.com>
15286 Chao-ying Fu <fu@mips.com>
15288 * doc/extend.texi: (micromips, nomicromips, nocompression):
15289 Document new function attributes.
15290 * doc/invoke.texi (minterlink-compressed, mmicromips,
15291 m14k, m14ke, m14kec): Document new options.
15292 (minterlink-mips16): Update documentation.
15293 * doc/md.texi (ZC, ZD): Document new constraints.
15294 * configure.ac (gcc_cv_as_micromips): Check if linker
15295 supports the .set micromips directive.
15296 * configure: Regenerate.
15297 * config.in: Regenerate.
15298 * config/mips/mips-tables.opt: Regenerate.
15299 * config/mips/micromips.md: New file.
15300 * constraints.md (ZC, ZD): New constraints.
15301 * config/mips/predicates.md (movep_src_register): New predicate.
15302 (movep_src_operand): New predicate.
15303 (non_volatile_mem_operand): New predicate.
15304 * config/mips/mips.md (multimem): New type.
15305 (length): Differentiate between 17-bit and 18-bit branch offsets.
15306 (MOVEP1, MOVEP2): New mode iterator.
15307 (mov_<load>l): Use ZC constraint.
15308 (mov_<load>r): Likewise.
15309 (mov_<store>l): Likewise.
15310 (mov_<store>r): Likewise.
15311 (*branch_equality<mode>_inverted): Add microMIPS support.
15312 (*branch_equality<mode>): Likewise.
15313 (*jump_absolute): Likewise.
15314 (indirect_jump_<mode>): Likewise.
15315 (tablejump_<mode>): Likewise.
15316 (<optab>_internal): Likewise.
15317 (sibcall_internal): Likewise.
15318 (sibcall_value_internal): Likewise.
15319 (prefetch): Use constraint ZD.
15320 * config/mips/mips.opt (minterlink-compressed): New option.
15321 (minterlink-mips16): Now an alias for minterlink-compressed.
15322 (mmicromips): New option.
15323 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
15324 (compare_and_swap_12): Likewise.
15325 (sync_add<mode>): Likewise.
15326 (sync_<optab>_12): Likewise.
15327 (sync_old_<optab>_12): Likewise.
15328 (sync_new_<optab>_12): Likewise.
15329 (sync_nand_12): Likewise.
15330 (sync_old_nand_12): Likewise.
15331 (sync_new_nand_12): Likewise.
15332 (sync_sub<mode>): Likewise.
15333 (sync_old_add<mode>): Likewise.
15334 (sync_old_sub<mode>): Likewise.
15335 (sync_new_add<mode>): Likewise.
15336 (sync_new_sub<mode>): Likewise.
15337 (sync_<optab><mode>): Likewise.
15338 (sync_old_<optab><mode>): Likewise.
15339 (sync_new_<optab><mode>): Likewise.
15340 (sync_nand<mode>): Likewise.
15341 (sync_old_nand<mode>): Likewise.
15342 (sync_new_nand<mode>): Likewise.
15343 (sync_lock_test_and_set<mode>): Likewise.
15344 (test_and_set_12): Likewise.
15345 (atomic_compare_and_swap<mode>): Likewise.
15346 (atomic_exchange<mode>_llsc): Likewise.
15347 (atomic_fetch_add<mode>_llsc): Likewise.
15348 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
15349 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
15350 (umips_save_restore_pattern_p): Likewise.
15351 (umips_load_store_pair_p): Likewise.
15352 (umips_output_load_store_pair): Likewise.
15353 (umips_movep_target_p): Likewise.
15354 (umips_12bit_offset_address_p): Likewise.
15355 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
15356 (mips_base_mips16): Rename this...
15357 (mips_base_compression_flags): ...to this. Update all uses.
15358 (mips_attribute_table): Add micromips, nomicromips and nocompression.
15359 (mips_mips16_decl_p): Delete.
15360 (mips_nomips16_decl_p): Delete.
15361 (mips_get_compress_on_flags): New function.
15362 (mips_get_compress_off_flags): New function.
15363 (mips_get_compress_mode): New function.
15364 (mips_get_compress_on_name): New function.
15365 (mips_get_compress_off_name): New function.
15366 (mips_insert_attributes): Support multiple compression types.
15367 (mips_merge_decl_attributes): Likewise.
15368 (umips_12bit_offset_address_p): New function.
15369 (mips_start_function_definition): Emit .set micromips directive.
15370 (mips_call_may_need_jalx_p): New function.
15371 (mips_function_ok_for_sibcall): Add microMIPS support.
15372 (mips_print_operand_punctuation): Support short delay slots and
15374 (umips_swm_mask, umips_swm_encoding): New.
15375 (umips_build_save_restore): New function.
15376 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
15377 (was_mips16_p): Remove.
15378 (old_compression_mode): New.
15379 (mips_set_compression_mode): New function.
15380 (mips_set_current_function): Add microMIPS support.
15381 (mips_option_override): Likewise.
15382 (umips_save_restore_pattern_p): New function.
15383 (umips_output_save_restore): New function.
15384 (umips_load_store_pair_p_1): New function.
15385 (umips_load_store_pair_p): New function.
15386 (umips_output_load_store_pair_1): New function.
15387 (umips_output_load_store_pair): New function.
15388 (umips_movep_target_p) New function.
15389 (mips_prepare_pch_save): Add microMIPS support.
15390 * config/mips/mips.h (TARGET_COMPRESSION): New.
15391 (TARGET_CPU_CPP_BUILTINS): Update macro
15392 to use new compression flags and to support microMIPS.
15393 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
15394 (MIPS_ARCH_FLOAT_SPEC): Likewise.
15395 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
15396 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
15397 (ASM_SPEC): Support mmicromips and mno-micromips.
15398 (M16STORE_REG_P): New macro.
15399 (MIPS_CALL): Support TARGET_MICROMIPS.
15400 (MICROMIPS_J): New macro.
15401 (mips_base_mips16): Rename this...
15402 (mips_base_compression_flags): ...to this.
15403 (UMIPS_12BIT_OFFSET_P): New macro.
15404 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
15405 (MULTILIB_DIRNAMES): Likewise.
15406 2013-03-20 Richard Biener <rguenther@suse.de>
15408 PR tree-optimization/56661
15409 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
15410 the result does not have to be distinct.
15412 2013-03-20 Richard Biener <rguenther@suse.de>
15414 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
15417 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15418 Steven Bosscher <steven@gcc.gnu.org>
15420 PR rtl-optimization/56605
15421 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
15423 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
15426 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
15427 that require movd instead of movq.
15429 2013-03-20 Richard Biener <rguenther@suse.de>
15431 * tree-ssa-structalias.c (struct variable_info): Add pointer
15432 to the first field of an aggregate with sub-vars. Make
15433 this and the pointer to the next subfield its ID.
15434 (vi_next): New function.
15435 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
15436 storedanything_id, integer_id): Increment by one.
15437 (new_var_info, get_call_vi, lookup_call_clobber_vi,
15438 get_call_clobber_vi): Adjust.
15439 (solution_set_expand): Simplify and speedup.
15440 (solution_set_add): Inline into ...
15441 (set_union_with_increment): ... this. Adjust accordingly.
15442 (do_sd_constraint): Likewise.
15443 (do_ds_constraint): Likewise.
15444 (do_complex_constraint): Simplify.
15445 (build_pred_graph): Adjust.
15446 (solve_graph): Likewise. Simplify and speedup.
15447 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
15448 get_constraint_for_component_ref, get_constraint_for_1,
15449 first_vi_for_offset, first_or_preceding_vi_for_offset,
15450 create_function_info_for, create_variable_info_for_1,
15451 create_variable_info_for, intra_create_variable_infos): Adjust.
15452 (init_base_vars): Push NULL for ID zero.
15453 (compute_points_to_sets): Adjust.
15455 2013-03-20 Richard Biener <rguenther@suse.de>
15457 * cfgloop.c (verify_loop_structure): Streamline and avoid
15458 ICEing on corrupt loop tree.
15459 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
15462 2013-03-20 Richard Biener <rguenther@suse.de>
15464 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
15465 check whether an SSA update is needed.
15467 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
15469 * config/mips/constraints.md (T): Rename to...
15473 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
15474 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
15476 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15478 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
15479 (*subsi3_carryin_uxtw): Likewise.
15481 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15483 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
15484 (*rorsi3_insn_uxtw): Likewise.
15486 2013-03-19 Ian Bolton <ian.bolton@arm.com>
15488 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
15489 (*extrsi5_insn_uxtw): Likewise.
15491 2013-03-19 Richard Biener <rguenther@suse.de>
15493 PR tree-optimization/56273
15494 * passes.c (init_optimization_passes): Move second VRP after DOM.
15496 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
15498 * config/i386/i386.md (*movti_internal): Merge from
15499 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
15500 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
15501 nox64 isa attributes.
15503 2013-03-18 Richard Biener <rguenther@suse.de>
15505 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
15507 (merge_node_constraints): Likewise.
15508 (build_succ_graph): Likewise.
15509 (valid_graph_edge): Inline into single caller.
15510 (unify_nodes): Likewise. Use bitmap_set_bit return value
15512 (scc_visit): Fix formatting and variable use.
15513 (do_sd_constraint): Use gcc_checking_assert.
15514 (do_ds_constraint): Likewise.
15515 (do_complex_constraint): Likewise.
15516 (condense_visit): Likewise. Cleanup.
15517 (dump_pred_graph): New function.
15518 (perform_var_substitution): Dump the pred-graph before
15519 variable substitution.
15520 (find_equivalent_node): Use gcc_checking_assert.
15521 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
15523 2013-03-18 Richard Biener <rguenther@suse.de>
15525 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
15526 Remove cond_expr_stmt_list argument and do not gimplify the
15528 (vect_loop_versioning): Adjust.
15529 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
15530 Cleanup to use less temporaries.
15531 (vect_create_data_ref_ptr): Cleanup.
15533 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15535 PR tree-optimization/56635
15536 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
15537 require types_compatible_p types.
15539 2013-03-18 Nick Clifton <nickc@redhat.com>
15541 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
15542 spurious backslash.
15544 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
15545 Add missing line to comment describing function.
15547 2013-03-18 Richard Biener <rguenther@suse.de>
15549 PR tree-optimization/56210
15550 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
15551 Handle string / character search functions.
15552 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
15554 2013-03-18 Richard Biener <rguenther@suse.de>
15556 PR middle-end/56483
15557 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
15558 and implement properly.
15559 * gimple.h (gimple_cond_single_var_p): Remove.
15561 2013-03-18 Richard Biener <rguenther@suse.de>
15563 * tree-data-ref.h (find_data_references_in_loop): Declare.
15564 * tree-data-ref.c (get_references_in_stmt): Use a stack
15565 vector pre-allocated in the callers.
15566 (find_data_references_in_stmt): Adjust.
15567 (graphite_find_data_references_in_stmt): Likewise.
15568 (create_rdg_vertices): Likewise.
15569 (find_data_references_in_loop): Export.
15570 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
15571 Compute dependences here...
15572 (vect_analyze_data_refs): ...not here. When we encounter
15573 a non-vectorizable data reference in basic-block vectorization
15574 truncate the data reference vector. Do not bother to
15575 fixup data-dependence information for gather loads.
15576 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
15577 of data references, as reported.
15579 2013-03-18 Richard Biener <rguenther@suse.de>
15581 PR tree-optimization/3713
15582 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
15583 has_constants and expr.
15584 (stmt_has_constants): Properly valueize SSA names when deciding
15585 whether the stmt has constants.
15587 2013-03-18 Richard Biener <rguenther@suse.de>
15589 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
15590 whole function when there is nothing to do.
15591 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
15592 * tree-vectorizer.c (vectorize_loops): Update virtual and
15593 loop-closed SSA once.
15594 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
15596 2013-03-18 Richard Biener <rguenther@suse.de>
15598 PR middle-end/56113
15599 * domwalk.c (bb_postorder): New global static.
15600 (cmp_bb_postorder): New function.
15601 (walk_dominator_tree): Replace scheme imposing an order for
15602 visiting dominator sons by one sorting them at the time they
15603 are pushed on the stack.
15605 2013-03-18 Richard Biener <rguenther@suse.de>
15607 PR tree-optimization/39326
15608 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
15609 (struct mem_ref): Replace mem member with ao_ref typed member.
15610 (MEM_ANALYZABLE): Adjust.
15611 (memref_eq): Likewise.
15612 (mem_ref_alloc): Likewise.
15613 (gather_mem_refs_stmt): Likewise.
15614 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
15615 (execute_sm_if_changed_flag_set): Adjust.
15616 (execute_sm): Likewise.
15617 (ref_always_accessed_p): Likewise.
15618 (refs_independent_p): Likewise.
15619 (can_sm_ref_p): Likewise.
15621 2013-03-18 Jakub Jelinek <jakub@redhat.com>
15624 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
15625 return 1 even for !unsignedp.
15627 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15629 * config/i386/i386.md (isa): Add x64 and nox64.
15630 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
15631 (*pushtf): Enable *roF alternative for x64 isa only.
15632 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
15633 mode attribute of integer alternatives to DImode for TARGET_64BIT.
15634 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
15635 (*movtf_internal): Merge from *movtf_internal_rex64 and
15636 *movtf_internal_sse. Use x64 and nox64 isa attributes.
15637 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
15638 nox64 isa attributes.
15639 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
15640 nox64 isa attributes.
15641 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
15643 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
15645 * config/alpha/alpha.c (TARGET_LRA_P): New define.
15647 2013-03-17 Jakub Jelinek <jakub@redhat.com>
15650 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
15651 class names. Remove trailing comma after "ALL_REGS".
15653 2013-03-16 Jan Hubicka <jh@suse.cz>
15655 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
15656 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
15657 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
15658 of cgraph_get_create_node.
15659 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
15661 2013-03-16 Jason Merrill <jason@redhat.com>
15664 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
15665 with DW_AT_default_value.
15667 2013-03-16 Jakub Jelinek <jakub@redhat.com>
15669 * BASE-VER: Set to 4.9.0.
15671 2013-03-14 Andi Kleen <ak@linux.intel.com>
15674 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
15675 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
15676 Document _x* TSX intrinsics.
15678 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
15679 David Holsgrove <david.holsgrove@xilinx.com>
15681 * configure.ac: Add MicroBlaze TLS support detection.
15682 * configure: Regenerate.
15683 * config/microblaze/microblaze-protos.h
15684 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
15685 symbol_mentioned_p, label_mentioned_p): Add prototypes.
15686 * config/microblaze/microblaze.c (microblaze_address_type): Add
15687 ADDRESS_TLS and tls_reloc address types.
15688 (microblaze_address_info): Add tls_reloc.
15689 (TARGET_HAVE_TLS): Define.
15690 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
15691 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
15692 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
15693 load_tls_operand, microblaze_call_tls_get_addr,
15694 microblaze_legitimize_tls_address): New functions.
15695 (microblaze_classify_unspec): Handle UNSPEC_TLS.
15696 (get_base_reg): Use microblaze_tls_symbol_p.
15697 (microblaze_classify_address): Handle TLS.
15698 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
15699 label_mentioned_p and microblaze_tls_referenced_p.
15700 (microblaze_legitimize_address): Handle TLS.
15701 (microblaze_address_insns): Handle ADDRESS_TLS.
15702 (pic_address_needs_scratch): Handle TLS.
15703 (print_operand_address): Handle TLS.
15704 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
15705 (microblaze_expand_move): Handle TLS.
15706 (microblaze_legitimate_constant_p): Check
15707 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
15708 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
15709 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
15710 (PIC_OFFSET_TABLE_REGNUM): Set.
15711 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
15712 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
15713 (addsi3, movsi_internal2, movdf_internal): Update constraints
15714 * config/microblaze/predicates.md (arith_plus_operand): Define
15715 (move_operand): Redefine as move_src_operand,
15716 check microblaze_tls_referenced_p.
15718 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15720 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
15721 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
15723 2013-03-14 Ian Bolton <ian.bolton@arm.com>
15725 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
15728 2013-03-14 Jakub Jelinek <jakub@redhat.com>
15730 PR tree-optimization/53265
15731 * common.opt (Waggressive-loop-optimizations): New option.
15732 * tree-ssa-loop-niter.c: Include tree-pass.h.
15733 (do_warn_aggressive_loop_optimizations): New function.
15734 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
15735 if number_of_latch_executions returned constant.
15736 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
15737 early. If number_of_latch_executions returned constant, set
15738 nb_iterations_upper_bound back to it.
15739 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
15741 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
15742 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
15744 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
15745 (MULTILIB_OSDIRNAMES): Set.
15746 * genmultilib: If defaultosdirname doesn't start with :: , set
15747 defaultosdirname2 instead, clear it and emit two . multilib_raw
15748 entries instead of just one.
15750 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
15752 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15753 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
15754 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
15755 (SUBTARGET_OVERRIDE_OPTIONS): New.
15757 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
15760 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
15761 (musermode): Convert to Var(TARGET_USERMODE).
15762 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
15763 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
15764 * config/sh/sh.c (sh_option_override): Use
15765 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
15766 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
15768 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
15770 (udivsi3_i4_single, divsi3_i4_single): Use
15771 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
15773 2013-03-13 Dave Korn <dave.korn.cygwin@....>
15775 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
15778 2013-03-13 Richard Biener <rguenther@suse.de>
15780 PR tree-optimization/56608
15781 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
15782 calls when vectorizing basic-blocks.
15784 2013-03-13 Jakub Jelinek <jakub@redhat.com>
15787 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
15790 2013-03-12 Jakub Jelinek <jakub@redhat.com>
15792 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
15794 2013-03-11 Jan Hubicka <jh@suse.cz>
15797 * lto-streamer-out.c (output_symbol_p): Skip references from
15798 constructors of external variables.
15800 2013-03-11 Jan Hubicka <jh@suse.cz>
15802 PR middle-end/56571
15803 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
15805 * emit-rtl.c (verify_rtx_sharing): Likewise.
15806 (copy_insn_1): Likewise.
15807 * rtl.c (copy_rtx): Likewise.
15809 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
15812 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
15813 output_operand_lossage message.
15815 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
15818 * arm.c (shift_op): Validate RTL pattern on the fly.
15819 (arm_print_operand, case 'S'): Don't use shift_operator to validate
15822 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15825 * config/pa/pa.md (call_value): Check for calls to powf and direct to
15826 new call patterns that clobber %fr12.
15827 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
15828 split and postreload patterns.
15829 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
15830 registers %fr12 and %fr12R as call used.
15832 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
15834 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
15835 (canon_address, record_store, replace_read, check_mem_read_rtx,
15836 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
15837 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
15838 rest_of_handle_dse): Likewise.
15840 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
15842 PR middle-end/56524
15843 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
15845 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
15846 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
15847 (save_optabs_if_changed): Replace with...
15848 (init_tree_optimization_optabs): ...this.
15849 * optabs.c (save_optabs_if_changed): Rename to...
15850 (init_tree_optimization_optabs): ...this. Take the optimization node
15851 as argument. Do nothing if the base optabs are already correct.
15852 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
15853 to recompute optabs.
15854 * function.h (function): Remove optabs field.
15855 * function.c (invoke_set_current_function_hook): Call
15856 init_tree_optimization_optabs. Use the result to initialize
15859 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
15861 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
15862 if GTMA_HAS_NO_INSTRUMENTATION.
15863 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
15864 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
15865 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
15866 * gimple-pretty-print.c (dump_gimple_transaction): Handle
15867 GTMA_HAS_NO_INSTRUMENTATION.
15869 2013-03-08 Jakub Jelinek <jakub@redhat.com>
15871 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
15874 2013-03-08 Marek Polacek <polacek@redhat.com>
15875 Jakub Jelinek <jakub@redhat.com>
15877 PR tree-optimization/56478
15878 * predict.c (is_comparison_with_loop_invariant_p): Change the
15879 type of loop_step to tree.
15880 (predict_loops): Adjust.
15881 (predict_iv_comparison): Perform the computations on double_ints.
15883 2013-03-08 Richard Biener <rguenther@suse.de>
15885 PR tree-optimization/56570
15886 * tree-cfg.c (verify_expr_location_1): Verify locations for
15888 * tree-sra.c (create_access_replacement): Strip locations
15889 from DECL_DEBUG_EXPRs.
15891 2013-03-08 Richard Biener <rguenther@suse.de>
15893 * tree-inline.c (expand_call_inline): Do not associate
15894 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
15895 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
15897 2013-03-08 Richard Biener <rguenther@suse.de>
15899 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
15900 or block changes with -Og. Fix for location / block encoding
15901 changes and PHI arguments with locations.
15903 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
15905 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
15907 (struct output_info): Likewise.
15908 (register_overhead): Remove bad gcc_assert.
15909 (bitmap_find_bit): If there is only a single bitmap element, do not
15910 count a miss as a search.
15911 (print_statistics): Update for counter type changes.
15912 (dump_bitmap_statistics): Likewise. Print headers such that they
15913 are properly lined up with the printed counters.
15915 2013-03-07 Jakub Jelinek <jakub@redhat.com>
15917 PR tree-optimization/56559
15918 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
15919 check that it has only a single use.
15921 2013-03-07 Richard Biener <rguenther@suse.de>
15923 * doc/invoke.texi (fwhole-program): Discourage use in combination
15926 2013-03-06 Jakub Jelinek <jakub@redhat.com>
15928 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
15930 PR tree-optimization/56539
15931 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
15932 instead of GSI_CONTINUE_LINKING as last argument to
15933 force_gimple_operand_gsi. Adjust function comment.
15935 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
15938 PR middle-end/56548
15939 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
15940 promoted mode, convert the result back to the original mode.
15942 2013-03-06 Richard Biener <rguenther@suse.de>
15944 PR middle-end/56294
15945 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
15946 (insert_updated_phi_nodes_compare_uids): New function.
15947 (update_ssa): Sort symbols_to_rename after UID before
15948 traversing it to insert PHI nodes.
15950 2013-03-06 Richard Biener <rguenther@suse.de>
15952 PR middle-end/50494
15953 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
15954 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
15957 2013-02-13 Richard Biener <rguenther@suse.de>
15960 * varasm.c (output_constant_def_1): Get the decl representing
15961 the constant as argument.
15962 (output_constant_def): Wrap output_constant_def_1.
15963 (make_decl_rtl): Use output_constant_def_1 with the decl
15964 representing the constant.
15965 (build_constant_desc): Optionally re-use a decl already
15966 representing the constant.
15967 (tree_output_constant_def): Adjust.
15969 2013-03-06 Joey Ye <joey.ye@arm.com>
15972 * gcc.c (convert_white_space): New function.
15973 (main): Handles white space in function name.
15975 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
15978 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
15979 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
15980 to SH_DIV_CALL_TABLE for TARGET_SH2.
15981 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
15983 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
15984 call-table options.
15986 2013-03-05 Sterling Augustine <saugustine@google.com>
15987 Cary Coutant <ccoutant@google.com>
15990 * dwarf2out.c (resolve_addr): Don't call
15991 remove_loc_list_addr_table_entries a second time for the same
15994 2013-03-05 Jakub Jelinek <jakub@redhat.com>
15997 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
15998 (avoid_complex_debug_insns): New function.
15999 (expand_debug_locations): Call it.
16001 PR rtl-optimization/56484
16002 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
16003 lifetimes of hard registers on small register class machines.
16005 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16007 * config/microblaze/microblaze-protos.h: Rename
16008 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
16009 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
16011 (microblaze_fast_interrupt_function_p): New function.
16012 (microblaze_is_interrupt_handler): Rename to
16013 microblaze_is_interrupt_variant and add fast_interrupt check.
16014 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
16015 (save_restore_insns): Likewise.
16016 (compute_frame_size): Likewise.
16017 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
16018 (microblaze_globalize_label): Likewise.
16019 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
16020 * config/microblaze/microblaze.md: Use wrapper
16021 microblaze_is_interrupt_variant.
16023 2013-03-05 Kai Tietz <ktietz@redhat.com>
16025 * sdbout.c (sdbout_one_type): Switch to current function's section
16026 supporting cold/hot.
16028 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
16030 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
16033 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16035 PR middle-end/56461
16036 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
16037 if VALGRIND_GET_VBITS is defined, temporarily make object
16038 memory all defined, and restore previous valgrind addressability
16039 and definability afterwards. Free this_object at the end.
16041 PR middle-end/56461
16042 * lra.c (lra): Call lra_clear_live_ranges if live_p,
16043 right before calling lra_create_live_ranges, also call it
16044 when clearing live_p. Only call lra_clear_live_ranges
16045 at the end if live_p.
16047 PR middle-end/56461
16048 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
16050 2013-03-05 Richard Biener <rguenther@suse.de>
16052 PR tree-optimization/56521
16053 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
16056 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
16059 * except.h (remove_unreachable_eh_regions): New prototype.
16060 * except.c (remove_eh_handler_splicer): New function, split out
16061 of remove_eh_handler.
16062 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
16063 warning about running it on many EH regions one at a time.
16064 (remove_unreachable_eh_regions_worker): New function, walk the
16065 EH tree in depth-first order and remove non-marked regions.
16066 (remove_unreachable_eh_regions): New function.
16067 * tree-eh.c (mark_reachable_handlers): New function, split out
16068 from remove_unreachable_handlers.
16069 (remove_unreachable_handlers): Use mark_reachable_handlers and
16070 remove_unreachable_eh_regions.
16071 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
16072 and remove_unreachable_eh_regions.
16074 2013-03-05 Richard Biener <rguenther@suse.de>
16076 PR middle-end/56525
16077 * loop-init.c (fix_loop_structure): Remove loops in two stages,
16078 not freeing them until the end.
16080 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16082 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
16084 2013-03-05 Richard Biener <rguenther@suse.de>
16086 PR tree-optimization/56270
16087 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
16088 of loads after scheduling an SLP instance.
16090 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16092 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
16094 (check_gcc_parallelize): Run guality.exp as a separate job from
16095 vect.exp with unsorted.exp and $(dg_target_exps) separately from
16096 struct-layout-1.exp with stackalign.exp.
16098 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
16100 PR middle-end/56461
16101 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
16102 load_index sbitmap even if some bit in it isn't set.
16104 PR middle-end/56461
16105 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
16106 (discover_iteration_bound_by_body_walk): Change queues to
16107 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
16108 spelling in comment. Call safe_push on queues[bound_index] directly.
16109 Release queues[queue_index] in every iteration unconditionally.
16110 Release bounds vector.
16112 PR middle-end/56461
16113 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
16114 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
16115 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
16118 2013-03-05 Richard Biener <rguenther@suse.de>
16121 * tree-inline.c (remap_blocks_to_null): New function.
16122 (expand_call_inline): When expanding a call stmt without
16123 an associated block inline remap all callee blocks to NULL.
16125 2013-03-05 Jakub Jelinek <jakub@redhat.com>
16127 PR rtl-optimization/56494
16128 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
16129 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
16130 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
16132 PR middle-end/56461
16133 * sel-sched-ir.c (free_sched_pools): Release
16134 succs_info_pool.stack[succs_info_pool.max_top] vectors too
16135 if succs_info_pool.max_top isn't -1.
16138 * opts.c (opts_obstack, opts_concat): Moved to...
16139 * opts-common.c (opts_obstack, opts_concat): ... here.
16141 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16143 PR middle-end/56461
16144 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
16146 2013-03-04 Martin Jambor <mjambor@suse.cz>
16148 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
16149 all appropriate places.
16151 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
16153 PR tree-optimization/56424
16154 * ipa-split.c (split_function): Do not set the RSO flag if result is
16155 not by reference and its type is a register type.
16157 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
16159 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
16160 (microblaze_legitimate_pic_operand): Likewise
16161 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
16162 new function microblaze_legitimate_pic_operand
16163 * config/microblaze/microblaze-protos.h
16164 (microblaze_legitimate_pic_operand): Declare.
16166 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16168 * config/microblaze/predicates.md (call_insn_simple_operand):
16169 New predicate for supported rtx code types.
16170 * config/microblaze/microblaze.md (call_internal1): Use
16171 call_insn_simple_operand predicate.
16173 2013-03-04 Jakub Jelinek <jakub@redhat.com>
16175 PR middle-end/56461
16176 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
16177 partitions.ordered_remove.
16179 PR middle-end/56461
16180 * tree-vect-stmts.c (vectorizable_conversion): Don't call
16181 vec_oprnds0.create (1) for modifier == NONE.
16183 PR middle-end/56461
16184 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
16185 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
16186 vec_oprnds1 right before pushing anything to it for
16189 PR middle-end/56461
16190 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
16191 set nbbs to 0 instead of having separate code path.
16192 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
16193 instead of false as last argument if returning NULL.
16195 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
16197 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
16198 the attribute is now called "target" instead of "option".
16199 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
16200 * doc/tm.texi.in (Target Attributes): Likewise document the correct
16201 attribute/pragma name for TARGET_OPTION_VALID_P and
16202 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
16203 * doc/tm.texi: Regenerated.
16205 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
16207 * config/microblaze/microblaze.c:
16208 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
16209 * config/microblaze/microblaze.h: Add -mxl-reorder to
16211 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
16212 instructions emitted if TARGET_REORDER.
16213 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
16214 or 0 for -m/-mno case, but initialises as 2 to detect default use case
16217 2013-03-01 Xinliang David Li <davidxl@google.com>
16219 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
16222 2013-03-01 Jakub Jelinek <jakub@redhat.com>
16224 PR middle-end/56461
16225 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
16226 vector even when returning true. Fix up function comment formatting.
16228 PR middle-end/56461
16229 * ira-build.c (ira_loop_nodes_count): New variable.
16230 (create_loop_tree_nodes): Initialize it.
16231 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
16233 PR middle-end/56461
16234 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
16235 method on dr_chain and result_chain.
16236 * tree-vect-stmts.c (vectorizable_store): Only call
16237 result_chain.create if j == 0.
16239 PR middle-end/56461
16240 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
16241 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
16242 before overwriting it.
16244 2013-03-01 Tobias Burnus <burnus@net-b.de>
16246 * doc/extended.texi (C Extensions): Change order in @menu
16248 (Other MIPS Built-in Functions): Move last MIPS entry before
16249 "picoChip Built-in Functions".
16250 (SH Built-in Functions): Move after RX Built-in Functions.
16251 * doc/gcc.texi (Introduction): Change order in @menu
16253 * doc/md.texi (Constraints): Ditto.
16254 * gty.texi (Type Information): Ditto.
16255 (User-provided marking routines for template types): Make
16257 * doc/invoke.texi (AArch64 Options): Move before
16258 "Adapteva Epiphany Options".
16260 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
16261 Jakub Jelinek <jakub@redhat.com>
16264 * asan.c (gate_asan): Lookup no_sanitize_address instead of
16265 no_address_safety_analysis attribute.
16266 * doc/extend.texi (no_address_safety_attribute): Rename to
16267 no_sanitize_address attribute, mention no_address_safety_analysis
16268 attribute as deprecated alias.
16270 2013-02-28 Jakub Jelinek <jakub@redhat.com>
16272 PR middle-end/56461
16273 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
16274 type to vec<vec<tree> > *.
16275 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
16276 to be vec<tree> instead of vec<tree> *, set vec_defs
16277 to vNULL and call vec_defs.create (number_of_vects), adjust other
16279 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
16280 vectorizable_condition): Adjust vect_get_slp_defs callers.
16282 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16284 * config/aarch64/aarch64.c
16285 (aarch64_float_const_representable): Remove unused variable.
16287 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16289 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
16291 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16293 * config/aarch64/aarch64-builtins.c
16294 (aarch64_init_simd_builtins): Make static.
16296 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
16298 * config/aarch64/aarch64.c
16299 (aarch64_simd_make_constant): Make static.
16301 2013-02-28 Martin Jambor <mjambor@suse.cz>
16303 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
16304 with no initialization to the RHS of debug statements.
16306 2013-02-28 Martin Jambor <mjambor@suse.cz>
16308 PR tree-optimization/56294
16309 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
16311 (get_access_replacement): Do not call create_access_replacement.
16312 Assert a replacement exists.
16313 (get_repl_default_def_ssa_name): Create the replacement declaration
16316 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16318 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
16319 final_end_function.
16321 2013-02-28 Marek Polacek <polacek@redhat.com>
16323 PR rtl-optimization/56466
16324 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
16325 if we're changing a loop.
16326 (peel_loops_completely): Likewise.
16328 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
16331 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
16333 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16336 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
16337 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
16338 INTX_FTYPE_FX, FX_FTYPE_INTX.
16339 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
16341 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16343 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
16344 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
16345 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
16346 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
16347 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
16348 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
16349 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
16350 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
16351 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
16352 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
16353 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
16354 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
16355 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
16356 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
16357 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
16358 (avrxmega6): Increase max flash segments from 5 to 6.
16359 * config/avr/t-multilib: Regenerate.
16360 * config/avr/avr-tables.opt: Regenerate.
16361 * doc/avr-mmcu.texi: Regenerate.
16363 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
16365 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
16366 (avr_device_to_arch): Rename to avr_device_to_ld.
16367 (avr_device_to_as): New prototype.
16368 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
16369 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
16370 * config/avr/driver-avr.c (avr_device_to_as): New.
16371 (avr_device_to_arch): Rename to avr_device_to_ld.
16373 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16375 PR middle-end/56461
16376 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
16377 method on dr_chain and result_chain.
16379 PR middle-end/56461
16380 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
16381 pointer_set_destroy on not_executed_last_iteration.
16383 PR middle-end/56461
16384 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
16386 PR middle-end/56461
16387 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
16388 FOR_EACH_DEFINED_FUNCTION when freeing state.
16390 PR middle-end/56461
16391 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
16393 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
16396 PR middle-end/56461
16397 * ipa-cp.c (decide_whether_version_node): Call vec_free on
16398 known_aggs[i].items and release known_aggs vector.
16400 PR middle-end/56461
16401 * ipa-reference.c (propagate): Free node_info even for alias nodes.
16403 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16405 * config/microblaze/microblaze.c (microblaze_emit_compare):
16406 Use xor for EQ/NE comparisions.
16407 * config/microblaze/microblaze.md (cstoresf4): Add constraints
16408 (cbranchsf4): Adjust operator to comparison_operator.
16410 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16412 PR middle-end/56461
16413 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
16415 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
16416 vec_safe_push, always update *slot.
16417 (redirect_edge_var_map_clear): Use vec_free.
16418 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
16419 (free_var_map_entry): Use vec_free.
16420 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
16421 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
16423 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
16425 PR middle-end/45472
16426 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
16427 when the may_trap_p bit of the exprs being merged differs.
16428 Reorder tests for speculativeness in the logical and operator.
16430 2013-02-27 Jakub Jelinek <jakub@redhat.com>
16432 * incpath.c (add_standard_paths): Use reconcat instead of concat
16433 where appropriate and avoid leaking memory.
16435 * opts.h: Include obstack.h.
16436 (opts_concat): New prototype.
16437 (opts_obstack): New declaration.
16438 * opts.c (opts_concat): New function.
16439 (opts_obstack): New variable.
16440 (init_options_struct): Call gcc_init_obstack on opts_obstack.
16441 (finish_options): Use opts_concat instead of concat
16442 and XOBNEWVEC instead of XNEWVEC.
16443 * opts-common.c (generate_canonical_option, decode_cmdline_option,
16444 generate_option): Likewise.
16445 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
16446 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
16449 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
16450 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
16452 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16454 PR middle-end/56461
16455 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
16457 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16459 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
16460 (arm_block_move_unaligned_straight): Likewise.
16461 (arm_adjust_block_mem): Likewise.
16463 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
16466 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
16467 temp, cond and label.
16468 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
16471 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
16472 * config/c6x/c6x.h (dbx_register_map): Update declaration.
16475 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
16476 of prologue/epilogue functions.
16479 * config/tilegx/tilegx.c (tilegx_expand_prologue):
16480 Remove unused variable cfa_offset.
16481 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
16484 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
16485 type promotion to unsigned.
16488 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
16489 for HOST_WIDE_INT of 32 bit / same size as int.
16490 (arm_block_move_unaligned_straight): Likewise.
16491 (arm_adjust_block_mem): Likewise.
16494 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
16497 2013-02-26 Marek Polacek <polacek@redhat.com>
16499 PR tree-optimization/56426
16500 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
16502 2013-02-26 Richard Biener <rguenther@suse.de>
16505 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
16506 unused variable loops.
16508 2013-02-26 Jakub Jelinek <jakub@redhat.com>
16510 PR tree-optimization/56448
16511 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
16512 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
16513 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
16514 later operands of the references, or even first operand for
16515 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
16517 PR tree-optimization/56443
16518 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
16519 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
16520 to type_for_mode langhook.
16522 2013-02-25 Matt Turner <mattst88@gmail.com>
16524 * doc/invoke.texi: Document r4700.
16526 2013-02-25 Richard Biener <rguenther@suse.de>
16528 PR tree-optimization/56175
16529 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
16531 (simplify_bitwise_binary): ... here. Also guard the conversion
16532 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
16534 2013-02-25 Catherine Moore <clm@codesourcery.com>
16537 2013-02-24 Catherine Moore <clm@codesourcery.com>
16538 Maciej W. Rozycki <macro@codesourcery.com>
16539 Tom de Vries <tom@codesourcery.com>
16540 Nathan Sidwell <nathan@codesourcery.com>
16541 Iain Sandoe <iain@codesourcery.com>
16542 Nathan Froyd <froydnj@codesourcery.com>
16543 Chao-ying Fu <fu@mips.com>
16545 * doc/extend.texi: (micromips, nomicromips, nocompression):
16546 Document new function attributes.
16547 * doc/invoke.texi (minterlink-compressed, mmicromips,
16548 m14k, m14ke, m14kec): Document new options.
16549 (minterlink-mips16): Update documentation.
16550 * doc/md.texi (ZC, ZD): Document new constraints.
16551 * configure.ac (gcc_cv_as_micromips): Check if linker
16552 supports the .set micromips directive.
16553 * configure: Regenerate.
16554 * config.in: Regenerate.
16555 * config/mips/mips-tables.opt: Regenerate.
16556 * config/mips/micromips.md: New file.
16557 * constraints.md (ZC, AD): New constraints.
16558 * config/mips/predicates.md (movep_src_register): New predicate.
16559 (movep_src_operand): New predicate.
16560 (non_volatile_mem_operand): New predicate.
16561 * config/mips/mips.md (multimem): New type.
16562 (length): Differentiate between 17-bit and 18-bit branch offsets.
16563 (MOVEP1, MOVEP2): New mode iterator.
16564 (mov_<load>l): Use ZC constraint.
16565 (mov_<load>r): Likewise.
16566 (mov_<store>l): Likewise.
16567 (mov_<store>r): Likewise.
16568 (*branch_equality<mode>_inverted): Add microMIPS support.
16569 (*branch_equality<mode>): Likewise.
16570 (*jump_absolute): Likewise.
16571 (indirect_jump_<mode>): Likewise.
16572 (tablejump_<mode>): Likewise.
16573 (<optab>_internal): Likewise.
16574 (sibcall_internal): Likewise.
16575 (sibcall_value_internal): Likewise.
16576 (prefetch): Use constraint ZD.
16577 * config/mips/mips.opt (minterlink-compressed): New option.
16578 (minterlink-mips16): Now an alias for minterlink-compressed.
16579 (mmicromips): New option.
16580 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16581 (compare_and_swap_12): Likewise.
16582 (sync_add<mode>): Likewise.
16583 (sync_<optab>_12): Likewise.
16584 (sync_old_<optab>_12): Likewise.
16585 (sync_new_<optab>_12): Likewise.
16586 (sync_nand_12): Likewise.
16587 (sync_old_nand_12): Likewise.
16588 (sync_new_nand_12): Likewise.
16589 (sync_sub<mode>): Likewise.
16590 (sync_old_add<mode>): Likewise.
16591 (sync_old_sub<mode>): Likewise.
16592 (sync_new_add<mode>): Likewise.
16593 (sync_new_sub<mode>): Likewise.
16594 (sync_<optab><mode>): Likewise.
16595 (sync_old_<optab><mode>): Likewise.
16596 (sync_new_<optab><mode>): Likewise.
16597 (sync_nand<mode>): Likewise.
16598 (sync_old_nand<mode>): Likewise.
16599 (sync_new_nand<mode>): Likewise.
16600 (sync_lock_test_and_set<mode>): Likewise.
16601 (test_and_set_12): Likewise.
16602 (atomic_compare_and_swap<mode>): Likewise.
16603 (atomic_exchange<mode>_llsc): Likewise.
16604 (atomic_fetch_add<mode>_llsc): Likewise.
16605 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16606 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16607 (umips_save_restore_pattern_p): Likewise.
16608 (umips_load_store_pair_p): Likewise.
16609 (umips_output_load_store_pair): Likewise.
16610 (umips_movep_target_p): Likewise.
16611 (umips_12bit_offset_address_p): Likewise.
16612 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16613 (mips_base_mips16): Rename this...
16614 (mips_base_compression_flags): ...to this. Update all uses.
16615 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16616 (mips_mips16_decl_p): Delete.
16617 (mips_nomips16_decl_p): Delete.
16618 (mips_get_compress_on_flags): New function.
16619 (mips_get_compress_off_flags): New function.
16620 (mips_get_compress_mode): New function.
16621 (mips_get_compress_on_name): New function.
16622 (mips_get_compress_off_name): New function.
16623 (mips_insert_attributes): Support multiple compression types.
16624 (mips_merge_decl_attributes): Likewise.
16625 (umips_12bit_offset_address_p): New function.
16626 (mips_start_function_definition): Emit .set micromips directive.
16627 (mips_call_may_need_jalx_p): New function.
16628 (mips_function_ok_for_sibcall): Add microMIPS support.
16629 (mips_print_operand_punctuation): Support short delay slots and
16631 (umips_swm_mask, umips_swm_encoding): New.
16632 (umips_build_save_restore): New function.
16633 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16634 (was_mips16_p): Remove.
16635 (old_compression_mode): New.
16636 (mips_set_compression_mode): New function.
16637 (mips_set_current_function): Add microMIPS support.
16638 (mips_option_override): Likewise.
16639 (umips_save_restore_pattern_p): New function.
16640 (umips_output_save_restore): New function.
16641 (umips_load_store_pair_p_1): New function.
16642 (umips_load_store_pair_p): New function.
16643 (umips_output_load_store_pair_1): New function.
16644 (umips_output_load_store_pair): New function.
16645 (umips_movep_target_p) New function.
16646 (mips_prepare_pch_save): Add microMIPS support.
16647 * config/mips/mips.h (TARGET_COMPRESSION): New.
16648 (TARGET_CPU_CPP_BUILTINS): Update macro
16649 to use new compression flags and to support microMIPS.
16650 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16651 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16652 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16653 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16654 (ASM_SPEC): Support mmicromips and mno-micromips.
16655 (M16STORE_REG_P): New macro.
16656 (MIPS_CALL): Support TARGET_MICROMIPS.
16657 (MICROMIPS_J): New macro.
16658 (mips_base_mips16): Rename this...
16659 (mips_base_compression_flags): ...to this.
16660 (UMIPS_12BIT_OFFSET_P): New macro.
16661 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16662 (MULTILIB_DIRNAMES): Likewise.
16664 2013-02-25 Tom de Vries <tom@codesourcery.com>
16666 PR rtl-optimization/56131
16667 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
16668 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
16669 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
16671 2013-02-25 Tobias Burnus <burnus@net-b.de>
16673 * doc/invoke.texi (-fsanitize=): Move from optimization
16674 to debugging options.
16676 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16678 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
16680 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
16681 Alexander Monakov <amonakov@ispras.ru>
16683 PR middle-end/56077
16684 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
16685 flush pending lists also on non-jumps. Adjust comment.
16687 2013-02-24 Catherine Moore <clm@codesourcery.com>
16688 Maciej W. Rozycki <macro@codesourcery.com>
16689 Tom de Vries <tom@codesourcery.com>
16690 Nathan Sidwell <nathan@codesourcery.com>
16691 Iain Sandoe <iain@codesourcery.com>
16692 Nathan Froyd <froydnj@codesourcery.com>
16693 Chao-ying Fu <fu@mips.com>
16695 * doc/extend.texi: (micromips, nomicromips, nocompression):
16696 Document new function attributes.
16697 * doc/invoke.texi (minterlink-compressed, mmicromips,
16698 m14k, m14ke, m14kec): Document new options.
16699 (minterlink-mips16): Update documentation.
16700 * doc/md.texi (ZC, ZD): Document new constraints.
16701 * configure.ac (gcc_cv_as_micromips): Check if linker
16702 supports the .set micromips directive.
16703 * configure: Regenerate.
16704 * config.in: Regenerate.
16705 * config/mips/mips-tables.opt: Regenerate.
16706 * config/mips/micromips.md: New file.
16707 * constraints.md (ZC, AD): New constraints.
16708 * config/mips/predicates.md (movep_src_register): New predicate.
16709 (movep_src_operand): New predicate.
16710 (non_volatile_mem_operand): New predicate.
16711 * config/mips/mips.md (multimem): New type.
16712 (length): Differentiate between 17-bit and 18-bit branch offsets.
16713 (MOVEP1, MOVEP2): New mode iterator.
16714 (mov_<load>l): Use ZC constraint.
16715 (mov_<load>r): Likewise.
16716 (mov_<store>l): Likewise.
16717 (mov_<store>r): Likewise.
16718 (*branch_equality<mode>_inverted): Add microMIPS support.
16719 (*branch_equality<mode>): Likewise.
16720 (*jump_absolute): Likewise.
16721 (indirect_jump_<mode>): Likewise.
16722 (tablejump_<mode>): Likewise.
16723 (<optab>_internal): Likewise.
16724 (sibcall_internal): Likewise.
16725 (sibcall_value_internal): Likewise.
16726 (prefetch): Use constraint ZD.
16727 * config/mips/mips.opt (minterlink-compressed): New option.
16728 (minterlink-mips16): Now an alias for minterlink-compressed.
16729 (mmicromips): New option.
16730 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
16731 (compare_and_swap_12): Likewise.
16732 (sync_add<mode>): Likewise.
16733 (sync_<optab>_12): Likewise.
16734 (sync_old_<optab>_12): Likewise.
16735 (sync_new_<optab>_12): Likewise.
16736 (sync_nand_12): Likewise.
16737 (sync_old_nand_12): Likewise.
16738 (sync_new_nand_12): Likewise.
16739 (sync_sub<mode>): Likewise.
16740 (sync_old_add<mode>): Likewise.
16741 (sync_old_sub<mode>): Likewise.
16742 (sync_new_add<mode>): Likewise.
16743 (sync_new_sub<mode>): Likewise.
16744 (sync_<optab><mode>): Likewise.
16745 (sync_old_<optab><mode>): Likewise.
16746 (sync_new_<optab><mode>): Likewise.
16747 (sync_nand<mode>): Likewise.
16748 (sync_old_nand<mode>): Likewise.
16749 (sync_new_nand<mode>): Likewise.
16750 (sync_lock_test_and_set<mode>): Likewise.
16751 (test_and_set_12): Likewise.
16752 (atomic_compare_and_swap<mode>): Likewise.
16753 (atomic_exchange<mode>_llsc): Likewise.
16754 (atomic_fetch_add<mode>_llsc): Likewise.
16755 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
16756 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
16757 (umips_save_restore_pattern_p): Likewise.
16758 (umips_load_store_pair_p): Likewise.
16759 (umips_output_load_store_pair): Likewise.
16760 (umips_movep_target_p): Likewise.
16761 (umips_12bit_offset_address_p): Likewise.
16762 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
16763 (mips_base_mips16): Rename this...
16764 (mips_base_compression_flags): ...to this. Update all uses.
16765 (mips_attribute_table): Add micromips, nomicromips and nocompression.
16766 (mips_mips16_decl_p): Delete.
16767 (mips_nomips16_decl_p): Delete.
16768 (mips_get_compress_on_flags): New function.
16769 (mips_get_compress_off_flags): New function.
16770 (mips_get_compress_mode): New function.
16771 (mips_get_compress_on_name): New function.
16772 (mips_get_compress_off_name): New function.
16773 (mips_insert_attributes): Support multiple compression types.
16774 (mips_merge_decl_attributes): Likewise.
16775 (umips_12bit_offset_address_p): New function.
16776 (mips_start_function_definition): Emit .set micromips directive.
16777 (mips_call_may_need_jalx_p): New function.
16778 (mips_function_ok_for_sibcall): Add microMIPS support.
16779 (mips_print_operand_punctuation): Support short delay slots and
16781 (umips_swm_mask, umips_swm_encoding): New.
16782 (umips_build_save_restore): New function.
16783 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
16784 (was_mips16_p): Remove.
16785 (old_compression_mode): New.
16786 (mips_set_compression_mode): New function.
16787 (mips_set_current_function): Add microMIPS support.
16788 (mips_option_override): Likewise.
16789 (umips_save_restore_pattern_p): New function.
16790 (umips_output_save_restore): New function.
16791 (umips_load_store_pair_p_1): New function.
16792 (umips_load_store_pair_p): New function.
16793 (umips_output_load_store_pair_1): New function.
16794 (umips_output_load_store_pair): New function.
16795 (umips_movep_target_p) New function.
16796 (mips_prepare_pch_save): Add microMIPS support.
16797 * config/mips/mips.h (TARGET_COMPRESSION): New.
16798 (TARGET_CPU_CPP_BUILTINS): Update macro
16799 to use new compression flags and to support microMIPS.
16800 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
16801 (MIPS_ARCH_FLOAT_SPEC): Likewise.
16802 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
16803 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
16804 (ASM_SPEC): Support mmicromips and mno-micromips.
16805 (M16STORE_REG_P): New macro.
16806 (MIPS_CALL): Support TARGET_MICROMIPS.
16807 (MICROMIPS_J): New macro.
16808 (mips_base_mips16): Rename this...
16809 (mips_base_compression_flags): ...to this.
16810 (UMIPS_12BIT_OFFSET_P): New macro.
16811 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
16812 (MULTILIB_DIRNAMES): Likewise.
16814 2013-02-24 Jakub Jelinek <jakub@redhat.com>
16817 * target-globals.c (save_target_globals): For init_reg_sets and
16818 target_reinit remporarily set this_fn_optabs to this_target_optabs.
16820 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
16822 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
16823 * config/aarch64/t-aarch64
16824 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
16826 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
16828 PR inline-asm/56148
16829 * lra-constraints.c (process_alt_operands): Reload operand
16830 conflicting with earlier clobber only if no more other conflicting
16833 2013-02-22 Jakub Jelinek <jakub@redhat.com>
16836 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
16837 if not linking a shared library.
16839 2013-02-22 Seth LaForge <sethml@google.com>
16841 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
16843 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
16845 * config/arm/arm.md (split for extendsidi): Update condition.
16846 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
16847 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
16848 (qhs_zextenddi_cstr): Likewise.
16850 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16852 PR middle-end/56420
16853 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
16854 avoid signed wrapping.
16855 (expand_mult): Handle properly multiplication by
16856 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
16857 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
16858 in the compiler if coeff is HOST_WIDE_INT_MIN.
16859 (expand_divmod): Don't make ext_op1 static, change it's type to
16860 uhwi. Avoid undefined behavior in -INTVAL (op1).
16862 PR rtl-optimization/50339
16863 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
16865 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
16866 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
16867 into splitting_ashiftrt field.
16868 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
16870 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
16873 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
16875 PR middle-end/56108
16876 * trans-mem.c (execute_tm_mark): Do not expand transactions that
16877 are sure to go irrevocable.
16879 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
16881 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
16882 scalars are valid operands.
16884 2013-02-21 Martin Jambor <mjambor@suse.cz>
16886 PR tree-optimization/56310
16887 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
16888 only matching indices and non-negative final offsets.
16889 (intersect_aggregates_with_edge): Pass src_idx to
16890 agg_replacements_to_vector. Pass src_idx insstead of index to
16891 intersect_with_agg_replacements.
16893 2013-02-21 Martin Jambor <mjambor@suse.cz>
16895 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
16896 instead of hard-wired defaults.
16898 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
16900 * doc/invoke.texi (MIPS Options): Update documentation of the
16901 floating-point multiply-accumulate instruction restrictions.
16903 2013-02-21 Kostya Serebryany <kcc@google.com>
16905 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
16906 asan_shadow_offset on x86_64 linux.
16908 2013-02-21 Richard Biener <rguenther@suse.de>
16910 PR tree-optimization/56415
16912 2013-02-11 Richard Biener <rguenther@suse.de>
16914 PR tree-optimization/56273
16915 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
16918 2013-02-21 Jakub Jelinek <jakub@redhat.com>
16921 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
16924 PR inline-asm/56405
16925 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
16926 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
16928 2013-02-20 Jan Hubicka <jh@suse.cz>
16930 PR tree-optimization/56265
16931 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
16932 when target is referenced for first time.
16934 2013-02-20 Richard Biener <rguenther@suse.de>
16936 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
16937 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
16938 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
16939 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
16940 not return anything.
16941 (rename_ssa_copies): Do not remove unused locals.
16942 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
16943 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
16944 * passes.c (execute_function_todo): Do not schedule unused locals
16945 removal if cleanup_tree_cfg did something.
16946 * tree-ssa-live.c (remove_unused_locals): Dump statistics
16947 about the number of removed locals.
16949 2013-02-20 Richard Biener <rguenther@suse.de>
16951 PR tree-optimization/56398
16952 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
16954 2013-02-20 Martin Jambor <mjambor@suse.cz>
16956 PR tree-optimization/55334
16957 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
16958 restricted pointers to arrays.
16960 2013-02-20 Richard Biener <rguenther@suse.de>
16961 Jakub Jelinek <jakub@redhat.com>
16963 PR tree-optimization/56396
16964 * tree-ssa-ccp.c (n_const_val): New static variable.
16965 (get_value): Return NULL for SSA names we don't have a lattice
16967 (ccp_initialize): Initialize n_const_val.
16968 * tree-ssa-copy.c (n_copy_of): New static variable.
16969 (init_copy_prop): Initialize n_copy_of.
16970 (get_value): Return NULL_TREE for SSA names we don't have a
16973 2013-02-20 Martin Jambor <mjambor@suse.cz>
16975 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
16977 2013-02-20 Richard Biener <rguenther@suse.de>
16979 * genpreds.c (write_lookup_constraint): Do not compare first
16980 letter of the constraint again.
16982 2013-02-20 Richard Biener <rguenther@suse.de>
16984 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
16986 (get_use_iv_cost): Terminate hashtable walk when coming across
16989 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
16991 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
16992 reassociation for avx2 targets.
16994 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
16996 * config/microblaze/microblaze.c: microblaze_has_clz = 0
16997 Add version check for v8.10.a to enable microblaze_has_clz
16998 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
16999 version and TARGET_PATTERN_COMPARE check
17000 * config/microblaze/microblaze.md: New clzsi2 instruction
17002 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17004 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
17005 function before branching.
17007 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17009 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
17011 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
17013 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
17015 PR middle-end/55889
17016 * sel-sched.c: Include ira.h.
17017 (implicit_clobber_conflict_p): New function.
17018 (moveup_expr): Use it.
17019 * Makefile.in (sel-sched.o): Depend on ira.h.
17021 2013-02-19 Richard Biener <rguenther@suse.de>
17023 PR tree-optimization/56384
17024 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
17025 (vn_hash_type): Split out from ...
17026 (vn_hash_constant_with_type): ... here.
17027 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
17028 (vn_phi_eq): Compare types from vn_phi_s structure.
17029 (vn_phi_lookup): Populate vn_phi_s type.
17030 (vn_phi_insert): Likewise.
17032 2013-02-19 Jakub Jelinek <jakub@redhat.com>
17034 PR tree-optimization/56350
17035 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
17036 if haven't found reduction or nested cycle operand, rather than
17037 asserting we must find it.
17039 PR tree-optimization/56381
17040 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
17043 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
17044 Jakub Jelinek <jakub@redhat.com>
17047 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
17048 (swap_optab_enable): Same.
17049 (init_all_optabs): Use argument instead of global.
17050 * tree.h (struct tree_optimization_option): New field target_optabs.
17051 * expr.h (init_all_optabs): Add argument to prototype.
17052 (TREE_OPTIMIZATION_OPTABS): New.
17053 (save_optabs_if_changed): Protoize.
17054 * optabs.h: Declare this_fn_optabs.
17055 * optabs.c (save_optabs_if_changed): New.
17056 Declare this_fn_optabs.
17057 (init_optabs): Add argument to init_all_optabs() call.
17058 * function.c (invoke_set_current_function_hook): Handle per
17060 * function.h (struct function): New field optabs.
17061 * config/mips/mips.c (mips_set_mips16_mode): Handle when
17062 optimization_current_node has changed.
17063 * target-globals.h (save_target_globals_default_opts): Protoize.
17064 * target-globals.c (save_target_globals_default_opts): New.
17066 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17069 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
17070 registers %fr12 and %fr12R as call used.
17073 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
17074 and HImode, require all displacements to be an integer multiple of
17076 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
17077 only allow QImode and HImode when reload is in progress and strict is
17078 true. Likewise for symbolic addresses. Use base14_operand to check
17079 displacements in REG+BASE addresses.
17081 2013-02-18 Richard Biener <rguenther@suse.de>
17083 PR tree-optimization/56366
17084 * tree-vect-loop.c (get_initial_def_for_induction): Properly
17085 handle sign-conversion of outer-loop initial induction value.
17087 2013-02-18 Richard Biener <rguenther@suse.de>
17089 PR middle-end/56349
17090 * cfghooks.c (merge_blocks): If we merge a latch into another
17091 block adjust references to it.
17092 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
17093 (verify_loop_structure): Verify that a recorded latch is in fact
17096 2013-02-18 Richard Biener <rguenther@suse.de>
17098 PR tree-optimization/56321
17099 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
17100 order SSA name release and virtual operand unlinking.
17102 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17104 * config/microblaze/microblaze.md (save_stack_block): Define.
17105 (restore_stack_block): Likewise.
17107 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17109 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
17110 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
17111 * config/microblaze/microblaze.c (microblaze_option_override):
17112 Bail out early for PIC modes when target does not support PIC.
17114 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17116 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
17117 Replace with a microblaze version.
17118 (microblaze_trampoline_init): Adapt for microblaze.
17119 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
17122 2013-02-16 Jakub Jelinek <jakub@redhat.com>
17123 Dodji Seketeli <dodji@redhat.com>
17126 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
17127 (instrument_mem_region_access): Do not forget to always put
17128 instrumentation of the of 'base' and 'base + len' in a "if (len !=
17129 0) statement, even for cases where either 'base' or 'base + len'
17130 are not instrumented -- because they have been previously
17131 instrumented. Simplify the logic by putting all the statements
17132 instrument 'base + len' inside a sequence, and then insert that
17133 sequence right before the current insertion point. Then, to
17134 instrument 'base + len', just get an iterator on that statement.
17135 And do not forget to update the pointer to iterator the function
17136 received as argument.
17138 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
17140 PR rtl-optimization/56348
17141 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
17143 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
17145 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
17146 (clean_graph_dump_file): Pass base to start_graph_dump.
17148 2013-02-14 Richard Henderson <rth@redhat.com>
17151 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
17153 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
17155 * collect2-aix.h: Define F_LOADONLY.
17157 2013-02-14 Richard Biener <rguenther@suse.de>
17160 * varasm.c (output_constant_def_1): Get the decl representing
17161 the constant as argument.
17162 (output_constant_def): Wrap output_constant_def_1.
17163 (make_decl_rtl): Use output_constant_def_1 with the decl
17164 representing the constant.
17165 (build_constant_desc): Optionally re-use a decl already
17166 representing the constant.
17167 (tree_output_constant_def): Adjust.
17169 2013-02-14 Dodji Seketeli <dodji@redhat.com>
17172 * asan.c (instrument_builtin_call): Really put the length of the
17173 second source argument into src1_len.
17175 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17177 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
17178 argument. If it is false, don't create edge from then_bb to
17180 (insert_if_then_before_iter): Pass true to it.
17181 (build_check_stmt): Pass false to it.
17182 (transform_statements): Flush hash table only on extended basic
17183 block boundaries, rather than at the beginning of every bb.
17184 Don't flush hash table on nonfreeing_call_p calls.
17185 * tree-flow.h (nonfreeing_call_p): New prototype.
17186 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
17188 2013-02-13 David S. Miller <davem@davemloft.net>
17190 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
17192 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
17195 * ira.c (max_regno_before_ira): Move from ...
17197 (fix_reg_equiv_init): Use max_regno_before_ira instead of
17200 2013-02-13 Jakub Jelinek <jakub@redhat.com>
17202 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
17204 2013-02-13 Richard Biener <rguenther@suse.de>
17207 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
17208 globals in MEM_REFs.
17210 2013-02-13 Richard Biener <rguenther@suse.de>
17212 * loop-init.c (loop_optimizer_init): Clear loop state when
17213 re-initializing preserved loops.
17214 * loop-unswitch.c (unswitch_single_loop): Return whether
17215 we unswitched the loop. Do not verify loop state here.
17216 (unswitch_loops): When we unswitched a loop discover new loops.
17218 2013-02-13 Kostya Serebryany <kcc@google.com>
17220 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
17222 * sanitizer.def: Rename __asan_init to __asan_init_v1.
17224 2013-02-12 Dodji Seketeli <dodji@redhat.com>
17226 Avoid instrumenting duplicated memory access in the same basic block
17227 * Makefile.in (asan.o): Add new dependency on hash-table.h
17228 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
17229 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
17230 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
17231 (free_mem_ref_resources, has_mem_ref_been_instrumented)
17232 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
17233 (get_mem_ref_of_assignment): New functions.
17234 (get_mem_refs_of_builtin_call): Extract from
17235 instrument_builtin_call and tweak a little bit to make it fit with
17237 (instrument_builtin_call): Use the new
17238 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
17239 of is_gimple_builtin_call.
17240 (instrument_derefs, instrument_mem_region_access): Insert the
17241 instrumented memory reference into the hash table.
17242 (maybe_instrument_assignment): Renamed instrument_assignment into
17243 this, and change it to advance the iterator when instrumentation
17244 actually happened and return true in that case. This makes it
17245 homogeneous with maybe_instrument_assignment, and thus give a
17246 chance to callers to be more 'regular'.
17247 (transform_statements): Clear the memory reference hash table
17248 whenever we enter a new BB, when we cross a function call, or when
17249 we are done transforming statements. Use
17250 maybe_instrument_assignment instead of instrumentation. No more
17251 need to special case maybe_instrument_assignment and advance the
17252 iterator after calling it; it's now handled just like
17253 maybe_instrument_call. Update comment.
17255 2013-02-13 Richard Biener <rguenther@suse.de>
17257 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
17258 Fix loop discovery code.
17260 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
17262 PR inline-asm/56148
17263 * lra-constraints.c (process_alt_operands): Match early clobber
17264 operand with itself. Check conflicts with earlyclobber only if
17265 the operand is not reloaded. Prefer to reload conflicting operand
17266 if earlyclobber and matching operands are the same.
17268 2013-02-12 Richard Biener <rguenther@suse.de>
17271 * lto-streamer-out.c (write_symbol): Do not output symbols
17272 for hard register variables.
17274 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
17277 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
17278 (umulsidi3_insn, mulsidi3_insn): New insns.
17280 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
17282 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
17283 (struct tune_params): Add vec_costs field.
17284 * config/arm/arm.c (arm_builtin_vectorization_cost)
17285 (arm_add_stmt_cost): New functions.
17286 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
17287 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
17288 (arm_default_vec_cost): New struct of type cpu_vec_costs.
17289 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
17290 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
17291 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
17292 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
17294 2013-02-12 Richard Biener <rguenther@suse.de>
17297 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
17298 decls again if possible.
17300 2013-02-12 Richard Biener <rguenther@suse.de>
17302 PR middle-end/56288
17303 * tree-ssa.c (verify_ssa_name): Fix check, move
17304 SSA_NAME_IN_FREE_LIST check up.
17306 2013-02-12 Jakub Jelinek <jakub@redhat.com>
17307 Steven Bosscher <steven@gcc.gnu.org>
17309 PR rtl-optimization/56151
17310 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
17311 equal to op0 or op1, and last_insn pattern is CODE operation
17312 with MEM dest and one of the operands matches that MEM.
17314 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
17316 * doc/extend.texi: Document Function Multiversioning and "default"
17317 parameter string to target attribute.
17318 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
17319 target attribute parameter is "default".
17320 (ix86_compare_version_priority): Remove checks for target attribute.
17321 (ix86_mangle_function_version_assembler_name): Change error to sorry.
17322 Remove check for target attribute equal to NULL. Add assert.
17323 (ix86_generate_version_dispatcher_body): Change error to sorry.
17325 2013-02-11 Iain Sandoe <iain@codesourcery.com>
17326 Jack Howarth <howarth@bromo.med.uc.edu>
17327 Patrick Marlier <patrick.marlier@gmail.com>
17330 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
17331 define ENDFILE_SPEC as TM_DESTRUCTOR.
17332 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
17334 2013-02-11 Alexander Potapenko <glider@google.com>
17335 Jack Howarth <howarth@bromo.med.uc.edu>
17336 Jakub Jelinek <jakub@redhat.com>
17339 * config/darwin.c (cdtor_record): Rename ctor_record.
17340 (sort_cdtor_records): Rename sort_ctor_records.
17341 (finalize_dtors): New routine to sort destructors by
17342 priority before use in assemble_integer.
17343 (machopic_asm_out_destructor): Use finalize_dtors if needed.
17345 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
17347 PR rtl-optimization/56275
17348 * simplify-rtx.c (avoid_constant_pool_reference): Check that
17349 offset is non-negative and less than cmode size before
17350 calling simplify_subreg.
17352 2013-02-11 Richard Biener <rguenther@suse.de>
17354 PR tree-optimization/56264
17355 * cfgloop.h (fix_loop_structure): Adjust prototype.
17356 * loop-init.c (fix_loop_structure): Return the number of
17357 newly discovered loops.
17358 * tree-cfgcleanup.c (repair_loop_structures): When new loops
17359 are discovered, do a full loop-closed SSA rewrite.
17361 2013-02-11 Richard Biener <rguenther@suse.de>
17363 PR tree-optimization/56273
17364 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
17366 (check_array_ref): Fix missing newline in dumps.
17367 (search_for_addr_array): Likewise.
17369 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
17371 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
17373 2013-02-09 Jakub Jelinek <jakub@redhat.com>
17376 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
17378 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
17380 PR rtl-optimization/56246
17381 * lra-constraints.c (simplify_operand_subreg): Try to reuse
17383 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
17384 constraints are satisfied.
17386 2013-02-08 Jeff Law <law@redhat.com>
17389 * emit-rtl.c (reg_is_parm_p): New function.
17390 * regs.h (reg_is_parm_p): New prototype.
17391 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
17392 callee-clobbered registers.
17394 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
17397 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
17398 If there is no implicit builtin declaration, just return NULL.
17400 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17402 * config/i386/sse.md (FMAMODEM): New mode iterator.
17403 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
17404 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
17406 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17408 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
17409 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
17410 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
17412 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
17414 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
17415 (microblaze*-*-elf): Likewise.
17416 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
17418 * config/microblaze/microblaze-c.c: Add builtin defines for
17419 _LITTLE_ENDIAN and _BIG_ENDIAN.
17420 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
17421 add to TARGET_DEFAULT flags.
17422 Expand ASM_SPEC and LINK_SPEC.
17423 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
17424 * config/microblaze/microblaze.md: Update extendsidi2 and
17425 movdi_internal instructions to use low-order / high-order reg
17427 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
17428 options and inversemask / mask of LITTLE_ENDIAN.
17429 * config/microblaze/t-microblaze: Expand multilib options to
17430 include mlittle-endian (le) and update exceptions patterns.
17432 2013-02-08 Jakub Jelinek <jakub@redhat.com>
17434 PR rtl-optimization/56195
17435 * lra-constraints.c (get_reload_reg): Don't reuse regs
17436 if they have smaller mode than requested, if they have
17437 wider mode than requested, try to return a SUBREG.
17439 PR tree-optimization/56250
17440 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
17441 if type is unsigned and code isn't MULT_EXPR.
17443 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17445 PR tree-optimization/56064
17446 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
17447 bits according to mode.
17448 * fixed-value.h (fixed_from_double_int)
17449 (const_fixed_from_double_int): Adjust comments.
17451 2013-02-08 Richard Biener <rguenther@suse.de>
17454 * lto-streamer.h (struct data_in): Remove current_file, current_line
17455 and current_col members.
17456 * lto-streamer-out.c (lto_output_location): Stream changed bits
17457 en-block for efficiency.
17458 * lto-streamer-in.c (clear_line_info): Remove.
17459 (lto_input_location): Cache current file, line and column
17460 globally via local statics. Read changed bits en-block.
17461 (input_function): Do not call clear_line_info.
17462 (lto_read_body): Likewise.
17463 (lto_input_toplevel_asms): Likewise.
17465 2013-02-08 Michael Matz <matz@suse.de>
17467 PR tree-optimization/52448
17468 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
17469 (nt_call_phase): New static.
17470 (add_or_mark_expr): Only mark accesses with newer phase than any
17472 (nonfreeing_call_p): New.
17473 (nt_init_block): Update nt_call_phase, mark blocks as visited.
17474 (nt_fini_block): Keep blocks marked as visited.
17475 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
17477 2013-02-08 Richard Biener <rguenther@suse.de>
17479 * ira.c (ira): Free broken dominator information.
17481 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
17483 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
17485 2013-02-08 Marek Polacek <polacek@redhat.com>
17487 * cfgloop.c (verify_loop_structure): Add more checking of headers.
17489 2013-02-08 Richard Biener <rguenther@suse.de>
17491 PR middle-end/56181
17492 * cfgloop.h (flow_loops_find): Adjust.
17493 (bb_loop_header_p): Declare.
17494 * cfgloop.c (bb_loop_header_p): New function split out from ...
17495 (flow_loops_find): ... here. Adjust function signature,
17496 support incremental loop structure update.
17497 (verify_loop_structure): Cleanup. Verify a loop is a loop.
17498 * cfgloopmanip.c (fix_loop_structure): Move ...
17499 * loop-init.c (fix_loop_structure): ... here.
17500 (apply_loop_flags): Split out from ...
17501 (loop_optimizer_init): ... here.
17502 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
17503 in incremental mode, only remove dead loops here.
17505 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
17508 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
17509 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
17510 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
17511 (*round<mode>3.libgcc): New insns for fixed-modes.
17512 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
17513 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
17514 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
17515 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
17516 implementations. Define to __builtin_avr_absFX,
17517 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
17518 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
17519 __builtin_avr_countlsFX, respectively.
17520 * config/avr/avr-c.c (target.h): Include it.
17521 (enum avr_builtin_id): New enum.
17522 (avr_resolve_overloaded_builtin): New static function.
17523 (avr_register_target_pragmas): Use it to set
17524 targetm.resolve_overloaded_builtin.
17525 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
17526 tree nodes used by DEF_BUILTIN.
17527 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
17528 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
17529 <AVR_BUILTIN_xxBITS>: Same.
17531 2013-02-08 Richard Biener <rguenther@suse.de>
17533 * cfgloop.c (verify_loop_structure): Properly handle
17534 a loop exiting to another loop header.
17535 * ira-int.h (ira_loops): Remove.
17536 * ira.c (ira_loops): Remove.
17537 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
17538 (do_reload): Use loop_optimizer_finalize.
17539 * ira-build.c (create_loop_tree_nodes): Use get_loops and
17540 number_of_loops to access the loop tree.
17541 (more_one_region_p): Likewise.
17542 (finish_loop_tree_nodes): Likewise.
17543 (rebuild_regno_allocno_maps): Likewise.
17544 (mark_loops_for_removal): Likewise.
17545 (mark_all_loops_for_removal): Likewise.
17546 (remove_unnecessary_regions): Likewise.
17547 (ira_build): Likewise.
17548 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17550 2013-02-08 Richard Biener <rguenther@suse.de>
17552 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
17553 * ipa-pure-const.c (analyze_function): Avoid calling
17554 mark_irreducible_loops twice.
17555 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
17557 2013-02-07 David S. Miller <davem@davemloft.net>
17559 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
17561 * var-tracking.c (vt_add_function_parameter): Test the presence of
17562 HAVE_window_save properly and do not remap argument registers when
17563 we have a leaf function.
17565 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
17568 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
17570 * config/i386/i386.c (ix86_print_operand): Ditto.
17572 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17574 * lra-constraints.c (process_alt_operands): Fix recently added comment.
17576 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
17578 PR rtl-optimization/56225
17579 * lra-constraints.c (process_alt_operands): Check that reload hard
17580 reg can hold value for strict_low_part.
17582 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17585 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
17586 dwarf2out_end_function.
17587 (in_first_function_p, maybe_at_text_label_p,
17588 first_loclabel_num_not_at_text_label): New variables.
17589 (dwarf2out_var_location): In the first function find out
17590 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
17591 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
17594 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
17596 PR rtl-optimization/56178
17597 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
17598 SUBREG of a register. Tidy up related block of code.
17599 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
17600 note if the source is a register or a SUBREG of a register.
17602 2013-02-07 Jakub Jelinek <jakub@redhat.com>
17605 * config/rs6000/rs6000.md (ptrm): New mode attr.
17606 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
17607 call_value_indirect_aix<pttrsize>,
17608 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
17611 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
17613 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
17614 if -bnortl. Convert to strcmp and strncmp.
17616 2013-02-07 Alan Modra <amodra@gmail.com>
17619 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
17620 addresses won't wrap when offsetting.
17621 (rs6000_secondary_reload): Provide secondary reloads needed for
17622 wrapping LO_SUM addresses.
17624 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
17626 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
17627 MACH, just __MACH__.
17629 2013-02-06 Richard Biener <rguenther@suse.de>
17631 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
17632 instead of calling fix_loop_structure.
17634 2013-02-06 Jakub Jelinek <jakub@redhat.com>
17636 PR middle-end/56217
17637 * omp-low.c (use_pointer_for_field): Return false if
17638 lower_send_shared_vars doesn't generate any copy-out code.
17640 2013-02-06 Tom de Vries <tom@codesourcery.com>
17642 PR rtl-optimization/56131
17643 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
17644 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
17645 of the label is NULL. Add comment.
17647 2013-02-05 Jakub Jelinek <jakub@redhat.com>
17649 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
17652 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
17653 (STATIC_LIBTSAN_LIBS): Likewise.
17654 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
17655 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
17656 is defined, don't add anything else beyond that.
17657 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
17658 (LINK_COMMAND_SPEC): Use them.
17660 PR tree-optimization/56205
17661 * tree-stdarg.c (check_all_va_list_escapes): Return true if
17662 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
17663 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
17665 2013-02-05 Richard Biener <rguenther@suse.de>
17667 PR tree-optimization/53342
17668 PR tree-optimization/53185
17669 * tree-vectorizer.h (vect_check_strided_load): Remove.
17670 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
17671 not disallow peeling for vectorized strided loads.
17672 (vect_check_strided_load): Make static and simplify.
17673 (vect_analyze_data_refs): Adjust.
17674 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
17675 correctly when vectorizing strided loads.
17677 2013-02-05 Richard Biener <rguenther@suse.de>
17679 * doc/install.texi: Refer to ISL, not PPL.
17681 2013-02-05 Jan Hubicka <jh@suse.cz>
17683 PR tree-optimization/55789
17684 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
17686 2013-02-05 Jan Hubicka <jh@suse.cz>
17688 PR tree-optimization/55789
17689 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
17690 the dead call anyway.
17692 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
17695 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
17697 2013-02-04 Alexander Potapenko <glider@google.com>
17698 Jack Howarth <howarth@bromo.med.uc.edu>
17699 Jakub Jelinek <jakub@redhat.com>
17702 * config/darwin.c (sort_ctor_records): Stabilized qsort
17703 on constructor priority by using original position.
17704 (finalize_ctors): New routine to sort constructors by
17705 priority before use in assemble_integer.
17706 (machopic_asm_out_constructor): Use finalize_ctors if needed.
17708 2013-02-04 Jakub Jelinek <jakub@redhat.com>
17711 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
17712 about visibility on artificial decls.
17713 * config/sol2.c (solaris_assemble_visibility): Likewise.
17715 2013-02-04 Kai Tietz <ktietz@redhat.com>
17718 * config/i386/i386.c (function_value_ms_64): Add additional valtype
17719 argument and improve checking of return-argument types for 16-byte
17721 (ix86_function_value_1): Add additional valtype argument on call
17722 of function_value_64.
17723 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
17724 handling infunction_value_64 function.
17726 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
17728 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
17730 2013-02-04 Richard Biener <rguenther@suse.de>
17732 PR tree-optimization/56188
17733 * tree-ssa-structalias.c (label_visit): Consider case with
17734 initially non-empty points-to set.
17735 (perform_var_substitution): Dump node mapping and clean up.
17737 2013-02-04 Richard Guenther <rguenther@suse.de>
17740 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
17741 node prevail as last resort.
17742 (lto_symtab_merge_decls): Remove guard on LTRANS here.
17743 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
17745 2013-02-04 Richard Biener <rguenther@suse.de>
17747 PR tree-optimization/56113
17748 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
17750 (equiv_class_lookup_or_add): ... this.
17751 (label_visit): Adjust and fix error in previous patch.
17752 (perform_var_substitution): Adjust.
17754 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
17756 * config/sh/divtab.c: Fix formatting and comments throughout the file.
17757 * config/sh/sh4-300.md: Likewise.
17758 * config/sh/sh4a.md: Likewise.
17759 * config/sh/constraints.md: Likewise.
17760 * config/sh/sh.md: Likewise.
17761 * config/sh/netbsd-elf.h: Likewise.
17762 * config/sh/predicates.md: Likewise.
17763 * config/sh/sh-protos.h: Likewise.
17764 * config/sh/ushmedia.h: Likewise.
17765 * config/sh/linux.h: Likewise.
17766 * config/sh/sh.c: Likewise.
17767 * config/sh/superh.h: Likewise.
17768 * config/sh/elf.h: Likewise.
17769 * config/sh/sh4.md: Likewise.
17770 * config/sh/sh.h: Likewise.
17772 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17774 * config/pa/constraints.md: Adjust unused letters. Change "T"
17775 constraint to match_test floating_point_store_memory_operand().
17776 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
17777 (base14_operand): New.
17778 (floating_point_store_memory_operand): New.
17779 (integer_store_memory_operand): Revise to use base14_operand and
17780 reg_plus_base_memory_operand.
17781 (move_dest_operand): Allow symbolic_memory_operands.
17782 (symbolic_memory_operand): Check for LO_SOM.
17783 (symbolic_operand): Change default case to break.
17784 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
17785 CONST_DOUBLE values to be reloaded by putting them into memory when
17786 the destination is a floating point register.
17787 (movdf): Remove code to handle CONST_DOUBLE.
17789 (reload_indf_r1): New.
17790 (reload_insf_r1): New.
17791 Consistently use "Q" and "T" constraints with integer and floating
17792 point move instructions, respectively.
17793 (movdi): Remove FAIL.
17794 Change predicate for source operand unamed DImode move from
17795 general_operand to move_src_operand.
17796 (umulsidi3): Change predicate for destination operand to
17798 Likewise for similar unamed patterns.
17799 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
17800 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
17801 (hppa_legitimize_address): Simplify mask calculation.
17802 (pa_emit_move_sequence): Revised handling of secondary reloads from
17803 REG+D addresses for floating point loads and stores. Directly handle
17804 loading CONST0_RTX (mode) to a floating point register.
17805 (pa_secondary_reload): Handle reloading DF and SFmode constant values
17806 to floating point registers. Don't restrict secondary reloads to
17807 floating point registers to integer modes. Revise some comments and
17809 (TARGET_LEGITIMATE_ADDRESS_P): Define.
17810 (pa_legitimate_address_p): New.
17811 (pa_legitimize_reload_address): New.
17812 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
17813 (STRICT_REG_OK_FOR_BASE_P): New.
17814 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
17815 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
17817 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
17818 Andrew Dixie <andrewd@gentrack.com>
17820 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
17823 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
17825 * expmed.c (extract_bit_field_1): Pass the full width of the
17826 structure to get_best_reg_extraction_insn.
17828 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17831 * configure.ac (use_cxa_atexit): Add AIX.
17832 * configure: Regenerate.
17834 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
17836 2013-02-01 Jakub Jelinek <jakub@redhat.com>
17839 * final.c (need_profile_function): New variable.
17840 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
17841 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
17842 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
17843 notes, targetm.asm_out.function_prologue doesn't emit anything,
17844 HAVE_prologue and profiler should be emitted before prologue,
17845 set need_profile_function instead of emitting it.
17846 (final_scan_insn): If need_profile_function, emit
17847 profile_function on the first NOTE_INSN_BASIC_BLOCK or
17848 NOTE_INSN_FUNCTION_BEG note.
17850 2013-02-01 Richard Henderson <rth@redhat.com>
17852 * config/rs6000/rs6000.md (smulditi3): New.
17855 * config/alpha/alpha.md (umulditi3): New.
17857 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
17859 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
17860 (ASM_OUTPUT_ALIGNED_LOCAL): New.
17862 2013-02-01 Richard Biener <rguenther@suse.de>
17864 PR tree-optimization/56113
17865 * tree-ssa-structalias.c (label_visit): Reduce work for
17866 single-predecessor nodes.
17868 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
17870 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
17871 range isn't testing for zero.
17873 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
17875 PR middle-end/56113
17876 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
17878 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
17879 Nick Clifton <nickc@redhat.com>
17881 * config/v850/constraints.md (Q): Define as a memory constraint.
17882 * config/v850/predicates.md (label_ref_operand): New predicate.
17883 (e3v5_shift_operand): New predicate.
17884 (ior_operator): New predicate.
17885 * config/v850/t-v850: Add e3v5 multilib.
17886 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
17887 (v850_gen_movdi): Prototype.
17888 * config/v850/v850.c: Add support for e3v5 architecture.
17889 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
17891 (construct_save_jarl): Add e3v5 long JARL support.
17892 (v850_adjust_insn_length): New function. Adjust length of call
17893 insns when using e3v5 instructions.
17894 (v850_gen_movdi): New function: Generate instructions to move a
17896 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
17897 (CPP_SPEC): Define __v850e3v5__ as appropriate.
17898 (TARGET_USE_FPU): Enable for e3v5.
17899 (CONST_OK_FOR_W): New macro.
17900 (ADJUST_INSN_LENGTH): Define.
17901 * config/v850/v850.md (UNSPEC_LOOP): Define.
17902 (attr cpu): Add v850e3v5.
17903 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
17904 (movdi): New pattern.
17905 (movdi_internal): New pattern.
17906 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
17907 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
17908 (cstoresf4): Likewise.
17909 (cstoredf4): Likewise.
17910 (insv): New pattern.
17911 (rotlso3_a): New pattern.
17912 (rotlsi3_b): New pattern
17913 (rotlsi3_v850e3v5): New pattern.
17914 (doloop_begin): New pattern.
17915 (fix_loop_counter): New pattern.
17916 (doloop_end): New pattern.
17917 (branch_normal): Add e3v5 long branch support.
17918 (branch_invert): Likewise.
17919 (branch_z_normal): Likewise.
17920 (branch_z_invert): Likewise.
17921 (branch_nz_normal): Likewise.
17922 (branch_nz_invert): Likewise.
17923 (call_internal_short): Add e3v5 register-indirect JARL support.
17924 (call_internal_long): Likewise.
17925 (call_value_internal_short): Likewise.
17926 (call_value_internal_long): Likewise.
17927 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
17928 (mloop): New option.
17929 * config.gcc: Add support for configuring v840e3v5 target.
17930 * doc/invoke.texi: Document new v850 specific command line options.
17932 2013-01-31 Paul Koning <ni1d@arrl.net>
17936 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
17937 children if parent is a class.
17938 (prune_unused_types_prune): Don't add DW_AT_declaration.
17940 2013-01-31 Richard Biener <rguenther@suse.de>
17942 PR tree-optimization/56157
17943 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
17944 match up operand with SLP child.
17946 2013-01-31 Jason Merrill <jason@redhat.com>
17949 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
17950 parameters the first time.
17951 (gen_scheduled_generic_parms_dies): Check completeness here.
17953 2013-01-31 Richard Biener <rguenther@suse.de>
17955 PR middle-end/53073
17956 * common.opt (faggressive-loop-optimizations): New flag,
17957 enabled by default.
17958 * doc/invoke.texi (faggressive-loop-optimizations): Document.
17959 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
17960 infer_loop_bounds_from_undefined by it.
17962 2013-01-31 Richard Biener <rguenther@suse.de>
17964 PR tree-optimization/56150
17965 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
17966 visit virtual operands.
17967 (find_uses_to_rename_bb): Likewise.
17969 2013-01-31 Richard Biener <rguenther@suse.de>
17971 PR tree-optimization/56150
17972 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
17973 mixed store non-store stmts.
17975 2013-01-30 Jakub Jelinek <jakub@redhat.com>
17978 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
17979 LIBASAN_EARLY_SPEC is defined.
17980 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
17981 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
17983 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
17986 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
17987 invalid args instead of ICEing on it.
17988 (ix86_valid_target_attribute_tree): Return error_mark_node if
17989 ix86_valid_target_attribute_inner_p failed.
17990 (ix86_valid_target_attribute_p): Return false only if
17991 ix86_valid_target_attribute_tree returned error_mark_node. Allow
17992 target("default") attribute.
17993 (sorted_attr_string): Change argument from const char * to tree,
17994 merge in all target attribute arguments rather than just one.
17995 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
17996 instead of free. Avoid using strcat.
17997 (ix86_mangle_function_version_assembler_name): Mangle
17998 target("default") as if no target attribute is present. Adjust
17999 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
18000 instead of xmalloc and XDELETEVEC instead of free.
18001 (ix86_function_versions): Don't return true if one of the decls
18002 doesn't have target attribute. If they don't and one of the decls
18003 is DECL_FUNCTION_VERSIONED, report an error. Adjust
18004 sorted_attr_string caller. Use XDELETEVEC instead of free.
18005 (ix86_supports_function_versions): Remove.
18006 (make_name): Fix up formatting.
18007 (make_dispatcher_decl): Remove resolver_name and its initialization.
18008 Avoid leaking memory.
18009 (is_function_default_version): Return true if there is
18010 target("default") attribute rather than no target attribute at all.
18011 (make_resolver_func): Avoid leaking memory.
18012 (ix86_generate_version_dispatcher_body): Likewise.
18013 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
18014 * target.def (supports_function_versions): Remove.
18015 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
18016 * doc/tm.texi: Regenerated.
18018 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
18020 PR rtl-optimization/56144
18021 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
18022 for values with side effects.
18024 2013-01-30 Richard Biener <rguenther@suse.de>
18026 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
18027 (sparseset_pop): Likewise.
18028 * cfganal.c (compute_idf): Likewise. Increase work-stack size
18029 to be able to use quick_push in the worker loop.
18031 2013-01-30 Marek Polacek <polacek@redhat.com>
18033 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
18035 2013-01-30 Richard Biener <rguenther@suse.de>
18038 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
18040 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
18042 PR tree-optimization/56064
18043 * fixed-value.c (fixed_from_double_int): New function.
18044 * fixed-value.h (fixed_from_double_int): New prototype.
18045 (const_fixed_from_double_int): New static inline function.
18046 * fold-const.c (native_interpret_fixed): New static function.
18047 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
18048 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
18049 (native_encode_fixed): New static function.
18050 (native_encode_expr) <FIXED_CST>: Use it.
18051 (native_interpret_int): Move double_int worker code to...
18052 * double-int.c (double_int::from_buffer): ...this new static method.
18053 * double-int.h (double_int::from_buffer): Prototype it.
18055 2013-01-30 Richard Biener <rguenther@suse.de>
18057 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
18058 New pointer-map and obstack.
18059 (init_alias_vars): Allocate pointer-map and obstack.
18060 (delete_points_to_sets): Free them.
18061 (find_what_var_points_to): Cache result.
18062 (find_what_p_points_to): Adjust for changed interface of
18063 find_what_var_points_to.
18064 (compute_points_to_sets): Likewise.
18065 (ipa_pta_execute): Likewise.
18067 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18069 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
18070 * configure: Regenerate.
18071 * config.in: Regenerate.
18072 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
18073 #nobits/#progbits if supported.
18075 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
18078 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
18079 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
18080 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
18082 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18084 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
18085 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
18087 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18089 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
18090 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
18092 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18094 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
18096 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
18097 * config/arm/cortex-a7.md: New bypasses using
18098 arm_mac_accumulator_is_result.
18100 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18102 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
18103 (cortex_a7_neon_mla): Likewise.
18104 (cortex_a7_fpfmad): New reservation.
18105 (cortex_a7_fpmacs): Use ffmas and update required units.
18106 (cortex_a7_fpmuld): Update required units and latency.
18107 (cortex_a7_fpmacd): Likewise.
18108 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
18109 (cortex_a7_neon). Likewise.
18110 (bypass) Update participating units.
18112 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
18114 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
18115 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
18117 * config/arm/vfp11.md (vfp_farith): Use ffmas.
18118 (vfp_fmul): Use ffmad.
18119 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
18120 (cortex_r4_fmacd): Use ffmad.
18121 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
18122 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
18123 (cortex_a9_fmacd): Use ffmad.
18124 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
18125 (cortex_a8_vfp_macd): Use ffmad.
18126 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
18127 (cortex_a5_fpmacd): Use ffmad.
18128 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
18129 (cortex_a15_vfp_macd): Use ffmad.
18130 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
18132 2013-01-29 Jason Merrill <jason@redhat.com>
18135 * varasm.c (default_assemble_visibility): Don't warn about
18136 visibility on artificial decls.
18138 2013-01-29 Richard Biener <rguenther@suse.de>
18140 PR tree-optimization/56113
18141 * tree-ssa-structalias.c (equiv_class_lookup): Also return
18143 (label_visit): Free duplicate bitmaps and record the leader instead.
18144 (perform_var_substitution): Adjust.
18146 2013-01-29 Richard Biener <rguenther@suse.de>
18148 PR tree-optimization/55270
18149 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
18150 the CFG, schedule loops for fixup.
18152 2013-01-29 Nick Clifton <nickc@redhat.com>
18154 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
18157 2013-01-28 Leif Ekblad <leif@rdos.net>
18159 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
18160 * config/i386/i386.h (TARGET_RDOS): New macro.
18161 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
18162 * config/i386/i386.c (ix86_option_override_internal): For 64bit
18163 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
18164 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
18165 DEFAULT_LARGE_SECTION_THRESHOLD.
18166 * config/i386/i386.md (R14_REG, R15_REG): New constants.
18167 * config/i386/rdos.h: New file.
18168 * config/i386/rdos64.h: New file.
18170 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
18173 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
18176 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18178 PR rtl-optimization/56117
18179 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
18180 call cselib_lookup_from_insn on the MEM before calling
18181 add_insn_mem_dependence.
18183 2013-01-28 Richard Biener <rguenther@suse.de>
18185 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
18186 to a stmt that didn't have one.
18187 (copy_phis_for_bb): Likewise for PHI arguments.
18188 (copy_debug_stmt): Likewise for debug stmts.
18190 2013-01-28 Richard Biener <rguenther@suse.de>
18192 PR tree-optimization/56034
18193 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
18194 (partition_builtin_p): Adjust.
18195 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
18196 it is the last partition.
18197 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
18198 up the vertex for the definition.
18199 (classify_partition): Classify whether a partition is a
18200 PKIND_REDUCTION, thus has uses outside of the loop.
18201 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
18202 Merge all PKIND_REDUCTION partitions into the last partition.
18203 (tree_loop_distribution): Seed partitions from reductions as well.
18205 2013-01-28 Jakub Jelinek <jakub@redhat.com>
18207 PR tree-optimization/56125
18208 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
18209 pow(x,c) into sqrt(x) * powi(x, n/2) or
18210 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
18211 optimizing for size.
18212 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
18213 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
18216 PR tree-optimization/56094
18217 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
18218 to UNKNOWN_LOCATION while gimplifying expr.
18220 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
18223 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
18224 operand 0 in movabs insn template for -masm=intel asm alternative.
18225 (*movabs<mode>_2): Ditto for operand 1.
18227 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
18230 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
18233 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18235 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
18238 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18240 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
18241 Undef to avoid warning.
18243 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
18245 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
18246 * configure: Regenerate.
18248 2013-01-25 Jakub Jelinek <jakub@redhat.com>
18250 PR tree-optimization/56098
18251 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
18252 for stmts with volatile ops.
18253 (cond_store_replacement): Don't optimize if assign has volatile ops.
18254 (cond_if_else_store_replacement_1): Don't optimize if either
18255 then_assign or else_assign have volatile ops.
18256 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
18259 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18261 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
18263 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18265 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
18266 missing ':' in asm example.
18268 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
18270 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
18271 entries into lane and laneq entries.
18272 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
18273 Remove AdvSIMD scalar modes.
18274 (aarch64_sq<r>dmulh_laneq<mode>): New.
18275 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
18277 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
18278 builtin implementations to relfect changes in RTL in aarch64-simd.md.
18279 * config/aarch64/iterators.md (VCOND): New.
18282 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
18285 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
18286 Add NULL LIBNAME argument to existing definitions.
18287 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
18288 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
18289 * config/avr/avr.c (DEF_BUILTIN): Same.
18290 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
18291 (avr_expand_builtin): Expand to a vanilla call if a libgcc
18292 implementation is available (DECL_ASSEMBLER_NAME is set).
18293 (avr_fold_absfx): New static function.
18294 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
18295 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
18296 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
18297 AVR_BUILTIN_ABSLLK.
18298 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
18299 (abshk, absk, abslk, absllk): Provide as static inline functions.
18301 2013-01-25 Marek Polacek <polacek@redhat.com>
18303 PR tree-optimization/56035
18304 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
18306 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
18308 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
18309 (*movtf_internal_rex64): Add (!o,C) alternative
18310 (*movxf_internal_rex64): Ditto.
18311 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
18313 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
18315 * doc/invoke.texi: fix typo.
18316 * doc/objc.texi: fix typo.
18318 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
18320 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
18321 for the first two alternatives.
18323 2013-01-24 Diego Novillo <dnovillo@google.com>
18325 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
18326 (ggc-zone.o): Remove.
18327 * configure.ac: Remove option --with-gc.
18328 * configure: Re-generate.
18329 * doc/install.texi: Remove documentation for --with-gc.
18330 * gengtype.c (write_enum_defn): Remove. Update all users.
18331 (write_Types_process_field): Remove generation of gt_e_* argument.
18332 (output_type_enum): Remove. Update all users.
18333 (write_enum_defn): Remove. Update all users.
18334 (enum alloc_zone): Remove. Update all users.
18335 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
18336 * ggc-common.c (ggc_splay_alloc): Remove first argument.
18337 Update all callers.
18338 (struct ptr_data): Remove field TYPE. Update all users.
18339 (gt_pch_note_object): Remove argument TYPE. Update all users.
18340 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
18342 * ggc-none.c (ggc_alloc_typed_stat): Remove.
18343 (struct alloc_zone): Remove.
18344 (ggc_internal_alloc_zone_stat): Remove.
18345 (ggc_internal_cleared_alloc_zone_stat): Remove.
18346 * ggc-page.c (ggc_alloc_typed_stat): Remove.
18347 (ggc_pch_count_object): Remove last argument. Update all users.
18348 (ggc_pch_alloc_object): Remove last argument. Update all users.
18349 (struct alloc_zone): Remove.
18350 * ggc-zone.c: Remove.
18351 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
18352 (struct alloc_zone): Remove.
18353 (ggc_alloc_typed_stat): Remove.
18354 (ggc_alloc_typed): Remove.
18355 (ggc_splay_alloc): Remove first argument.
18356 (rtl_zone): Remove. Update all users.
18357 (tree_zone): Remove. Update all users.
18358 (tree_id_zone): Remove. Update all users.
18359 (ggc_internal_zone_alloc_stat): Remove. Update all users.
18360 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
18361 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
18362 * tree-ssanames.c: Remove references to zone allocator in comments.
18364 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
18366 * config/avr/avr.c (avr_out_fract): Make register numbers that
18367 might be outside of source operand signed.
18369 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
18371 * config/i386/constraints.md (Yf): New constraint.
18372 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
18373 of f constraint to conditionaly disable x87 register preferences.
18374 (*movdf_internal): Ditto.
18375 (*movsf_internal): Ditto.
18377 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
18379 PR inline-asm/55934
18380 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
18381 that have operands with impossible constraints.
18382 Add a FIXME for a speed-up opportunity.
18383 * lra-constraints.c (process_alt_operands): Verify that a class
18384 selected from constraints on asms is valid for the operand mode.
18385 (curr_insn_transform): Remove incorrect comment.
18387 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
18389 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
18390 TOC operand is a valid symbol ref in the constant pool.
18392 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18394 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
18396 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
18399 * config/avr/stdfix.h: New file.
18400 * t-avr (stdfix-gcc.h): New rule to build it.
18401 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
18403 2013-01-23 Kostya Serebryany <kcc@google.com>
18405 * config/darwin.h: remove dependency on
18406 CoreFoundation (asan on Mac OS).
18408 2013-01-23 Jakub Jelinek <jakub@redhat.com>
18411 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
18412 instead of cmpdi_operand for first comparison operand.
18413 Don't assert that comparison operands aren't both constants.
18415 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18417 * doc/install.texi (Downloading the Source): Update references to
18418 downloading separate components.
18420 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
18422 * doc/extend.texi (__int128): Improve grammar.
18424 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
18427 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
18428 alternative to (o,r).
18429 (*movdi_internal_rex64): Remove (!o,n) alternative.
18430 (DImode immediate->memory splitter): Remove.
18431 (DImode immediate->memory peephole2): Remove.
18432 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
18433 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
18434 alternative to (!o,*r).
18435 (*movtf_internal_sse): New pattern.
18436 (*movxf_internal_rex64): New pattern.
18437 (*movxf_internal): Disable for TARGET_64BIT.
18438 (*movdf_internal_rex64): Remove (!o,F) alternative.
18440 2013-01-22 Jakub Jelinek <jakub@redhat.com>
18442 PR middle-end/56074
18443 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
18444 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
18445 * tree-vect-loop-manip.c (find_loop_location): Also ignore
18446 stmt locations where LOCATION_LOCUS of the stmt location is
18447 UNKNOWN_LOCATION or BUILTINS_LOCATION.
18450 * config/i386/i386.md (UNSPEC_STOS): New.
18451 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
18452 *strsetqi_1): Add UNSPEC_STOS.
18454 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
18457 * doc/invoke.texi: Remove left over -Wsynth example.
18459 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18461 PR tree-optimization/56051
18462 * fold-const.c (fold_binary_loc): Don't fold
18463 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
18464 a narrowing conversion, or widening conversion from signed
18467 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18469 PR rtl-optimization/56023
18470 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
18471 dependent on debug instruction.
18473 2013-01-21 Martin Jambor <mjambor@suse.cz>
18475 PR middle-end/56022
18476 * function.c (allocate_struct_function): Call
18477 invoke_set_current_function_hook earlier.
18479 2013-01-21 Jakub Jelinek <jakub@redhat.com>
18481 * reload1.c (init_reload): Only initialize reload_obstack
18482 during the first call.
18484 2013-01-21 Marek Polacek <polacek@redhat.com>
18486 * cfgloop.c (verify_loop_structure): Fix up grammar.
18488 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
18490 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
18491 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
18493 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18496 * config/arm/marvell-pj4.md: Update copyright year.
18497 Fix up use of alu to alu_reg and simple_alu_imm.
18499 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
18501 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
18503 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
18506 * lra-constraints.c (curr_insn_transform): Don't reuse original
18507 insn for secondary memory move when memory mode should be different.
18509 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18511 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
18512 atomic_storedi_1): New patterns.
18514 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
18516 btver2 pipeline descriptions.
18517 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
18519 * config/i386/i386.md (btver2_decode): New type attributes.
18520 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
18522 * config/i386/btver2.md: New file describing btver2 pipelines.
18524 2013-01-19 Andrew Pinski <apinski@cavium.com>
18526 PR tree-optimization/52631
18527 * tree-ssa-sccvn (visit_use): Before looking up the original
18528 statement, try looking up the simplified expression.
18530 2013-01-19 Anthony Green <green@moxielogic.com>
18532 * config/moxie/moxie.c (moxie_expand_prologue): Set
18533 current_function_static_stack_size.
18535 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18537 PR tree-optimization/56029
18538 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
18539 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
18541 2013-01-18 Sharad Singhai <singhai@google.com>
18543 PR tree-optimization/55995
18544 * dumpfile.c (dump_loc): Print location only if available.
18545 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
18547 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
18550 * lra-constraints.c (curr_insn_transform): Reuse original insn for
18551 secondary memory move.
18552 (inherit_reload_reg): Use rclass instead of cl for
18553 check_secondary_memory_needed_p.
18555 2013-01-18 Jakub Jelinek <jakub@redhat.com>
18557 PR middle-end/56015
18558 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
18559 the case where writing real complex part of target modifies op1.
18561 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
18563 * config/aarch64/aarch64-simd.md
18564 (aarch64_vcond_internal<mode>): Handle unordered cases.
18565 * config/aarch64/iterators.md (v_cmp_result): New.
18567 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
18568 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18570 * config/arm/marvell-pj4.md: New file.
18571 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
18572 * config/arm/arm.md (generic_sched): Add marvell_pj4.
18573 (generic_vfp): Likewise.
18574 * config/arm/arm-cores.def: Add marvell-pj4.
18575 * config/arm/arm-tune.md: Regenerate.
18576 * config/arm/arm-tables.opt: Regenerate.
18577 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
18578 * doc/invoke.texi: Document marvell-pj4.
18580 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
18582 * config/aarch64/arm_neon.h: Map scalar types to standard types.
18584 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18589 * var-tracking.c (negative_power_of_two_p): New.
18590 (global_get_addr_cache, local_get_addr_cache): New.
18591 (get_addr_from_global_cache, get_addr_from_local_cache): New.
18592 (vt_canonicalize_addr): Rewrite using the above. Adjust the
18594 (vt_stack_offset_p): Remove.
18595 (vt_canon_true_dep): Always canonicalize loc's address.
18596 (clobber_overlapping_mems): Make sure we have a MEM.
18597 (local_get_addr_clear_given_value): New.
18598 (val_reset): Clear local cached entries.
18599 (compute_bb_dataflow): Create and release the local cache.
18600 Disable duplicate MEMs clobbering.
18601 (emit_notes_in_bb): Clobber MEMs likewise.
18602 (vt_emit_notes): Create and release the local cache.
18603 (vt_initialize, vt_finalize): Create and release the global
18604 cache, respectively.
18605 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
18607 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18609 PR libmudflap/53359
18610 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
18611 not found in the symtab.
18613 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
18616 PR rtl-optimization/55547
18617 PR rtl-optimization/53827
18620 * alias.c (offset_overlap_p): New, factored out of...
18621 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
18622 the conservative special case for symbolic constants. Don't
18623 adjust zero sizes on alignment.
18625 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
18627 PR rtl-optimization/52573
18628 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
18629 REG_UNUSED for the same register.
18631 2013-01-17 Richard Biener <rguenther@suse.de>
18632 Marek Polacek <polacek@redhat.com>
18634 PR rtl-optimization/55833
18635 * loop-unswitch.c (unswitch_loops): Move loop verification...
18636 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
18637 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
18638 Set it to true when we're removing a loop from hierarchy tree in
18639 an irreducible region.
18640 (fix_bb_placements): Adjust caller.
18641 (fix_loop_placements): Likewise.
18643 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
18645 * config/avr/builtins.def (DEF_BUILTIN): Factor out
18646 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
18647 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
18648 Remove ID. Adjust comments.
18649 * config/avr/avr-c.c (avr_builtin_name): Remove.
18650 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
18651 * config/avr/avr.c (avr_tolower): New static function.
18652 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
18653 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
18654 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
18657 2013-01-17 Jan Hubicka <jh@suse.cz>
18659 PR tree-optimization/55273
18660 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
18662 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
18665 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
18666 store through atomic_store<mode>_1.
18667 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
18669 2013-01-17 Martin Jambor <mjambor@suse.cz>
18671 PR tree-optimizations/55264
18672 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
18673 for virtual methods.
18674 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
18675 virtual methods before inlining is over.
18676 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
18678 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
18681 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
18683 PR rtl-optimization/56005
18684 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
18685 pending reads for prefetch.
18687 2013-01-16 Ian Bolton <ian.bolton@arm.com>
18689 * config/aarch64/aarch64.md
18690 (*cstoresi_neg_uxtw): New pattern.
18691 (*cmovsi_insn_uxtw): New pattern.
18692 (*<optab>si3_uxtw): New pattern.
18693 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
18694 (*<optab>si3_insn_uxtw): New pattern.
18695 (*bswapsi2_uxtw): New pattern.
18697 2013-01-16 Richard Biener <rguenther@suse.de>
18699 * tree-inline.c (tree_function_versioning): Remove set but
18700 never used variable.
18702 2013-01-16 Richard Biener <rguenther@suse.de>
18704 PR tree-optimization/55964
18705 * tree-flow.h (rename_variables_in_loop): Remove.
18706 (rename_variables_in_bb): Likewise.
18707 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
18708 (copy_loop_before): Adjust and delete update-ssa status.
18709 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
18710 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
18711 (rename_variables_in_loop): Remove.
18712 (slpeel_update_phis_for_duplicate_loop): Likewise.
18713 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
18714 use available cfg machinery instead of duplicating it.
18715 Update PHI nodes and perform poor-mans SSA update here.
18716 (slpeel_tree_peel_loop_to_edge): Adjust.
18718 2013-01-16 Richard Biener <rguenther@suse.de>
18720 PR tree-optimization/54767
18721 PR tree-optimization/53465
18722 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
18723 (vrp_visit_phi_node): For PHI arguments coming via backedges
18724 drop all symbolical range information.
18725 (execute_vrp): Compute backedges.
18727 2013-01-16 Richard Biener <rguenther@suse.de>
18729 * doc/install.texi: Update CLooG and ISL requirements to
18732 2013-01-16 Christian Bruel <christian.bruel@st.com>
18735 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
18736 (broken_move): Handle UNSPECV_SP_SWITCH_B.
18737 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
18739 2013-01-16 DJ Delorie <dj@redhat.com>
18741 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
18742 (UNSPECV_SP_SWITCH_E): New.
18743 (sp_switch_1): Change to an unspec.
18744 (sp_switch_2): Change to an unspec. Don't use post-inc when we
18747 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
18749 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
18750 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
18751 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
18752 (expand_mem_thread_fence): Ditto.
18753 (expand_mem_signal_fence): Ditto.
18754 (expand_atomic_load): Ditto.
18755 (expand_atomic_store): Ditto.
18757 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
18759 PR rtl-optimization/55547
18760 PR rtl-optimization/53827
18763 * alias.c (memrefs_conflict_p): Set sizes to negative after
18766 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18769 * function.c (thread_prologue_and_epilogue_insns): Always
18770 add crtl->drap_reg to set_up_by_prologue.set, even if
18771 stack_realign_drap is false.
18773 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
18775 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
18776 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
18777 *call): Fix indention.
18779 2013-01-15 Tom de Vries <tom@codesourcery.com>
18782 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
18785 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
18787 PR rtl-optimization/55153
18788 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
18790 2013-01-15 Martin Jambor <mjambor@suse.cz>
18792 PR tree-optimization/55920
18793 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
18794 accesses as grp_to_be_debug_replaced.
18796 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18798 PR tree-optimization/55920
18799 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
18800 there is non-useless type conversion needed from debug rhs to lhs,
18801 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
18803 2013-01-15 Joseph Myers <joseph@codesourcery.com>
18804 Mikael Pettersson <mikpe@it.uu.se>
18807 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
18809 (ASM_OUTPUT_CASE_LABEL): Remove.
18810 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
18811 * final.c (shorten_branches): Update alignment of labels before
18812 jump tables if CASE_VECTOR_SHORTEN_MODE.
18814 2013-01-15 Richard Biener <rguenther@suse.de>
18817 * system.h: Do not include gmp.h for building host tools.
18819 2013-01-15 Richard Biener <rguenther@suse.de>
18821 PR middle-end/55882
18822 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
18823 account for bitpos when computing alignment.
18825 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
18827 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
18828 (ix86_target_macros_internal): Likewise.
18830 * config/i386/i386.c (m_CORE2I7): Removed.
18831 (m_CORE_HASWELL): New macro.
18832 (m_CORE_ALL): Likewise.
18833 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
18834 (initial_ix86_arch_features): Likewise.
18835 (processor_target_table): Initializations for Core avx2.
18836 (cpu_names): New names "core-avx2".
18837 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
18838 PROCESSOR_CORE_HASWELL.
18839 (ix86_issue_rate): New case.
18840 (ia32_multipass_dfa_lookahead): Likewise.
18841 (ix86_sched_init_global): Likewise.
18843 * config/i386/i386.h (TARGET_HASWELL): New macro.
18844 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
18845 (processor_type): New PROCESSOR_HASWELL.
18847 2013-01-15 Jakub Jelinek <jakub@redhat.com>
18849 PR tree-optimization/55955
18850 * tree-vect-loop.c (vectorizable_reduction): Give up early on
18851 *SHIFT_EXPR and *ROTATE_EXPR codes.
18853 PR tree-optimization/48766
18854 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
18855 -ftrapv disable -fwrapv.
18857 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18860 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
18861 etc. to 1 and not to __flash.
18862 Use LL suffix for __INT24_MAX__ with -mint8.
18863 Use ULL suffix for __UINT24_MAX__ with -mint8.
18865 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18867 * config/avr/avr-arch.h
18868 (struct base_arch_s): Use typedef avr_arch_t instead.
18869 (struct arch_info_s): Use typedef avr_arch_info_t instead.
18870 (struct mcu_type_s): Use typedef avr_mcu_t instead.
18871 * config/avr/avr.c: Same.
18872 * config/avr/avr-devices.c: Same.
18873 * config/avr/driver-avr.c: Same.
18874 * config/avr/gen-avr-mmcu-texi.c: Same.
18875 * config/avr/avr-mcus.def: Adjust comment.
18877 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
18879 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
18880 * config/aarch64/iterators.md (VALLDI): New.
18882 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
18883 Andi Kleen <ak@linux.intel.com>
18886 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
18887 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
18890 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18892 * config/avr/avr-stdint.h: Remove trailing blanks.
18893 * config/avr/avr-log.h: Same.
18894 * config/avr/avr-arch.h: Same.
18895 * config/avr/avr-devices.c: Same.
18896 * config/avr/avr-dimode.md: Same.
18897 * config/avr/predicates.md: Same.
18898 * config/avr/avr-c.c: Same. And fix typo.
18900 * config/avr/avr-protos.h: Same. And:
18901 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
18902 (init_cumulative_args): Rename to avr_init_cumulative_args.
18903 (expand_prologue): Rename to avr_expand_prologue.
18904 (expand_epilogue): Rename to avr_expand_epilogue.
18905 (adjust_insn_length): Rename to avr_adjust_insn_length.
18906 (notice_update_cc): Rename to avr_notice_update_cc.
18907 (final_prescan_insn): Rename to avr_final_prescan_insn.
18908 * config/avr/avr.c: Same.
18909 * config/avr/avr.h: Same.
18910 * config/avr/avr.md: Remove trailing blanks.
18911 (prologue): Use avr_expand_prologue.
18912 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
18914 2013-01-14 Richard Biener <rguenther@suse.de>
18916 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
18917 verify_location, collect_subblocks): New functions.
18918 (verify_gimple_in_cfg): Verify that locations only reference
18919 BLOCKs in the functions BLOCK tree.
18921 2013-01-14 Richard Biener <rguenther@suse.de>
18923 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
18925 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
18927 (insert_out_of_ssa_copy_on_edge): Likewise.
18928 (rewrite_close_phi_out_of_ssa): Likewise.
18929 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
18931 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
18932 propagated constants.
18933 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
18936 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
18938 * config/avr/avr-modes.def: Add GPL copyright notice.
18940 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
18942 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
18943 MEMMODEL_MASK to determine memory model.
18944 (atomic_store<mode>): Ditto from operands[2].
18945 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
18947 2013-01-13 Jakub Jelinek <jakub@redhat.com>
18950 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
18951 (fold_gimple_assign): Don't call unshare_expr here.
18952 (fold_ctor_reference): Call unshare_expr.
18954 2013-01-13 Terry Guo <terry.guo@arm.com>
18956 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
18957 * doc/fragments.texi: Document MULTILIB_REUSE.
18958 * gcc.c (multilib_reuse): New internal spec.
18959 (set_multilib_dir): Also search multilib from multilib_reuse.
18960 * genmultilib (tmpmultilib3): Refactor code.
18961 (tmpmultilib4): Ditto.
18962 (multilib_reuse): New multilib argument.
18964 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
18966 * Makefile.in: Update copyright.
18968 2013-01-12 Tom de Vries <tom@codesourcery.com>
18970 PR middle-end/55890
18971 * calls.c (expand_call): Check if arg_nr is valid.
18973 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18975 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
18976 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
18977 documentation. Add missing '__' in front of
18978 __builtin_ia32_packssdw256.
18980 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18983 * config/s390/s390.c (s390_preferred_reload_class): Do not return
18984 NO_REGS for larl operands.
18985 (s390_reload_larl_operand): Use s390_load_address instead of
18988 2013-01-11 Richard Biener <rguenther@suse.de>
18990 * tree-cfg.c (verify_node_sharing_1): Split out from ...
18991 (verify_node_sharing): ... here.
18992 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
18994 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
18996 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
18997 Substitute TREECHECKING.
18998 * configure: Regenerate.
18999 * Makefile.in (TREECHECKING): New.
19001 2013-01-11 Richard Guenther <rguenther@suse.de>
19003 PR tree-optimization/44061
19004 * tree-vrp.c (extract_range_basic): Compute zero as
19005 value-range for __builtin_constant_p of function parameters.
19007 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
19009 Update copyright years.
19011 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
19013 PR rtl-optimization/55672
19014 * lra-eliminations.c (mark_not_eliminable): Permit addition with
19015 const to be eliminable.
19017 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
19019 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
19020 * configure: Regenerate.
19022 2013-01-10 Richard Biener <rguenther@suse.de>
19024 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
19026 2013-01-10 Richard Biener <rguenther@suse.de>
19029 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
19030 locations for virtual PHI arguments.
19031 (rewrite_update_phi_arguments): Likewise.
19033 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
19035 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
19038 2013-01-10 Jakub Jelinek <jakub@redhat.com>
19040 PR tree-optimization/55921
19041 * tree-complex.c (expand_complex_asm): New function.
19042 (expand_complex_operations_1): Call it for GIMPLE_ASM.
19044 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19047 * config/s390/s390.c (s390_symref_operand_p)
19048 (s390_loadrelative_operand_p): Merge the two functions.
19049 (s390_check_qrst_address, print_operand_address): Add parameters
19050 to s390_loadrelative_operand_p invokation.
19051 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
19052 (s390_reload_larl_operand, s390_secondary_reload): Use
19053 s390_loadrelative_operand_p instead of s390_symref_operand_p.
19054 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
19056 2013-01-09 Mike Stump <mikestump@comcast.net>
19058 * dse.c (record_store): Remove unnecessary assert.
19060 2013-01-09 Jan Hubicka <jh@suse.cz>
19062 PR tree-optimization/55569
19063 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
19064 * cfgloop.h (scale_loop_profile): Likewise.
19066 2013-01-09 Jan Hubicka <jh@suse.cz>
19069 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
19071 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
19073 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
19075 PR middle-end/55114
19076 * expr.h (maybe_emit_group_store): Declare.
19077 * expr.c (maybe_emit_group_store): New function.
19078 * builtins.c (expand_builtin_int_roundingfn): Call it.
19079 (expand_builtin_int_roundingfn_2): Likewise.
19081 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
19083 PR rtl-optimization/55829
19084 * lra-constraints.c (match_reload): Add code for absent output.
19085 (curr_insn_transform): Add code for reloads of matched inputs
19088 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
19090 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
19091 attribute of movddup insn to DF.
19092 (*vec_interleave_lowv2df): Ditto.
19093 (vec_dupv2df): Ditto.
19095 2013-01-09 Jan Hubicka <jh@suse.cz>
19097 PR tree-optimiation/55875
19098 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
19099 EVERY_ITERATION parameter.
19100 (number_of_iterations_exit): Check if exit is executed every iteration.
19101 (idx_infer_loop_bounds): Similarly here.
19102 (n_of_executions_at_most): Simplify
19103 to only test for cases where statement is dominated by the
19104 particular bound; handle correctly the "postdominance" test.
19105 (scev_probably_wraps_p): Use max loop iterations info
19106 as a global bound first.
19108 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
19109 Nick Clifton <nickc@redhat.com>
19111 * config/v850/v850.md (cbranchsf4): New pattern.
19112 (cstoresf4): New pattern.
19113 (cbranchdf4): New pattern.
19114 (cstoredf4): New pattern.
19115 (movsicc): Disallow floating point comparisons.
19116 (cmpsf_le_insn): Fix order of operators.
19117 (cmpsf_lt_insn): Likewise.
19118 (cmpsf_eq_insn): Likewise.
19119 (cmpdf_le_insn): Likewise.
19120 (cmpdf_lt_insn): Likewise.
19121 (cmpdf_eq_insn): Likewise.
19122 (cmpsf_ge_insn): Use LE comparison.
19123 (cmpdf_ge_insn): Likewise.
19124 (cmpsf_gt_insn): Use LT comparison.
19125 (cmpdf_gt_insn): Likewise.
19126 (cmpsf_ne_insn): Delete pattern.
19127 (cmpdf_ne_insn): Delete pattern.
19128 * config/v850/v850.c (v850_gen_float_compare): Use
19129 gen_cmpdf_eq_insn for NE comparison.
19130 (v850_float_z_comparison_operator)
19131 (v850_float_nz_comparison_operator): Move from here ...
19132 * config/v850/predicates.md: ... to here. Move GT and GE
19133 comparisons into v850_float_z_comparison_operator.
19134 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
19136 (v850_float_nz_comparison_operator): Likewise.
19138 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19140 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
19141 with calls to gen_insvsi/gen_insvdi.
19143 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
19145 * config/i386/i386.c (initial_ix86_tune_features): Set up
19146 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
19148 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
19149 Jakub Jelinek <jakub@redhat.com>
19151 PR tree-optimization/48189
19152 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
19153 If nitercst is 0, don't predict the exit edge.
19155 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19157 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
19158 in asm_fprintf with reg_names.
19159 (aarch64_print_operand_address): Likewise.
19160 (aarch64_return_addr): Likewise.
19161 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
19163 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19165 * config/pa/pa.h (VAL_U6_BITS_P): Define.
19166 (INT_U6_BITS): Likewise.
19167 * config/pa/predicates.md (uint6_operand): New predicate.
19168 (shift5_operand, shift6_operand): Likewise.
19169 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
19171 (lshrdi3): Use shift6_operand.
19172 (shrpsi4, shrpdi4): New insn patterns.
19173 (extzv): Delete expander.
19174 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
19175 predicates in unamed zero extract patterns. Tighten common constraint.
19176 (extv): Delete expander.
19177 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
19178 predicates in unamed sign extract patterns. Tighten common constraint.
19179 (insv): Delete expander.
19180 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
19181 predicates in unamed insert patterns. Tighten common constraint.
19182 Change uint32_operand predicate to uint6_operand predicate in unamed
19183 DImode pattern to insert constant values of type 1...1xxxx.
19185 2013-01-04 Jan Hubicka <jh@suse.cz>
19187 PR tree-optimization/55823
19188 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
19191 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19192 Uros Bizjak <ubizjak@gmail.com>
19194 PR rtl-optimization/55845
19195 * df-problems.c (can_move_insns_across): Stop scanning at
19196 volatile_insn_p source instruction or give up if
19197 across_from .. across_to range contains any volatile_insn_p
19200 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19202 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
19203 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
19205 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
19206 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
19208 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19211 * asan.c (asan_clear_shadow): New function.
19212 (asan_emit_stack_protection): Use it.
19214 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
19216 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
19217 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
19218 with tab instead of space.
19220 2013-01-08 Nick Clifton <nickc@redhat.com>
19222 * config/rl78/rl78.c (rl78_expand_prologue): Always select
19223 register bank 0 at the start of an interrupt handler.
19224 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
19227 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19229 * config/aarch64/aarch64-simd.md
19230 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
19231 (aarch64_simd_bsl): Likewise.
19232 (aarch64_vcond_internal<mode>): Likewise.
19233 (vcond<mode><mode>): Likewise.
19234 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
19235 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
19237 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
19239 * config/aarch64/aarch64-builtins.c
19240 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
19242 2013-01-08 Martin Jambor <mjambor@suse.cz>
19245 * tree-sra.c (analyze_access_subtree): Return true also after
19246 potentially creating a debug-only replacement.
19248 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19250 PR middle-end/55890
19251 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
19253 PR tree-optimization/54120
19254 * tree-vrp.c (range_fits_type_p): Don't allow
19255 src_precision < precision from signed vr to unsigned_p
19256 if vr->min or vr->max is negative.
19257 (simplify_float_conversion_using_ranges): Test can_float_p
19258 against CODE_FOR_nothing.
19260 2013-01-08 Jakub Jelinek <jakub@redhat.com>
19261 Richard Biener <rguenther@suse.de>
19263 PR middle-end/55851
19264 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
19265 types instead of just INTEGER_TYPE types.
19267 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
19269 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
19272 2013-01-07 Steve Ellcey <sellcey@mips.com>
19275 * config/mips/mips.opt: Change mad to mmad to match documentation.
19277 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19280 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
19281 .progmemx.data now.
19283 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19286 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
19287 (avr_addrspace_t): Add .section_name field.
19288 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
19290 (avr_addrspace): Same. Initialize .section_name. Remove last
19291 NULL entry. Put __memx into .progmemx.data.
19292 (progmem_section_prefix): Remove.
19293 (avr_asm_init_sections): No need to initialize progmem_section.
19294 (avr_asm_named_section): Use avr_addrspace[].section_name to get
19295 section name prefix.
19296 (avr_asm_select_section): Ditto. And use get_unnamed_section to
19297 retrieve the progmem section.
19298 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
19299 boundary to run over avr_addrspace[].
19300 (avr_register_target_pragmas): Ditto.
19302 2013-01-06 Jakub Jelinek <jakub@redhat.com>
19304 * varasm.c (output_constant_def_contents): For asan_protect_global
19305 protected strings, adjust DECL_ALIGN if needed, before testing for
19307 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
19308 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
19310 (output_object_block): For asan protected decls, emit asan padding
19311 after their contents.
19312 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
19313 (asan_finish_file): Test it here instead.
19315 2013-01-07 Nick Clifton <nickc@redhat.com>
19316 Matthias Klose <doko@debian.org>
19317 Doug Kwan <dougkwan@google.com>
19318 H.J. Lu <hongjiu.lu@intel.com>
19321 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
19323 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
19325 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
19327 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
19329 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
19331 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19334 * doc/install.texi (Cross-Compiler-Specific Options): Document
19337 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
19339 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
19340 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
19341 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
19342 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
19343 vqmovun_high_s64): Fix source operand number and update copyright.
19345 2013-01-07 Richard Biener <rguenther@suse.de>
19347 PR middle-end/55890
19348 * gimple.h (gimple_call_builtin_p): New overload.
19349 * gimple.c (validate_call): New function.
19350 (gimple_call_builtin_p): Likewise.
19351 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
19352 Use gimple_call_builtin_p.
19353 (find_func_clobbers): Likewise.
19354 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
19355 (strlen_optimize_stmt): Likewise.
19357 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19359 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
19360 (vld1q_dup_*): Likewise.
19361 (vld1_*): Likewise.
19362 (vld1q_*): Likewise.
19363 (vld1_lane_*): Likewise.
19364 (vld1q_lane_*): Likewise.
19366 2013-01-07 Richard Biener <rguenther@suse.de>
19368 * lto-streamer.h (LTO_minor_version): Bump to 2.
19370 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
19372 * config/aarch64/aarch64-protos.h
19373 (aarch64_const_double_zero_rtx_p): Rename to...
19374 (aarch64_float_const_zero_rtx_p): ...this.
19375 (aarch64_float_const_representable_p): New.
19376 (aarch64_output_simd_mov_immediate): Likewise.
19377 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
19378 move immediate case.
19379 * config/aarch64/aarch64.c
19380 (aarch64_const_double_zero_rtx_p): Rename to...
19381 (aarch64_float_const_zero_rtx_p): ...this.
19382 (aarch64_print_operand): Allow printing of new constants.
19383 (aarch64_valid_floating_const): New.
19384 (aarch64_legitimate_constant_p): Check for valid floating-point
19386 (aarch64_simd_valid_immediate): Likewise.
19387 (aarch64_vect_float_const_representable_p): New.
19388 (aarch64_float_const_representable_p): Likewise.
19389 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
19390 (aarch64_output_simd_mov_immediate): New.
19391 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
19392 (*movdf_aarch64): Likewise.
19393 * config/aarch64/constraints.md (Ufc): New.
19394 (Y): call aarch64_float_const_zero_rtx.
19395 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
19397 2013-01-07 Richard Biener <rguenther@suse.de>
19399 PR tree-optimization/55888
19400 PR tree-optimization/55862
19401 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
19402 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
19403 not if it is contained therein.
19405 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19407 * config/avr/t-avr: Typo.
19409 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
19412 * config/avr/t-avr: Don't automatically rebuild
19413 $(srcdir)/config/avr/t-multilib
19414 $(srcdir)/config/avr/avr-tables.opt
19415 $(srcdir)/doc/avr-mmcu.texi
19416 (avr-mcus): New phony target to build them on request.
19417 (s-avr-mlib, s-avr-mmcu-texi): Remove.
19418 * avr/avr-mcus.def: Adjust comments.
19420 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
19422 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
19424 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19426 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
19428 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
19430 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
19432 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
19434 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
19435 to generate profiling.
19436 * config/rs6000/aix64.h (LIB_SPEC): Same.
19438 2013-01-04 Andrew Pinski <apinski@cavium.com>
19440 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
19442 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
19444 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
19446 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
19448 (ix86_expand_move): Ditto.
19449 (ix86_zero_extend_to_Pmode): Ditto.
19450 (ix86_expand_call): Ditto.
19451 (ix86_expand_special_args_builtin): Ditto.
19452 (ix86_expand_builtin): Ditto.
19454 2013-01-04 Richard Biener <rguenther@suse.de>
19456 PR tree-optimization/55862
19457 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
19458 translating them through PHI nodes.
19460 2013-01-04 Martin Jambor <mjambor@suse.cz>
19462 PR tree-optimization/55755
19463 * tree-sra.c (sra_modify_assign): Do not check that an access has no
19464 children when trying to avoid producing a VIEW_CONVERT_EXPR.
19466 2013-01-04 Marek Polacek <polacek@redhat.com>
19468 PR middle-end/55859
19469 * opts.c (default_options_optimization): Clarify error message.
19471 2013-01-04 Richard Biener <rguenther@suse.de>
19473 PR middle-end/55863
19474 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
19477 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19480 * config/pa/pa.md (movsi): Revert previous change.
19481 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
19484 2013-01-03 Richard Henderson <rth@redhat.com>
19486 * config/i386/i386.c (ix86_expand_move): Always assign to op1
19487 after eliminating TLS symbols.
19489 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
19492 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
19493 * graphite-poly.c (debug_gmp_value): Likewise.
19495 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
19498 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
19499 selected code model, define __code_mode_small__, __code_model_medium__,
19500 __code_model_large__, __code_model_32__ or __code_model_kernel__.
19501 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
19502 xchg temporary register with %k. Declare temporary register as
19504 [__x86_64__]: For medium and large code models, preserve %rbx register.
19506 2013-01-03 Richard Biener <rguenther@suse.de>
19508 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
19509 (dump_subscript): Adjust.
19510 (finalize_ddr_dependent): Do not dump redundant info.
19511 (analyze_siv_subscript): Adjust.
19512 (subscript_dependence_tester): Likewise.
19513 (compute_affine_dependence): Likewise.
19515 2013-01-03 Richard Biener <rguenther@suse.de>
19518 2013-01-03 Richard Biener <rguenther@suse.de>
19520 PR tree-optimization/55857
19521 * tree-vect-stmts.c (vectorizable_load): Do not setup
19522 re-alignment for invariant loads.
19524 2013-01-02 Richard Biener <rguenther@suse.de>
19526 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19527 invariant load do not generate a vector load from the scalar location.
19529 2013-01-03 Richard Biener <rguenther@suse.de>
19531 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
19532 for not vectorizing.
19533 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
19534 not build INDIRECT_REFs, call get_name once only.
19535 (vect_create_data_ref_ptr): Likewise. Dump base object kind
19536 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
19538 2013-01-03 Richard Biener <rguenther@suse.de>
19540 PR tree-optimization/55857
19541 * tree-vect-stmts.c (vectorizable_load): Do not setup
19542 re-alignment for invariant loads.
19544 2013-01-03 Richard Biener <rguenther@suse.de>
19547 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
19548 prefer a built-in decl.
19550 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19552 * gcc.c (process_command): Update copyright notice dates.
19553 * gcov.c (print_version): Likewise.
19554 * gcov-dump.c (print_version): Likewise.
19556 PR rtl-optimization/55838
19557 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
19558 iv0.step, iv1.step and step.
19560 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19561 Marc Glisse <marc.glisse@inria.fr>
19563 PR tree-optimization/55832
19564 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
19565 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
19566 integer_{one,zero}_node.
19568 2013-01-03 Jakub Jelinek <jakub@redhat.com>
19571 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
19572 * var-tracking.c (reverse_op): Don't add reverse ops to
19573 VALUEs that have already
19574 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
19576 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
19578 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
19580 2013-01-02 Teresa Johnson <tejohnson@google.com>
19582 * dumpfile.c (dump_loc): Print filename with location.
19583 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
19584 new location_t parameter to emit complete unroll message with
19585 new dump framework.
19586 (canonicalize_loop_induction_variables): Compute loops location
19587 and pass to try_unroll_loop_completely.
19588 * loop-unroll.c (report_unroll_peel): New function.
19589 (peel_loops_completely): Use new dump format with location
19590 for main dumpfile message, and invoke report_unroll_peel on success.
19591 (decide_unrolling_and_peeling): Ditto.
19592 (decide_peel_once_rolling): Remove old dumpfile message subsumed
19593 by report_unroll_peel.
19594 (decide_peel_completely): Ditto.
19595 (decide_unroll_constant_iterations): Ditto.
19596 (decide_unroll_runtime_iterations): Ditto.
19597 (decide_peel_simple): Ditto.
19598 (decide_unroll_stupid): Ditto.
19599 * cfgloop.c (get_loop_location): New function.
19600 * cfgloop.h (get_loop_location): Declare.
19602 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19604 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
19607 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19609 PR middle-end/55198
19610 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
19611 BLKmode objects when EXPAND_MEMORY is specified.
19613 2013-01-02 Sriraman Tallam <tmsriram@google.com>
19615 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
19617 (fold_builtin_cpu): Do not share cpu model decls across statements.
19619 2013-01-02 Jason Merrill <jason@redhat.com>
19622 * tree.c (build_array_type_1): Revert earlier change.
19624 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
19626 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
19628 * config/aarch64/aarch64-tune.md: Re-generate.
19630 2013-01-02 Richard Biener <rguenther@suse.de>
19632 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
19633 invariant load do not generate a vector load from the scalar location.
19635 2013-01-02 Richard Biener <rguenther@suse.de>
19638 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
19639 * configure: Regenerate.
19641 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19643 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
19644 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
19645 (expand_builtin_int_roundingfn_2): Keep the original target around
19646 for the fallback case.
19648 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
19650 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
19651 to be clear for sign changes.
19653 2013-01-01 Jan Hubicka <jh@suse.cz>
19655 * ipa-inline-analysis.c: Fix formatting.
19657 2013-01-01 Jakub Jelinek <jakub@redhat.com>
19659 PR tree-optimization/55831
19660 * tree-vect-loop.c (get_initial_def_for_induction): Use
19661 gsi_after_labels instead of gsi_start_bb.
19663 Copyright (C) 2013 Free Software Foundation, Inc.
19665 Copying and distribution of this file, with or without modification,
19666 are permitted in any medium without royalty provided the copyright
19667 notice and this notice are preserved.