1 2013-11-12 Marek Polacek <polacek@redhat.com>
3 * final.c (update_alignments): Initialize label to NULL_RTX.
5 2013-11-12 Jeff Law <law@redhat.com>
7 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
8 (insert_trap_and_remove_trailing_statements): New argument OP which
9 is the NULL pointer. Emit the trap after the load/store through
10 the NULL pointer. Simplify the RHS of a store through a NULL pointer
11 when trivial to do so.
12 (isolate_path): Corresponding changes.
13 (gimple_ssa_isolate_erroneous_path): Likewise.
15 2013-11-12 Teresa Johnson <tejohnson@google.com>
16 Jan Hubicka <jh@suse.cz>
18 * predict.c (drop_profile): New function.
19 (handle_missing_profiles): Ditto.
20 (counts_to_freqs): Don't overwrite estimated frequencies
21 when function has no profile counts.
22 * predict.h (handle_missing_profiles): Declare.
23 * tree-inline.c (freqs_to_counts): New function.
24 (copy_cfg_body): Invoke freqs_to_counts as needed.
25 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
27 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
30 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
32 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
34 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
37 * config/i386/i386.c (ix86_option_override_internal): Check
38 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
39 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
40 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
41 MASK_AVX256_SPLIT_UNALIGNED_STORE.
43 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
44 Clear m_COREI7_AVX and update comments.
45 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
47 2013-11-12 Martin Jambor <mjambor@suse.cz>
49 PR rtl-optimization/10474
50 * ira.c (interesting_dest_for_shprep): New function.
51 (split_live_ranges_for_shrink_wrap): Likewise.
52 (find_moveable_pseudos): Move calculation of dominance info,
53 df_analysios and the final anlyses to...
54 (ira): ...here, call split_live_ranges_for_shrink_wrap.
56 2013-11-12 Bin Cheng <bin.cheng@arm.com>
58 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
59 Handle type conversion.
61 2013-11-11 Martin Liska <marxin.liska@gmail.com>
62 Jan Hubicka <jh@suse.cz>
64 * cgraph.c (dump_cgraph_node): Profile dump added.
65 * cgraph.h (struct cgraph_node): New time profile variable added.
66 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
67 * gcov-io.h (gcov_type): New profiler type introduced.
68 * ipa-profile.c (lto_output_node): Streaming for time profile added.
69 (input_node): Time profiler is read from LTO stream.
70 * predict.c (maybe_hot_count_p): Hot prediction changed.
71 * profile.c (instrument_values): New case for time profiler added.
72 (compute_value_histograms): Read of time profile.
73 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
74 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
75 (gimple_init_edge_profiler): TP function instrumentation.
76 (gimple_gen_time_profiler): New.
77 * value-prof.c (gimple_add_histogram_value): Support for time profiler
79 (dump_histogram_value): TP type added to dumps.
80 (visit_hist): More sensitive check that takes TP into account.
81 (gimple_find_values_to_profile): TP instrumentation.
82 * value-prof.h (hist_type): New histogram type added.
83 (struct histogram_value_t): Pointer to struct function added.
84 * libgcc/Makefile.in: New GCOV merge function for TP added.
85 * libgcov.c: function_counter variable introduced.
86 (_gcov_merge_time_profile): New.
87 (_gcov_time_profiler): New.
89 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
90 Jeff Law <law@redhat.com>
92 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
93 ao_ref_init_from_ptr_and_size for builtins.
95 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
96 H.J. Lu <hongjiu.lu@intel.com>
99 * config/i386/x86-tune.def
100 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
101 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
103 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
104 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
105 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
106 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
107 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
108 misaligned_prologue_used. Check that
109 desired_aling <= epilogue_size_needed.
111 2013-11-11 Cong Hou <congh@google.com>
113 PR tree-optimization/59050
114 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
116 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
119 * expmed.c (emit_store_flag): Fail for const-const comparison.
121 2013-11-11 Tristan Gingold <gingold@adacore.com>
122 Eric Botcazou <ebotcazou@adacore.com>
124 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
125 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
126 * tree.def (CONSTRUCTOR): Likewise.
127 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
128 * gimplify.c (gimplify_init_constructor): Do not clear the object when
129 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
131 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
133 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
136 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
138 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
141 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
143 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
145 2013-11-08 Jeff Law <law@redhat.com>
147 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
148 threading paths first, then perform PHI node checks if applicable.
150 2013-11-10 Karlson2k <k2k@narod.ru>
151 Kai Tietz <ktietz@redhat.com>
154 * configure.ac: Adding for exported symbols check
155 and for rdynamic-check executable-extension.
156 * configure: Regenerated.
158 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
160 * mode-switching.c (optimize_mode_switching): Mark block as
161 nontransparent, if last_mode at block exit is different from no_mode.
163 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
165 * function.c (NAME__MAIN): Move to...
166 * cfgexpand.c (NAME__MAIN): ...here.
168 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
170 * target.def (can_use_doloop_p): New hook.
171 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
172 * doc/tm.texi: Regenerate.
173 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
174 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
175 * hooks.c (hook_bool_dint_dint_uint_true): New function.
176 * targhooks.h (can_use_doloop_if_innermost): Declare.
177 * targhooks.c (can_use_doloop_if_innermost): New function.
178 * target.h: Include double-int.h.
179 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
180 Remove iteration count, maximum iteration count, loop depth and
181 enter-at-top inputs from doloop_begin and doloop_end.
182 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
184 * config/arc/arc.c (arc_can_use_doloop_p): New function.
185 (TARGET_CAN_USE_DOLOOP_P): Define.
186 * config/arm/thumb2.md (doloop_end): Update for new interface.
187 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
188 * config/bfin/bfin.md (doloop_end): Update for new interface.
189 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
190 (TARGET_CAN_USE_DOLOOP_P): Define.
191 * config/c6x/c6x.md (doloop_end): Update for new interface.
192 * config/ia64/ia64.md (doloop_end): Update for new interface.
193 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
194 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
196 * config/mep/mep.c (mep_emit_doloop): Likewise.
197 (TARGET_CAN_USE_DOLOOP_P): Define.
198 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
199 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
200 * config/s390/s390.md (doloop_end): Update for new interface.
201 * config/sh/sh.md (doloop_end): Likewise.
202 * config/spu/spu.md (doloop_end): Likewise.
203 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
204 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
205 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
206 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
207 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
208 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
210 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
212 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
215 * doc/extend.texi: Move Cilk Plus Builtins node before Other
218 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
219 Joseph Myers <joseph@codesourcery.com>
221 * ginclude/stdatomic.h: New file.
222 * Makefile.in (USER_H): Add stdatomic.h.
224 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
226 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
229 2013-11-08 Jeff Law <law@redhat.com>
231 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
232 * tree-ssa-threadupdate.c (delete_thread_path): New function.
233 (ssa_redirect_edges, thread_block_1): Use it.
234 (thread_through_loop_header, mark_threaded_blocks): Likewise.
235 (thread_through_all_blocks, register_jump_thread): Likewise.
236 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
238 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
240 * config/arm/aarch-common.c
241 (search_term): New typedef.
242 (shift_rtx_costs): New array.
243 (arm_rtx_shift_left_p): New.
244 (arm_find_sub_rtx_with_search_term): Likewise.
245 (arm_find_sub_rtx_with_code): Likewise.
246 (arm_early_load_addr_dep): Add sanity checking.
247 (arm_no_early_alu_shift_dep): Likewise.
248 (arm_no_early_alu_shift_value_dep): Likewise.
249 (arm_no_early_mul_dep): Likewise.
250 (arm_no_early_store_addr_dep): Likewise.
252 2013-11-08 Richard Biener <rguenther@suse.de>
254 PR tree-optimization/59047
255 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
257 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
259 * common.opt (fcheck-pointer-bounds): Move to ...
260 * c-family/c.opt: ... here.
261 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
262 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
263 * langhooks.h (lang_hooks): Remove chkp_supported field.
264 * toplev.c (process_options): Remove chkp_supported check.
266 2013-11-08 Richard Biener <rguenther@suse.de>
268 PR tree-optimization/59038
269 PR tree-optimization/58955
270 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
271 previous change. Handle known dependences correctly.
273 2013-11-08 Tom de Vries <tom@codesourcery.com>
275 * config/rs6000/t-xilinx: Remove duplicate contents.
277 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
278 Joseph Myers <joseph@codesourcery.com>
280 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
281 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
282 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
283 (struct tree_base): Add atomic_flag field.
284 * tree.h (TYPE_ATOMIC): New accessor macro.
285 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
286 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
287 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
288 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
289 * tree.c (set_type_quals): Set TYPE_ATOMIC.
290 (find_atomic_core_type): New function.
291 (build_qualified_type): Adjust alignment for qualified types.
292 (build_atomic_base): New function
293 (build_common_tree_nodes): Build atomicQI_type_node,
294 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
296 * print-tree.c (print_node): Print atomic qualifier.
297 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
298 * target.def (atomic_assign_expand_fenv): New hook.
299 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
300 * doc/tm.texi: Regenerate.
301 * targhooks.c (default_atomic_assign_expand_fenv): New function.
302 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
303 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
304 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
306 * config/i386/i386.c (enum ix86_builtins): Add
307 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
309 (bdesc_special_args): Add __builtin_ia32_fnstenv,
310 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
311 (ix86_expand_builtin): Handle the new built-in functions.
312 (ix86_atomic_assign_expand_fenv): New function.
313 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
314 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
315 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
316 (fnstenv, fldenv, fnstsw, fnclex): New insns.
318 2013-11-07 Steve Ellcey <sellcey@mips.com>
320 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
321 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
322 (MULTILIB_DIRNAMES): Add fp64 directory.
323 (MULTILIB_EXCEPTIONS): Add new exclusions.
325 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
327 * gimplify.c (gimple_regimplify_operands): Do not set
329 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
330 (rewrite_close_phi_out_of_ssa): Same.
331 (rewrite_phi_out_of_ssa): Same.
332 (rewrite_degenerate_phi): Same.
333 (handle_scalar_deps_crossing_scop_limits): Same.
334 * tree-if-conv.c (predicate_scalar_phi): Same.
335 * tree-parloops.c (create_loads_for_reductions): Same.
336 (create_final_loads_for_reduction): Same.
337 (create_loads_and_stores_for_name): Same.
338 (transform_to_exit_first_loop): Same.
339 (create_parallel_loop): Same.
341 (move_computations_dom_walker::before_dom_children): Same.
342 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
343 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
344 * tree-ssa-propagate.c (substitute_and_fold): Same.
345 * tree-vect-loop.c (vect_finalize_reduction): Same.
346 * tree-vect-stmts.c (vectorizable_call): Same.
348 2013-11-07 Mike Stump <mikestump@comcast.net>
350 * config/pdp11/pdp11.c: Include dbxout.h.
351 * config/picochip/picochip.c: Likewise.
353 2013-11-07 Cong Hou <congh@google.com>
355 PR tree-optimization/56764
356 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
357 Combine alias checks if it is possible to amortize the runtime
358 overhead. Return the number of alias checks after merging.
359 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
360 Use the function vect_create_cond_for_alias_checks () to check
361 the number of alias checks.
363 2013-11-07 Jeff Law <law@redhat.com>
365 * varpool.c (ctor_for_folding): Fix typo in comment.
367 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
369 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
370 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
372 2013-11-07 Jeff Law <law@redhat.com>
374 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
376 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
377 No longer check if we have __builtin_trap, assume it's available.
379 2013-11-07 Diego Novillo <dnovillo@google.com>
381 * attribs.c (lookup_scoped_attribute_spec): Make static.
382 (get_attribute_namespace): Likewise.
383 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
384 (validate_arglist): Move earlier in the file. Make static.
385 (expand_stack_restore): Move from stmt.c
386 (expand_stack_save): Move from stmt.c
387 (rewrite_call_expr_array): Move earlier in the file.
388 (rewrite_call_expr_valist): Likewise.
389 * cfgexpand.c: Include hard-reg-set.h before tree.h
392 (expand_asm_loc): Move from stmt.c.
393 (n_occurrences): Move from stmt.c.
394 (check_operand_nalternatives): Move from stmt.c.
395 (tree_conflicts_with_clobbers_p): Move from stmt.c.
396 (expand_asm_operands): Move from stmt.c
397 (expand_asm_stmt): Move from stmt.c
398 (expand_computed_goto): Move from stmt.c
399 (expand_goto): Move from stmt.c
400 (expand_null_return_1): Move from stmt.c
401 (expand_null_return): Move from stmt.c
402 (expand_value_return): Move from stmt.c
403 (expand_return): Move from stmt.c
404 (expand_main_function): Move from function.c
405 (stack_protect_prologue): Move from function.c
406 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
407 (build_function_decl_skip_args): Move from tree.c.
408 * explow.c (tree_expr_size): Move from tree.c.
409 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
410 (fields_length): Move from tree.c.
411 * fold-const.c (size_low_cst): Move from tree.c.
412 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
413 (tree_expr_nonzero_p): Make static. Move earlier in the file.
414 (fold_build3_initializer_loc): Remove.
415 (tree_invalid_nonnegative_warnv_p): Make static.
416 * function.c (expand_main_function): Move to cfgexpand.c.
417 (stack_protect_prologue): Move to cfgexpand.c.
418 (set_insn_locations): Move earlier in the file.
419 * gimple-fold.c: Include langhooks.h.
420 (truth_type_for): Move from tree.c.
421 * print-tree.c (print_vec_tree): Remove.
422 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
423 (expand_goto): Move to cfgexpand.c.
424 (n_occurrences): Move to cfgexpand.c.
425 (expand_asm_loc): Move to cfgexpand.c
426 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
427 (expand_asm_operands): Move to cfgexpand.c.
428 (expand_asm_stmt): Move to cfgexpand.c.
429 (check_operand_nalternatives): Move to cfgexpand.c
430 (expand_null_return): Move to cfgexpand.c.
431 (expand_value_return): Move to cfgexpand.c.
432 (expand_null_return_1): Move to cfgexpand.c.
433 (expand_return): Move to cfgexpand.c.
434 (expand_stack_save): Move to builtins.c.
435 (expand_stack_restore): Move to builtins.c
436 * symtab.c: Include output.h.
437 (decl_assembler_name_hash): Move from tree.c.
438 (decl_assembler_name_equal): Move from tree.c.
439 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
440 * tree-eh.c (in_array_bounds_p): Move from tree.c.
441 (range_in_array_bounds_p): Move from tree.c.
442 * tree-object-size.c (fini_object_sizes): Make static.
443 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
444 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
445 * tree.c (decl_assembler_name_equal): Move to symtab.c.
446 (tree_expr_size): Move to explow.c.
447 (decl_assembler_name_hash): Move to symtab.c.
449 (tree_expr_size): Move to explow.c.
450 (stabilize_reference_1): Move earlier in the file. Make static.
451 (omp_remove_redundant_declare_simd_attrs): Remove.
452 (simple_cst_list_equal): Move earlier in the file. Make static.
453 (size_low_cst): Move to fold-const.c.
454 (build_type_no_quals): Remove.
455 (build_function_type_skip_args): Move to cgraphclones.c.
456 (build_function_decl_skip_args): Move to cgraphclones.c.
457 (in_array_bounds_p): Move to tree-eh.c.
458 (range_in_array_bounds_p): Move to tree-eh.c.
459 (truth_type_for): Move to gimple-fold.c.
460 (list_equal_p): Remove.
461 * tree.h (decl_assembler_name_equal): Remove.
462 (decl_assembler_name_hash): Remove.
463 (truth_type_for): Remove.
464 (build_type_no_quals): Remove.
465 (build_function_decl_skip_args): Remove.
466 (in_array_bounds_p): Remove.
467 (range_in_array_bounds_p): Remove.
468 (size_low_cst): Remove.
469 (omp_remove_redundant_declare_simd_attrs): Remove.
470 (tree_expr_size): Remove.
471 (fields_length): Remove.
472 (stabilize_reference_1): Remove.
473 (expand_goto): Remove.
474 (expand_stack_save): Remove.
475 (expand_stack_restore): Remove.
476 (expand_return): Remove.
477 (fold_build3_initializer_loc): Remove.
478 (tree_expr_nonzero_p): Remove.
479 (tree_invalid_nonnegative_warnv_p): Remove.
480 (tree_expr_nonzero_warnv_p): Remove.
481 (fold_builtin_snprintf_chk): Remove.
482 (validate_arglist): Remove.
483 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
484 (simple_cst_list_equal): Remove.
486 (expand_main_function): Remove.
487 (stack_protect_prologue): Remove.
488 (print_vec_tree): Remove.
489 (lookup_scoped_attribute_spec): Remove.
490 (get_attribute_namespace): Remove.
491 (expand_computed_goto): Remove.
492 (expand_asm_stmt): Remove.
493 (list_equal_p): Remove.
494 (ssa_name_nonnegative_p): Remove.
495 (fini_object_sizes): Remove.
496 (addr_expr_of_non_mem_decl_p): Remove.
497 (is_tm_safe_or_pure): Move to trans-mem.c.
498 (more_const_call_expr_args_p): Remove.
499 (save_vtable_map_decl): Remove.
501 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
503 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
506 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
507 link to libgomp and its dependencies.
508 * config/ia64/hpux.h (LIB_SPEC): Likewise.
509 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
510 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
511 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
513 2013-11-07 Jakub Jelinek <jakub@redhat.com>
515 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
516 (determine_value_range): Add loop argument. Use get_range_info to
518 (bound_difference): Adjust caller.
520 2013-11-07 Richard Biener <rguenther@suse.de>
521 Jakub Jelinek <jakub@redhat.com>
523 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
524 latches from header PHI arguments from the latch edge.
526 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
529 * varasm.c (output_constant): Handle NULLPTR_TYPE.
531 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
533 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
534 misaligned_prologue_used when it has been set.
536 2013-11-07 Yury Gribov <y.gribov@samsung.com>
537 Jakub Jelinek <jakub@redhat.com>
540 * asan.c (get_mem_refs_of_builtin_call): Allow
541 integer literals as addresses in instrumented builtins.
543 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
545 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
546 Explain why plus_constant is not used.
548 2013-11-07 Richard Biener <rguenther@suse.de>
550 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
551 (canonicalize_value): ... this. Also handle stripping of
553 (get_value, set_lattice_value, get_value_for_expr): Adjust.
554 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
555 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
557 2013-11-07 Richard Biener <rguenther@suse.de>
559 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
561 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
564 * config/i386/i386.md (push peepholer/splitter): Use Pmode
565 with stack_pointer_rtx.
567 2013-11-07 Bin Cheng <bin.cheng@arm.com>
569 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
570 using operand_equal_p.
572 2013-11-07 Bin Cheng <bin.cheng@arm.com>
574 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
575 * tree-affine.c (get_inner_reference_aff): Return base.
576 * tree-affine.h (get_inner_reference_aff): Change prototype.
578 2013-11-06 Tobias Burnus <burnus@net-b.de>
580 * doc/invoke.texi (Wdate-time): Fix typo.
582 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
584 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
585 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
586 fp_arith_reg_operand instead of arith_reg_operand.
588 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
590 * config/sh/sh.md (adddi3): Remove empty constraints.
591 Remove can_create_pseudo_p and arith_reg_operand check.
592 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
595 2013-11-06 Jeff Law <law@redhat.com>
596 Tom Tromey <tromey@redhat.com>
598 * gdbinit.in: Disable strict type checking.
600 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
602 * tree-pass.h (make_pass_live_range_shrinkage): New external.
603 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
604 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
605 (rest_of_handle_live_range_shrinkage): Ditto
606 (class pass_live_range_shrinkage): Ditto.
607 (pass_data_live_range_shrinkage): Ditto.
608 (make_pass_live_range_shrinkage): Ditto.
609 * sched-int.h (initialize_live_range_shrinkage): New prototype.
610 (finish_live_range_shrinkage): Ditto.
611 * sched-deps.c (create_insn_reg_set): Make void return value.
612 * passes.def: Add pass_live_range_shrinkage.
613 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
614 * haifa-sched.c (live_range_shrinkage_p): New.
615 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
617 (rank_for_schedule): Add code for pressure relief through live
619 (schedule_insn): Print more debug info.
620 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
621 through live range shrinkage.
622 * doc/invoke.texi (-flive-range-shrinkage): New.
623 * common.opt (flive-range-shrinkage): New.
625 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
628 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
629 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
630 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
631 RTXes that return in AVX256 register.
633 2013-11-06 Richard Biener <rguenther@suse.de>
635 PR tree-optimization/58653
636 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
637 (prepare_initializers_chain): Adjust.
639 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
641 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
642 (gimple_in_transaction): Use bb_in_transaction.
643 * basic-block.h (bb_in_transaction): Relocate here and rename.
644 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
646 2013-11-06 Richard Biener <rguenther@suse.de>
648 * tree.c (drop_tree_overflow): New function.
649 * tree.h (drop_tree_overflow): Declare.
650 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
651 * tree-vrp.c (range_int_cst_singleton_p): Use
652 is_overflow_infinity instead of testing TREE_OVERFLOW.
653 (extract_range_from_assert): Likewise.
654 (zero_nonzero_bits_from_vr): Likewise.
655 (extract_range_basic): Likewise.
656 (register_new_assert_for): Use drop_tree_overflow.
657 (vrp_visit_phi_node): Likewise.
659 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
661 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
662 checking for leaf functions without dynamic stack allocation.
663 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
664 (ia64_expand_prologue): Likewise.
665 * config/mips/mips.c (mips_expand_prologue): Likewise.
666 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
667 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
668 (sparc_flat_expand_prologue): Likewise.
670 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
672 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
673 (__ST3_LANE_FUNC): Likewise.
674 (__ST4_LANE_FUNC): Likewise.
676 2013-11-06 Nick Clifton <nickc@redhat.com>
678 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
679 name returned by msp430_mcu_name.
680 (LIB_SPEC): If a -T option has not been specified then set a
681 default, mcu-specific, linker script.
682 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
683 * config/msp430/msp430.c (msp430x_names): Likewise.
684 Alpha sort the names for ease of comparison.
685 (msp430_mcu_name): New function: Returns a string suitable for
686 use as a C preprocessor symbol based upon the name of the MCU
688 (msp430_option_override): Accept msp430x and msp430xv2 as generic
690 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
692 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
693 unless it is the last part of a spec substring.
694 * doc/invoke.texi (Spec Files): Document space insertion
697 2013-11-06 Christian Bruel <christian.bruel@st.com>
699 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
700 Factorize probabilities, Use adjust_address instead of
701 adjust_automodify_address when possible. Enable for optimize.
702 (sh_expand_strlen): New function.
703 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
704 * config/sh/sh.md (strlensi): New pattern.
705 (UNSPEC_BUILTIN_STRLEN): Define.
707 2013-11-06 Jakub Jelinek <jakub@redhat.com>
710 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
711 (expand_assignment): If *bitpos is negative, set *offset
712 and adjust *bitpos, so that it is not negative.
714 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
716 * config/i386/bdver3.md : Added two additional decoder units
717 to support issue rate of 4 and remodeled vector unit.
718 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
719 architectures is set to 4.
720 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
721 lookahead is set to 4 for BD architectures.
723 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
725 * config/rs6000/rs6000.c (rs6000_option_override_internal):
726 Remove restriction against use of VSX instructions when generating
727 code for little endian mode.
729 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
731 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
732 for both big and little endian.
733 (mulv8hi3): Swap input operands for merge high and merge low
734 instructions for little endian.
736 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
738 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
739 define_insn to define_expand that uses even patterns for big
740 endian and odd patterns for little endian.
741 (vec_widen_smult_even_v16qi): Likewise.
742 (vec_widen_umult_even_v8hi): Likewise.
743 (vec_widen_smult_even_v8hi): Likewise.
744 (vec_widen_umult_odd_v16qi): Likewise.
745 (vec_widen_smult_odd_v16qi): Likewise.
746 (vec_widen_umult_odd_v8hi): Likewise.
747 (vec_widen_smult_odd_v8hi): Likewise.
748 (altivec_vmuleub): New define_insn.
749 (altivec_vmuloub): Likewise.
750 (altivec_vmulesb): Likewise.
751 (altivec_vmulosb): Likewise.
752 (altivec_vmuleuh): Likewise.
753 (altivec_vmulouh): Likewise.
754 (altivec_vmulesh): Likewise.
755 (altivec_vmulosh): Likewise.
757 2013-11-05 Mike Stump <mikestump@comcast.net>
759 * Makefile.in (mostlyclean): Remove c-family objects.
761 2013-11-05 Ian Lance Taylor <iant@google.com>
763 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
764 If possible, add .cfi directives to record change to bx.
765 * config/i386/i386.c (ix86_emit_cfi): New function.
766 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
768 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
770 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
771 immediately after a label for a tablejump pattern.
773 * config/arm/arm.c (is_jump_table): Remove.
774 (create_fix_barrier): Use tablejump_p instead.
775 (arm_reorg): Likewise.
776 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
777 (thumb2_output_casesi): Likewise.
778 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
779 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
780 casesi_shift_media, casesi_load_media): Likewise.
781 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
782 * config/microblaze/microblaze.md: Likewise.
784 2013-11-05 Tobias Burnus <burnus@net-b.de>
786 * doc/invoke.texi (-Wdate-time): Document.
788 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
790 * double-int.c (lshift_double, rshift_double): Remove
791 SHIFT_COUNT_TRUNCATED handling.
793 2013-11-05 Jeff Law <law@redhat.com>
795 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
796 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
797 * gimple-ssa-isolate-paths.c: New file.
798 * gimple.c (check_loadstore): New function.
799 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
800 Verify OP is in the argument list and the argument corresponding
801 to OP is a pointer type. Use operand_equal_p rather than
802 pointer equality when testing if OP is on the nonnull list.
803 Use check_loadstore rather than count_ptr_derefs. Handle
804 GIMPLE_RETURN statements.
805 * tree-vrp.c (infer_nonnull_range): Remove.
806 * gimple.h (infer_nonnull_range): Declare.
807 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
808 * passes.def: Add pass_isolate_erroneous_paths.
809 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
810 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
811 * tree-ssa.c (struct count_ptr_d): Remove.
812 (count_ptr_derefs, count_uses_and_derefs): Remove.
813 * tree-ssa.h (count_uses_and_derefs): Remove.
815 2013-11-05 Jakub Jelinek <jakub@redhat.com>
817 PR rtl-optimization/58997
818 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
819 get_iv_value to be in iv->mode rather than iv->extend_mode.
820 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
821 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
822 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
825 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
827 * gimple.h: Move some prototypes to gimple-expr.h and add to include
829 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
830 Move to gimple-expr.h.
831 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
832 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
833 is_gimple_id, virtual_operand_p, is_gimple_addressable,
834 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
836 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
837 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
838 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
839 is_gimple_address, is_gimple_invariant_address,
840 is_gimple_ip_invariant_address, is_gimple_min_invariant,
841 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
842 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
843 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
844 gimple_decl_printable_name, useless_type_conversion_p,
845 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
847 * gimple-expr.c: New File.
848 (useless_type_conversion_p, gimple_set_body, gimple_body,
849 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
850 gimple_can_coalesce_p, extract_ops_from_tree_1,
851 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
852 is_gimple_address, is_gimple_invariant_address,
853 is_gimple_ip_invariant_address, is_gimple_min_invariant,
854 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
855 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
856 is_gimple_mem_ref_addr): Relocate here.
857 * Makefile.in (OBJS): Add gimple-expr.o.
859 2013-11-05 David Malcolm <dmalcolm@redhat.com>
861 * gengtype-parse.c (struct_field_seq): Support empty structs.
863 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
865 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
867 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
869 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
871 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
873 (LONG_DOUBLE_TYPE_SIZE): New define.
874 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
876 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
879 * doc/extend.texi [visibility ("visibility_type")]: Add example
880 about visibility attribute on namespace declaration.
882 2013-11-05 Richard Biener <rguenther@suse.de>
885 * passes.def (all_passes): Start with pass_fixup_cfg again.
887 2013-11-05 Richard Biener <rguenther@suse.de>
889 PR tree-optimization/58955
890 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
893 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
895 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
897 (vec_pack_ufix_trunc_v2df): Likewise.
899 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
902 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
903 pattern, instead of word_mode.
905 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
907 (set_storage_via_setmem): Likewise.
909 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
911 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
912 gimple.h and the rest of the condition in eliminate_build.
913 (eliminate_build): Call new routine.
914 * gimple.h (phi_ssa_name_p): Delete.
916 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
918 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
919 case of no prefix and reserving zero slots, because when that's the
920 case we'll never get here.
921 * vec.h (va_heap::reserve): Don't try and handle
922 vec_prefix::calculate_allocation returning zero because that should
925 2013-11-05 Richard Biener <rguenther@suse.de>
928 * tree-dfa.c (get_ref_base_and_extent): Merge common code
929 in MEM_REF and TARGET_MEM_REF handling. Make sure to
930 process trailing array detection before diving into the
931 view-converted object (and possibly apply some extra offset).
933 2013-11-05 Joseph Myers <joseph@codesourcery.com>
935 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
937 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
939 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
941 PR tree-optimization/58958
942 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
943 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
945 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
947 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
948 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
950 2013-11-05 Jakub Jelinek <jakub@redhat.com>
952 PR tree-optimization/58984
953 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
954 set *SIZE_P if non-NULL on success.
955 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
957 (ipcp_transform_function): Likewise. Punt if size of access
958 is different from TYPE_SIZE on v->value's type.
960 2013-11-05 Tobias Burnus <burnus@net-b.de>
962 * doc/invoke.texi (-fopenmp-simd): Document new option.
963 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
964 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
965 * tree.c (attribute_value_equal): Ditto.
967 2013-11-04 Wei Mi <wmi@google.com>
969 * sched-rgn.c (add_branch_dependences): Keep insns in
970 a SCHED_GROUP at the end of BB to remain their location.
972 2013-11-04 Wei Mi <wmi@google.com>
974 * config/i386/i386.c (memory_address_length): Extract a part
975 of code to rip_relative_addr_p.
976 (rip_relative_addr_p): New Function.
977 (ix86_macro_fusion_p): Ditto.
978 (ix86_macro_fusion_pair_p): Ditto.
979 * config/i386/i386.h: Add new tune features about macro-fusion.
980 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
981 * doc/tm.texi: Generated.
982 * doc/tm.texi.in: Ditto.
983 * haifa-sched.c (try_group_insn): New Function.
984 (group_insns_for_macro_fusion): Ditto.
985 (sched_init): Call group_insns_for_macro_fusion.
986 * target.def: Add two hooks: macro_fusion_p and
989 2013-11-04 Kostya Serebryany <kcc@google.com>
991 Update to match the changed asan API.
992 * asan.c (asan_function_start): New function.
993 (asan_emit_stack_protection): Update the string stored in the
994 stack red zone to match new API. Store the PC of the current
995 function in the red zone.
996 (asan_global_struct): Update the __asan_global definition to match
998 (asan_add_global): Ditto.
999 * asan.h (asan_function_start): New prototype.
1000 * final.c (final_start_function): Call asan_function_start.
1001 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
1004 2013-11-04 Wei Mi <wmi@google.com>
1006 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
1007 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
1008 * config/i386/i386.c (ix86_option_override_internal): Ditto.
1009 (ix86_issue_rate): Ditto.
1010 (ix86_adjust_cost): Ditto.
1011 (ia32_multipass_dfa_lookahead): Ditto.
1012 (ix86_sched_init_global): Ditto.
1013 (get_builtin_code_for_version): Ditto.
1014 * config/i386/i386.h (enum target_cpu_default): Ditto.
1015 (enum processor_type): Ditto.
1016 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1018 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1020 PR rtl-optimization/58967
1021 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
1022 !lra_in_progress for mode sizes bigger word.
1024 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1026 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
1027 arguments to merge instruction for little endian.
1028 (vec_widen_umult_lo_v16qi): Likewise.
1029 (vec_widen_smult_hi_v16qi): Likewise.
1030 (vec_widen_smult_lo_v16qi): Likewise.
1031 (vec_widen_umult_hi_v8hi): Likewise.
1032 (vec_widen_umult_lo_v8hi): Likewise.
1033 (vec_widen_smult_hi_v8hi): Likewise.
1034 (vec_widen_smult_lo_v8hi): Likewise.
1036 2013-11-04 Ian Lance Taylor <iant@google.com>
1038 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
1039 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
1040 builtins that take pointers.
1041 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
1042 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
1043 (run_gcc): Pass -fnon-call-exceptions.
1045 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1047 * optabs.c (expand_vec_perm): Revert one incorrect line from
1050 PR tree-optimization/58978
1051 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
1052 use_stmt by single_imm_use directly. Only call single_imm_use
1055 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1057 PR rtl-optimization/58968
1058 * lra-spills.c (return_regno_p): New function.
1059 (lra_final_code_change): Use it.
1061 2013-11-04 Joseph Myers <joseph@codesourcery.com>
1063 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
1064 * target.def (float_exceptions_rounding_supported_p): New hook.
1065 * targhooks.c (default_float_exceptions_rounding_supported_p): New
1067 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
1068 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
1070 * doc/tm.texi: Regenerate.
1071 * config.gcc (powerpc*-*-linux*): Set extra_objs.
1072 * config/rs6000/rs6000-linux.c: New file.
1073 * config/rs6000/rs6000-protos.h
1074 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
1075 * config/rs6000/linux.h
1076 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
1077 * config/rs6000/linux64.h
1078 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
1079 * config/rs6000/t-linux (rs6000-linux.o): New rule.
1080 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
1082 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1084 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
1085 Replace the define_insn_and_split with a define_insn and two
1086 define_splits, with the split after reload re-permuting the source
1087 register to its original value.
1088 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
1089 (*vsx_le_perm_store_v8hi): Likewise.
1090 (*vsx_le_perm_store_v16qi): Likewise.
1092 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1094 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
1097 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1099 PR tree-optimization/58946
1100 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
1101 bbs with bbinfo[idx].op != NULL before all blocks with
1102 bbinfo[idx].op == NULL.
1104 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
1106 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
1107 (avr_dump_double_int_hex): Likewise.
1108 (avr_log_vadump): Remove %D and %X handling.
1109 * config/avr/avr.c (avr_double_int_push_digit): Delete.
1110 (avr_map_op_t): Change map from double_int to unsigned int.
1111 (avr_map_op): Update accordingly.
1112 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
1113 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
1114 unsigned ints rather than double_ints.
1116 2013-11-03 Marek Polacek <polacek@redhat.com>
1118 Implement -fsanitize=vla-bound.
1119 * opts.c (common_handle_option): Handle vla-bound.
1120 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
1121 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
1122 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
1124 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1126 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
1129 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
1131 * config/i386/constraints.md (Ts, Tv): New address constrains.
1132 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
1133 constraint for address_no_seg_operand.
1134 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
1135 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
1136 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
1137 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
1138 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
1139 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
1140 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
1141 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
1142 constraint for vsib_address_operand.
1144 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
1146 * gcse.c (pre_delete): Remove references to regmove from comments.
1147 * recog.c: (validate_replace_rtx_1): Likewise.
1148 * config/rl78/rl78.c: Likewise.
1149 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
1150 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
1151 * common/config/mmix/mmix-common.c: Likewise.
1153 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
1155 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
1156 * gimplify.c (gimplify_compound_lval): Likewise.
1157 * graphite-clast-to-gimple.c (gloog): Likewise.
1158 * graphite-dependences.c (loop_is_parallel_p): Likewise.
1159 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
1160 (limit_scop); Likewise.
1161 (build_scops): Likewise.
1162 (dot_scop): Likewise.
1163 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1164 (build_scop_drs): Likewise.
1165 (insert_stmts): Likewise.
1166 (insert_out_of_ssa_copy): Likewise.
1167 (remove_phi): Likewise.
1168 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
1169 * hw-doloop.c (discover_loop): Likewise.
1170 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
1171 * tree-dfa.c (dump_enumerated_decls): Likewise.
1172 * tree-if-conv.c (if_convertable_loop_p): Likewise.
1173 * tree-inline.c (tree_function_versioning): Likewise.
1174 * tree-loop-distribution.c (build_rdg): Likewise.
1175 (rdg_flag_vertex_and_dependent): Likewise.
1176 (distribute_loop): Likewise.
1177 * tree-parloops.c (loop_parallel_p): Likewise.
1178 (eliminate_local_variables): Likewise.
1179 (separate_decls_in_region): Likewise.
1180 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
1181 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
1182 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1183 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
1184 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
1185 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
1186 (vectorizable_condition): Likewise.
1188 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
1190 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
1191 * configure: Regenerate.
1192 * config/i386/i386.md (*movdi_internal): Change
1193 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
1194 (*movdf_internal): Ditto.
1195 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
1196 * config/i386/sse.md (vec_concatv2di): Output interunit movq
1197 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
1199 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
1201 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
1204 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1206 Automated part of renaming of symtab_node_base to symtab_node.
1208 Patch autogenerated by rename_symtab.py from
1209 https://github.com/davidmalcolm/gcc-refactoring-scripts
1210 revision 58bb219cc090b2f4516a9297d868c245495ee622
1211 with ChangeLog entry fixed up by hand.
1213 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
1215 (cgraph_node_for_asm): Likewise.
1216 * cgraph.h (symtab_node_base): Likewise.
1217 (cgraph_node): Likewise.
1218 (varpool_node): Likewise.
1219 (is_a_helper <cgraph_node>::test): Likewise.
1220 (is_a_helper <varpool_node>::test): Likewise.
1221 (symtab_nodes): Likewise.
1222 (symtab_register_node): Likewise.
1223 (symtab_unregister_node): Likewise.
1224 (symtab_remove_node): Likewise.
1225 (symtab_get_node): Likewise.
1226 (symtab_node_for_asm): Likewise.
1227 (symtab_node_asm_name): Likewise.
1228 (symtab_node_name): Likewise.
1229 (symtab_insert_node_to_hashtable): Likewise.
1230 (symtab_add_to_same_comdat_group): Likewise.
1231 (symtab_dissolve_same_comdat_group_list): Likewise.
1232 (dump_symtab_node): Likewise.
1233 (debug_symtab_node): Likewise.
1234 (dump_symtab_base): Likewise.
1235 (verify_symtab_node): Likewise.
1236 (verify_symtab_base): Likewise.
1237 (symtab_used_from_object_file_p): Likewise.
1238 (symtab_alias_ultimate_target): Likewise.
1239 (symtab_resolve_alias): Likewise.
1240 (fixup_same_cpp_alias_visibility): Likewise.
1241 (symtab_for_node_and_aliases): Likewise.
1242 (symtab_nonoverwritable_alias): Likewise.
1243 (availability symtab_node_availability): Likewise.
1244 (symtab_semantically_equivalent_p): Likewise.
1245 (fixup_same_cpp_alias_visibility): Likewise.
1246 (symtab_prevail_in_asm_name_hash): Likewise.
1248 (varpool): Likewise.
1249 (varpool_first_variable): Likewise.
1250 (varpool_next_variable): Likewise.
1251 (varpool_first_static_initializer): Likewise.
1252 (varpool_next_static_initializer): Likewise.
1253 (varpool_first_defined_variable): Likewise.
1254 (varpool_next_defined_variable): Likewise.
1255 (cgraph_first_defined_function): Likewise.
1256 (cgraph_next_defined_function): Likewise.
1257 (cgraph_first_function): Likewise.
1258 (cgraph_next_function): Likewise.
1259 (cgraph_first_function_with_gimple_body): Likewise.
1260 (cgraph_next_function_with_gimple_body): Likewise.
1261 (symtab_alias_target): Likewise.
1262 (symtab_real_symbol_p): Likewise.
1263 (symtab_can_be_discarded): Likewise.
1264 * cgraphbuild.c (mark_address): Likewise.
1265 (mark_load): Likewise.
1266 (mark_store): Likewise.
1267 * cgraphunit.c (decide_is_symbol_needed): Likewise.
1269 (enqueue_node): Likewise.
1270 (referred_to_p): Likewise.
1271 (cgraph_process_same_body_aliases): Likewise.
1272 (analyze_functions): Likewise.
1273 (handle_alias_pairs): Likewise.
1274 (output_weakrefs): Likewise.
1275 (compile): Likewise.
1276 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1277 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1278 * ipa-prop.c (remove_described_reference): Likewise.
1279 (try_decrement_rdesc_refcount): Likewise.
1280 (ipa_edge_duplication_hook): Likewise.
1281 * ipa-ref.c (ipa_record_reference): Likewise.
1282 (ipa_maybe_record_reference): Likewise.
1283 (ipa_clone_ref): Likewise.
1284 (ipa_clone_references): Likewise.
1285 (ipa_clone_referring): Likewise.
1286 (ipa_find_reference): Likewise.
1287 (ipa_remove_stmt_references): Likewise.
1288 (ipa_clear_stmts_in_references): Likewise.
1289 * ipa-ref.h (symtab_node_base): Likewise.
1290 (ipa_ref): Likewise.
1291 (ipa_record_reference): Likewise.
1292 (ipa_maybe_record_reference): Likewise.
1293 (ipa_clone_references): Likewise.
1294 (ipa_clone_referring): Likewise.
1295 (ipa_clone_ref): Likewise.
1296 (ipa_find_reference): Likewise.
1297 (ipa_remove_stmt_references): Likewise.
1298 (ipa_clear_stmts_in_references): Likewise.
1299 * ipa-reference.c (ipa_reference_write_optimization_summary):
1301 * ipa.c (enqueue_node): Likewise.
1302 (process_references): Likewise.
1303 (walk_polymorphic_call_targets): Likewise.
1304 (symtab_remove_unreachable_nodes): Likewise.
1305 (address_taken_from_non_vtable_p): Likewise.
1306 (comdat_can_be_unshared_p_1): Likewise.
1307 (comdat_can_be_unshared_p): Likewise.
1308 (can_replace_by_local_alias): Likewise.
1309 (function_and_variable_visibility): Likewise.
1310 * is-a.h: Likewise (within example in comment).
1311 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1312 (lto_symtab_encoder_encode): Likewise.
1313 (lto_symtab_encoder_delete_node): Likewise.
1314 (lto_symtab_encoder_in_partition_p): Likewise.
1315 (lto_set_symtab_encoder_in_partition): Likewise.
1316 (output_refs): Likewise.
1317 (compute_ltrans_boundary): Likewise.
1318 (output_symtab): Likewise.
1319 (input_node): Likewise.
1320 (input_ref): Likewise.
1321 (input_edge): Likewise.
1322 (input_cgraph_1): Likewise.
1323 (input_refs): Likewise.
1324 (output_cgraph_opt_summary): Likewise.
1325 (input_node_opt_summary): Likewise.
1326 (input_cgraph_opt_section): Likewise.
1327 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1329 * lto-streamer-out.c (lto_output): Likewise.
1330 (output_symbol_p): Likewise.
1331 (produce_symtab): Likewise.
1332 * lto-streamer.h (lto_encoder_entry): Likewise.
1333 (lto_free_function_in_decl_state_for_node): Likewise.
1334 (lto_symtab_encoder_encode): Likewise.
1335 (lto_symtab_encoder_delete_node): Likewise.
1336 (lto_symtab_encoder_in_partition_p): Likewise.
1337 (lto_set_symtab_encoder_in_partition): Likewise.
1338 (lto_symtab_encoder_lookup): Likewise.
1339 (lsei_node): Likewise.
1340 (lto_symtab_encoder_deref): Likewise.
1341 * symtab.c (symtab_hash): Likewise.
1342 (assembler_name_hash): Likewise.
1343 (symtab_nodes): Likewise.
1344 (hash_node): Likewise.
1345 (eq_node): Likewise.
1346 (hash_node_by_assembler_name): Likewise.
1347 (eq_assembler_name): Likewise.
1348 (insert_to_assembler_name_hash): Likewise.
1349 (unlink_from_assembler_name_hash): Likewise.
1350 (symtab_prevail_in_asm_name_hash): Likewise.
1351 (symtab_register_node): Likewise.
1352 (symtab_insert_node_to_hashtable): Likewise.
1353 (symtab_unregister_node): Likewise.
1354 (symtab_get_node): Likewise.
1355 (symtab_remove_node): Likewise.
1356 (symtab_initialize_asm_name_hash): Likewise.
1357 (symtab_node_for_asm): Likewise.
1358 (symtab_add_to_same_comdat_group): Likewise.
1359 (symtab_dissolve_same_comdat_group_list): Likewise.
1360 (symtab_node_asm_name): Likewise.
1361 (symtab_node_name): Likewise.
1362 (dump_symtab_base): Likewise.
1363 (dump_symtab_node): Likewise.
1364 (dump_symtab): Likewise.
1365 (debug_symtab_node): Likewise.
1366 (verify_symtab_base): Likewise.
1367 (verify_symtab_node): Likewise.
1368 (verify_symtab): Likewise.
1369 (symtab_used_from_object_file_p): Likewise.
1370 (symtab_node_availability): Likewise.
1371 (symtab_alias_ultimate_target): Likewise.
1372 (fixup_same_cpp_alias_visibility): Likewise.
1373 (symtab_resolve_alias): Likewise.
1374 (symtab_for_node_and_aliases): Likewise.
1375 (symtab_for_node_and_aliases): Likewise.
1376 (symtab_nonoverwritable_alias_1): Likewise.
1377 (symtab_nonoverwritable_alias): Likewise.
1378 (symtab_semantically_equivalent_p): Likewise.
1379 * value-prof.c (init_node_map): Likewise.
1380 * varasm.c (find_decl): Likewise.
1381 * varpool.c (varpool_node_for_asm): Likewise.
1382 (varpool_remove_unreferenced_decls): Likewise.
1384 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1386 Manual part of renaming of symtab_node_base to symtab_node.
1388 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1389 clashes with the preferred name for the base class.
1390 (const_symtab_node): Remove redundant typedef.
1392 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1394 * optabs.c (expand_vec_perm): Avoid vector mode punning
1395 SUBREGs in SET_DEST.
1396 * expmed.c (store_bit_field_1): Likewise.
1397 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1398 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1399 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1400 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1401 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1402 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1403 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1404 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1405 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1406 expand_vec_perm_vpshufb2_vpermq,
1407 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1408 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1409 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1410 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1413 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1415 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1416 Use gcc_unreachable for unhandled modes. Do not check results of
1417 expand_simple_binop. If not expanded to target, move the result.
1419 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1420 Shiva Chen <shiva0217@gmail.com>
1422 * config.gcc (nds32*-*-*): Add nds32 target.
1423 * config/nds32/nds32.c: New file.
1424 * config/nds32/nds32.h: New file.
1425 * config/nds32/nds32.md: New file.
1426 * config/nds32/constants.md: New file.
1427 * config/nds32/constraints.md: New file.
1428 * config/nds32/iterators.md: New file.
1429 * config/nds32/nds32-doubleword.md: New file.
1430 * config/nds32/nds32-intrinsic.md: New file.
1431 * config/nds32/nds32_intrinsic.h: New file.
1432 * config/nds32/nds32-modes.def: New file.
1433 * config/nds32/nds32-multiple.md: New file.
1434 * config/nds32/nds32.opt: New file.
1435 * config/nds32/nds32-opts.h: New file.
1436 * config/nds32/nds32-protos.h: New file.
1437 * config/nds32/nds32-peephole2.md: New file.
1438 * config/nds32/pipelines.md: New file.
1439 * config/nds32/predicates.md: New file.
1440 * config/nds32/t-mlibs: New file.
1441 * common/config/nds32: New directory and files.
1443 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1444 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1445 * doc/install.texi (Cross-Compiler-Specific Options): Document
1446 --with-nds32-lib for nds32 target.
1447 * doc/extend.texi (Function Attributes, Target Builtins): Document
1448 nds32 specific attributes.
1450 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1452 * lra-constraints (process_alt_operands): Use the result
1453 elimination register for operand when matching constraints.
1455 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1457 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1458 (remove_range_assertions): Call it.
1460 * tree.c (tree_ctz): New function.
1461 * tree.h (tree_ctz): New prototype.
1462 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1463 first argument from tree to const_tree.
1464 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1465 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1466 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1468 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1469 * tree-vect-loop.c (vect_analyze_loop_operations,
1470 vect_transform_loop): Don't force scalar loop for bound just because
1471 number of iterations is unknown, only do it if it is not known to be
1472 a multiple of vectorization_factor.
1473 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1475 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1476 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1477 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1478 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1479 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1480 a default def isn't partially constant.
1481 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1482 is known to be partially zero.
1483 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1484 to see if a default def isn't partially constant.
1485 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1486 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1488 (set_nonzero_bits, get_nonzero_bits): New functions.
1490 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1491 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1492 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1493 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1494 uses but in the condition and ASSERT_EXPR and the other successor of
1495 the predecessor bb is __builtin_unreachable (), set_range_info of the
1496 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1498 2013-10-31 Martin Jambor <mjambor@suse.cz>
1500 PR rtl-optimization/58934
1502 2013-10-30 Martin Jambor <mjambor@suse.cz>
1503 PR rtl-optimization/10474
1504 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1506 (interesting_dest_for_shprep): New function.
1507 (split_live_ranges_for_shrink_wrap): Likewise.
1508 (ira): Calculate dominance info and df analysis. Call
1509 split_live_ranges_for_shrink_wrap.
1511 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1512 Yury Gribov <y.gribov@samsung.com>
1515 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1516 shadow pointer to avoid clobbering the main one.
1518 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1520 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1522 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1525 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1526 it for calling update_costs_from_allocno.
1527 (assign_hard_reg): Call restore_costs_from_copies only for
1528 !retry_p. Pass new argument to update_costs_from_copies.
1529 (color_pass): Pass new argument to update_costs_from_copies.
1530 (ira_mark_allocation_change): Ditto.
1532 2013-10-30 Sharad Singhai <singhai@google.com>
1535 * opts.c (common_handle_option): Remove deprecated option
1536 -ftree-vectorizer-verbose.
1537 * doc/invoke.texi (Debugging Options): Ditto.
1538 * opts-global.c (handle_common_deferred_options): Ditto.
1539 (dump_remap_tree_vectorizer_verbose): Delete.
1540 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1542 2013-10-30 DJ Delorie <dj@redhat.com>
1544 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1545 that take no arguments.
1547 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1550 * reload1.c (update_eliminables_and_spill): New function, broken
1552 (reload): Use it. Check for frame size change after frame size
1553 alignment, and call update_eliminables_and_spill first if continue-ing.
1555 2013-10-30 Cong Hou <congh@google.com>
1558 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1559 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1560 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1562 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1564 PR rtl-optimization/58369
1565 * reload1.c (compute_reload_subreg_offset): New function.
1566 (choose_reload_regs): Use it to pass endian-correct
1567 offset to subreg_regno_offset.
1569 2013-10-30 Tobias Burnus <burnus@net-b.de>
1572 * tree-cfg.c (replace_loop_annotate): Replace warning by
1575 2013-10-30 Jason Merrill <jason@redhat.com>
1577 * configure.ac (loose_warn): Add -Wno-format if
1578 --disable-build-format-warnings.
1580 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1582 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1583 onto multiple lines to make things easier for rename_symtab.py.
1585 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1586 (symtab_semantically_equivalent_p): Likewise.
1588 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1591 * lra.c (check_rtl): Remove address check before LRA work.
1593 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1595 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1596 POINTER_PLUS_EXPR in the defining statement.
1598 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1600 * regmove.c: Remove.
1601 * tree-pass.h (make_pass_regmove): Remove.
1602 * timevar.def (TV_REGMOVE): Remove.
1603 * passes.def (pass_regmove): Remove.
1604 * opts.c (default_options_table): Remove entry for regmove.
1605 * doc/passes.texi: Remove regmove pass description.
1606 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1608 (-fdump-rtl-regmove): Ditto.
1609 * common.opt (foptimize-register-move, fregmove): Ignore.
1610 * Makefile.in (OBJS): Remove regmove.o.
1611 * regmove.c: Remove.
1612 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1614 (struct ira_allocno) New member allocno_prefs.
1615 (ALLOCNO_PREFS): New macro.
1616 (ira_prefs, ira_prefs_num): New external vars.
1617 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1618 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1619 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1621 (ira_add_allocno_copy_to_list): Remove prototype.
1622 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1623 (ira_pref_iterator): New type.
1624 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1625 (FOR_EACH_PREF): New macro.
1626 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1627 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1629 (ira_setup_alts): New function.
1630 (decrease_live_ranges_number): New function.
1631 (ira): Call the above function.
1632 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1633 (ira_create_allocno): Initialize allocno prefs.
1634 (pref_pool, pref_vec): New static vars.
1635 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1637 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1638 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1639 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1640 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1641 (ira_add_allocno_copy_to_list): Make static. Rename to
1642 add_allocno_copy_to_list.
1643 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1644 swap_allocno_copy_ends_if_necessary.
1645 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1646 ira_remove_allocno_prefs.
1647 (ira_flattening): Ditto.
1648 (ira_build): Call initiate_prefs, print_prefs.
1649 (ira_destroy): Call finish_prefs.
1650 * ira-color.c (struct update_cost_record): New.
1651 (struct allocno_color_data): Add new member update_cost_records.
1652 (update_cost_record_pool): New static var.
1653 (init_update_cost_records, get_update_cost_record): New functions.
1654 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1655 (struct update_cost_queue_elem): Add member from.
1656 (initiate_cost_update): Call init_update_cost_records.
1657 (finish_cost_update): Call finish_update_cost_records.
1658 (queue_update_cost, get_next_update_cost): Add new param from.
1659 (Update_allocno_cost, update_costs_from_allocno): New functions.
1660 (update_costs_from_prefs): Ditto.
1661 (update_copy_costs): Rename to update_costs_from_copies.
1662 (restore_costs_from_copies): New function.
1663 (update_conflict_hard_regno_costs): Don't go back.
1664 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1666 (pop_allocnos): Add priniting more debug info.
1667 (color_allocnos): Remove prefs for conflicting hard regs.
1668 Call update_costs_from_prefs.
1669 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1670 (get_dup_num): Rename, modify, and move to ira.c
1671 (process_regs_for_copy): Add prefs.
1672 (add_insn_allocno_copies): Put src as first arg of
1673 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1674 * ira-costs.c (record_reg_classes): Modify and move code into
1675 record_operands_costs.
1676 (find_costs_and_classes): Create prefs for the hard reg of small
1678 (process_bb_node_for_hard_reg_moves): Add prefs.
1680 2013-10-30 Richard Biener <rguenther@suse.de>
1683 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1684 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1686 (pre_and_rev_post_order_compute): ... which now wraps it.
1687 * graph.c (draw_cfg_nodes_no_loops): Use
1688 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1691 2013-10-30 Christian Bruel <christian.bruel@st.com>
1693 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1694 (sh_expand_cmpstr): Handle known align and schedule improvements.
1695 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1696 * config/sh/sh.md (cmpstrnsi): New pattern.
1698 2013-10-30 Martin Jambor <mjambor@suse.cz>
1700 PR rtl-optimization/10474
1701 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1703 (interesting_dest_for_shprep): New function.
1704 (split_live_ranges_for_shrink_wrap): Likewise.
1705 (ira): Calculate dominance info and df analysis. Call
1706 split_live_ranges_for_shrink_wrap.
1708 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1711 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1713 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1715 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1716 * tree.h (POINTER_BOUNDS_P): New.
1717 (BOUNDED_TYPE_P): New.
1719 (pointer_bounds_type_node): New.
1720 * tree.c (build_common_tree_nodes): Initialize
1721 pointer_bounds_type_node.
1722 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1723 (gimple_call_num_nobnd_args): New.
1724 (gimple_call_nobnd_arg): New.
1725 (gimple_return_retbnd): New.
1726 (gimple_return_set_retbnd): New
1727 * gimple.c (gimple_build_return): Increase number of ops
1728 for return statement.
1729 (gimple_call_get_nobnd_arg_index): New.
1730 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1732 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1734 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1735 with "chkp ctor" and "bnd_legacy" attributes.
1736 * gimplify.c (gimplify_init_constructor): Avoid infinite
1737 loop during gimplification of bounds initializer.
1739 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1741 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1742 (handle_bnd_legacy): New.
1743 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1744 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1747 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1749 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1750 (BT_FN_PTR_CONST_PTR): New.
1751 (BT_FN_CONST_PTR_CONST_PTR): New.
1752 (BT_FN_PTR_CONST_PTR_SIZE): New.
1753 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1754 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1755 (BT_FN_VOID_CONST_PTR_SIZE): New.
1756 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1757 * chkp-builtins.def: New.
1758 * builtins.def: include chkp-builtins.def.
1759 (DEF_CHKP_BUILTIN): New.
1760 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1761 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1762 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1763 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1764 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1765 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1766 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1767 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1768 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1769 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1770 * common.opt (fcheck-pointer-bounds): New.
1771 * toplev.c (process_options): Check Pointer Bounds Checker is
1773 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1775 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1777 * target.def (builtin_chkp_function): New.
1778 (chkp_bound_type): New.
1779 (chkp_bound_mode): New.
1780 (fn_abi_va_list_bounds_size): New.
1781 (load_bounds_for_arg): New.
1782 (store_bounds_for_arg): New.
1783 * targhooks.h (default_load_bounds_for_arg): New.
1784 (default_store_bounds_for_arg): New.
1785 (default_fn_abi_va_list_bounds_size): New.
1786 (default_chkp_bound_type): New.
1787 (default_chkp_bound_mode): New.
1788 (default_builtin_chkp_function): New.
1789 * targhooks.c (default_load_bounds_for_arg): New.
1790 (default_store_bounds_for_arg): New.
1791 (default_fn_abi_va_list_bounds_size): New.
1792 (default_chkp_bound_type): New.
1793 (default_chkp_bound_mode); New.
1794 (default_builtin_chkp_function): New.
1795 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1796 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1797 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1798 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1799 (TARGET_CHKP_BOUND_TYPE): New.
1800 (TARGET_CHKP_BOUND_MODE): New.
1801 * doc/tm.texi: Regenerated.
1802 * langhooks.h (lang_hooks): Add chkp_supported field.
1803 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1804 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1806 2013-10-29 Andrew Pinski <apinski@cavium.com>
1808 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1809 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1810 fails, combining the branches anyways.
1811 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1812 increases the number of combinings.
1813 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1815 2013-10-29 Mike Stump <mikestump@comcast.net>
1817 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1818 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1819 (make_vector_mode): Increase namebuf to 16.
1820 (emit_insn_modes_h): When processing BImode, don't
1821 also match partial int modes.
1822 (emit_class_narrowest_mode): Likewise.
1824 * config/bfin/bfin-modes.def: Add precision to PDI.
1825 * config/m32c/m32c-modes.def: Add precision to PSI.
1826 * config/msp430/msp430-modes.def: Add precision to PSI.
1827 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1828 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1830 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1833 * config/sh/sh.md (*addc): Rename existing variations to ...
1834 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1835 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1836 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1837 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1839 2013-10-29 Teresa Johnson <tejohnson@google.com>
1842 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1843 insanities when updating probabilities.
1845 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1847 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1848 prettyprinter for cgraph_node to reflect the conversion of the
1849 symtable types to a C++ class hierarchy: it now *is* a
1850 symtab_node_base, rather than having one (named "symbol").
1852 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1854 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1855 __cilkrts_pop_frame. If matched, then return true for built-in
1857 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1858 BUILT_IN_CILK_POP_FRAME case.
1859 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1860 (lhs_cilk_detect_spawn): Likewise.
1861 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1862 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1863 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1864 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1865 (LANG_HOOKS_CILKPLUS): Likewise.
1866 * tree.h (CILK_SPAWN_FN): Likewise.
1867 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1868 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1869 (OBJS): Added cilk-common.o.
1870 (BUILTINS_DEF): Added cilk-builtins.def.
1871 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1872 (lhd_cilk_detect_spawn): Likewise.
1873 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1874 (struct lang_hooks): Added new field called "cilkplus."
1875 * cilk-common.c: New file.
1877 * cilk-builtins.def: Likewise.
1878 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1879 "__cilk" macro and set it to 200.
1880 * function.h (struct function::cilk_frame_decl): New field.
1881 (struct function::is_cilk_function): Likewise.
1882 (struct function::calls_cilk_spawn): Likewise.
1883 * gimplify.c (gimplify_call_expr): Added a check if the function call
1884 being gimplified is a spawn detach point. If so, then add pop_frame
1885 and detach function calls.
1886 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1887 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1888 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1889 function. If so, added a _Cilk_sync and gimplified it.
1890 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1891 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1892 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1894 (can_inline_edge_p): Prevent inling of spawnee function.
1895 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1896 for functions that use Cilk keywords.
1897 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1898 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1899 CILK_SYNC_STMT cases.
1900 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1902 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1903 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1904 * passes.texi (Cilk Keywords): New section that describes the compiler
1905 code changes for handling Cilk Keywords.
1907 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1909 Patch autogenerated by refactor_symtab.py from
1910 https://github.com/davidmalcolm/gcc-refactoring-scripts
1911 revision 58bb219cc090b2f4516a9297d868c245495ee622
1913 * asan.c (asan_finish_file): Update for conversion of symtab types to
1914 a true class hierarchy.
1915 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1916 * cgraph.c (cgraph_get_body): Likewise.
1917 (cgraph_get_create_real_symbol_node): Likewise.
1918 (verify_cgraph_node): Likewise.
1919 (verify_edge_corresponds_to_fndecl): Likewise.
1920 (verify_edge_count_and_frequency): Likewise.
1921 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1922 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1923 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1924 (cgraph_node_cannot_return): Likewise.
1925 (cgraph_set_pure_flag_1): Likewise.
1926 (cgraph_set_const_flag_1): Likewise.
1927 (cgraph_set_nothrow_flag_1): Likewise.
1928 (cgraph_make_node_local_1): Likewise.
1929 (cgraph_for_node_and_aliases): Likewise.
1930 (cgraph_for_node_thunks_and_aliases): Likewise.
1931 (cgraph_node_can_be_local_p): Likewise.
1932 (cgraph_node_cannot_be_local_p_1): Likewise.
1933 (cgraph_function_body_availability): Likewise.
1934 (dump_cgraph_node): Likewise.
1935 (cgraph_rtl_info): Likewise.
1936 (cgraph_mark_address_taken_node): Likewise.
1937 (cgraph_remove_node): Likewise.
1938 (cgraph_release_function_body): Likewise.
1939 (cgraph_update_edges_for_call_stmt_node): Likewise.
1940 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1941 (cgraph_make_edge_direct): Likewise.
1942 (cgraph_resolve_speculation): Likewise.
1943 (cgraph_speculative_call_info): Likewise.
1944 (cgraph_turn_edge_to_speculative): Likewise.
1945 (cgraph_create_edge_1): Likewise.
1946 (cgraph_set_call_stmt): Likewise.
1947 (cgraph_node_for_asm): Likewise.
1948 (cgraph_add_thunk): Likewise.
1949 (cgraph_same_body_alias): Likewise.
1950 (cgraph_create_function_alias): Likewise.
1951 (cgraph_create_node): Likewise.
1952 (cgraph_create_empty_node): Likewise.
1953 (record_function_versions): Likewise.
1954 (used_from_object_file_p): Likewise.
1955 * cgraph.h (symtab_can_be_discarded): Likewise.
1956 (symtab_real_symbol_p): Likewise.
1957 (cgraph_mark_force_output_node): Likewise.
1958 (cgraph_edge_recursive_p): Likewise.
1959 (symtab_alias_target): Likewise.
1960 (varpool_all_refs_explicit_p): Likewise.
1961 (varpool_can_remove_if_no_refs): Likewise.
1962 (cgraph_only_called_directly_or_aliased_p): Likewise.
1963 (cgraph_next_function_with_gimple_body): Likewise.
1964 (cgraph_first_function_with_gimple_body): Likewise.
1965 (cgraph_function_with_gimple_body_p): Likewise.
1966 (cgraph_next_function): Likewise.
1967 (cgraph_first_function): Likewise.
1968 (cgraph_next_defined_function): Likewise.
1969 (cgraph_first_defined_function): Likewise.
1970 (varpool_next_defined_variable): Likewise.
1971 (varpool_first_defined_variable): Likewise.
1972 (varpool_next_static_initializer): Likewise.
1973 (varpool_first_static_initializer): Likewise.
1974 (varpool_next_variable): Likewise.
1975 (varpool_first_variable): Likewise.
1976 (varpool_node_name): Likewise.
1977 (varpool): Likewise.
1979 (is_a_helper <varpool_node>::test): Likewise.
1980 (is_a_helper <cgraph_node>::test): Likewise.
1981 (varpool_variable_node): Likewise.
1982 (cgraph_function_or_thunk_node): Likewise.
1983 (varpool_alias_target): Likewise.
1984 (cgraph_alias_target): Likewise.
1985 (cgraph_node_name): Likewise.
1986 (varpool_node_asm_name): Likewise.
1987 (cgraph_node_asm_name): Likewise.
1988 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1989 (cgraph_rebuild_references): Likewise.
1990 (rebuild_cgraph_edges): Likewise.
1991 (record_eh_tables): Likewise.
1992 (build_cgraph_edges): Likewise.
1993 (mark_store): Likewise.
1994 (mark_load): Likewise.
1995 (mark_address): Likewise.
1996 (record_type_list): Likewise.
1997 (record_reference): Likewise.
1998 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1999 (cgraph_materialize_clone): Likewise.
2000 (cgraph_function_versioning): Likewise.
2001 (cgraph_copy_node_for_versioning): Likewise.
2002 (update_call_expr): Likewise.
2003 (cgraph_find_replacement_node): Likewise.
2004 (cgraph_create_virtual_clone): Likewise.
2005 (cgraph_clone_node): Likewise.
2006 * cgraphunit.c (compile): Likewise.
2007 (output_weakrefs): Likewise.
2008 (output_in_order): Likewise.
2009 (expand_function): Likewise.
2010 (assemble_thunks_and_aliases): Likewise.
2011 (expand_thunk): Likewise.
2012 (mark_functions_to_output): Likewise.
2013 (handle_alias_pairs): Likewise.
2014 (analyze_functions): Likewise.
2015 (walk_polymorphic_call_targets): Likewise.
2016 (varpool_finalize_decl): Likewise.
2017 (process_function_and_variable_attributes): Likewise.
2018 (cgraph_process_same_body_aliases): Likewise.
2019 (analyze_function): Likewise.
2020 (cgraph_add_new_function): Likewise.
2021 (cgraph_finalize_function): Likewise.
2022 (referred_to_p): Likewise.
2023 (cgraph_reset_node): Likewise.
2024 (cgraph_process_new_functions): Likewise.
2025 (enqueue_node): Likewise.
2026 (decide_is_symbol_needed): Likewise.
2027 * coverage.c (coverage_compute_profile_id): Likewise.
2028 * dbxout.c (dbxout_expand_expr): Likewise.
2029 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
2030 (reference_to_unused): Likewise.
2031 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2032 * gimplify.c (unvisit_body): Likewise.
2033 (unshare_body): Likewise.
2034 * ipa-cp.c (ipcp_generate_summary): Likewise.
2035 (ipcp_decision_stage): Likewise.
2036 (identify_dead_nodes): Likewise.
2037 (decide_whether_version_node): Likewise.
2038 (decide_about_value): Likewise.
2039 (perhaps_add_new_callers): Likewise.
2040 (create_specialized_node): Likewise.
2041 (update_profiling_info): Likewise.
2042 (ipcp_propagate_stage): Likewise.
2043 (estimate_local_effects): Likewise.
2044 (good_cloning_opportunity_p): Likewise.
2045 (devirtualization_time_bonus): Likewise.
2046 (propagate_constants_accross_call): Likewise.
2047 (initialize_node_lattices): Likewise.
2048 (ipcp_cloning_candidate_p): Likewise.
2049 (determine_versionability): Likewise.
2050 (print_all_lattices): Likewise.
2051 (print_lattice): Likewise.
2052 (ipcp_discover_new_direct_edges): Likewise.
2053 * ipa-devirt.c (ipa_devirt): Likewise.
2054 (likely_target_p): Likewise.
2055 (update_type_inheritance_graph): Likewise.
2056 (possible_polymorphic_call_target_p): Likewise.
2057 (dump_possible_polymorphic_call_targets): Likewise.
2058 (devirt_variable_node_removal_hook): Likewise.
2059 (record_binfo): Likewise.
2060 (maybe_record_node): Likewise.
2061 (build_type_inheritance_graph): Likewise.
2062 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2063 (inline_generate_summary): Likewise.
2064 (inline_analyze_function): Likewise.
2065 (do_estimate_growth): Likewise.
2066 (simple_edge_hints): Likewise.
2067 (estimate_node_size_and_time): Likewise.
2068 (estimate_edge_devirt_benefit): Likewise.
2069 (compute_inline_parameters): Likewise.
2070 (estimate_function_body_sizes): Likewise.
2071 (compute_bb_predicates): Likewise.
2072 (initialize_inline_failed): Likewise.
2073 (dump_inline_summary): Likewise.
2074 (dump_inline_edge_summary): Likewise.
2075 * ipa-inline-transform.c (inline_transform): Likewise.
2076 (preserve_function_body_p): Likewise.
2077 (save_inline_function_body): Likewise.
2078 (inline_call): Likewise.
2079 (clone_inlined_nodes): Likewise.
2080 (can_remove_node_now_p): Likewise.
2081 (can_remove_node_now_p_1): Likewise.
2082 * ipa-inline.c (early_inliner): Likewise.
2083 (early_inline_small_functions): Likewise.
2084 (inline_always_inline_functions): Likewise.
2085 (ipa_inline): Likewise.
2086 (flatten_function): Likewise.
2087 (inline_small_functions): Likewise.
2088 (speculation_useful_p): Likewise.
2089 (recursive_inlining): Likewise.
2090 (update_caller_keys): Likewise.
2091 (reset_edge_caches): Likewise.
2092 (update_edge_key): Likewise.
2093 (edge_badness): Likewise.
2094 (relative_time_benefit): Likewise.
2095 (want_inline_self_recursive_call_p): Likewise.
2096 (want_inline_small_function_p): Likewise.
2097 (want_early_inline_function_p): Likewise.
2098 (num_calls): Likewise.
2099 (can_early_inline_edge_p): Likewise.
2100 (can_inline_edge_p): Likewise.
2101 (report_inline_failed_reason): Likewise.
2102 * ipa-profile.c (ipa_profile): Likewise.
2103 (ipa_propagate_frequency): Likewise.
2104 (ipa_propagate_frequency_1): Likewise.
2105 (ipa_profile_generate_summary): Likewise.
2106 * ipa-prop.c (ipcp_transform_function): Likewise.
2107 (read_replacements_section): Likewise.
2108 (ipa_prop_read_section): Likewise.
2109 (ipa_modify_call_arguments): Likewise.
2110 (ipa_print_node_params): Likewise.
2111 (propagate_controlled_uses): Likewise.
2112 (update_indirect_edges_after_inlining): Likewise.
2113 (remove_described_reference): Likewise.
2114 (ipa_make_edge_direct_to_target): Likewise.
2115 (ipa_analyze_node): Likewise.
2116 (ipa_analyze_params_uses): Likewise.
2117 (ipa_compute_jump_functions): Likewise.
2118 (ipa_get_callee_param_type): Likewise.
2119 (ipa_print_node_jump_functions): Likewise.
2120 (ipa_initialize_node_params): Likewise.
2121 (ipa_populate_param_decls): Likewise.
2122 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
2123 (write_agg_replacement_chain): Likewise.
2124 (ipa_write_node_info): Likewise.
2125 (ipa_edge_duplication_hook): Likewise.
2126 (try_decrement_rdesc_refcount): Likewise.
2127 * ipa-pure-const.c (propagate_nothrow): Likewise.
2128 (propagate_pure_const): Likewise.
2129 (pure_const_read_summary): Likewise.
2130 (pure_const_write_summary): Likewise.
2131 (analyze_function): Likewise.
2132 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
2133 (ipa_ref_referring_ref_list): Likewise.
2134 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
2135 (ipa_remove_stmt_references): Likewise.
2136 (ipa_find_reference): Likewise.
2137 (ipa_dump_referring): Likewise.
2138 (ipa_dump_references): Likewise.
2139 (ipa_record_reference): Likewise.
2140 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
2141 (ipa_reference_write_optimization_summary): Likewise.
2142 (write_node_summary_p): Likewise.
2143 (propagate): Likewise.
2144 (read_write_all_from_decl): Likewise.
2145 (generate_summary): Likewise.
2146 (analyze_function): Likewise.
2147 (propagate_bits): Likewise.
2148 (ipa_reference_get_not_written_global): Likewise.
2149 (ipa_reference_get_not_read_global): Likewise.
2150 * ipa-split.c (execute_split_functions): Likewise.
2151 (split_function): Likewise.
2152 * ipa-utils.c (ipa_merge_profiles): Likewise.
2153 (dump_cgraph_node_set): Likewise.
2154 (ipa_reverse_postorder): Likewise.
2155 (ipa_edge_within_scc): Likewise.
2156 (ipa_get_nodes_in_cycle): Likewise.
2157 (ipa_free_postorder_info): Likewise.
2158 (ipa_reduced_postorder): Likewise.
2159 (searchc): Likewise.
2160 (recursive_call_p): Likewise.
2161 * ipa.c (ipa_cdtor_merge): Likewise.
2162 (record_cdtor_fn): Likewise.
2163 (function_and_variable_visibility): Likewise.
2164 (varpool_externally_visible_p): Likewise.
2165 (cgraph_externally_visible_p): Likewise.
2166 (comdat_can_be_unshared_p): Likewise.
2167 (comdat_can_be_unshared_p_1): Likewise.
2168 (address_taken_from_non_vtable_p): Likewise.
2169 (ipa_discover_readonly_nonaddressable_vars): Likewise.
2170 (symtab_remove_unreachable_nodes): Likewise.
2171 (walk_polymorphic_call_targets): Likewise.
2172 (process_references): Likewise.
2173 (enqueue_node): Likewise.
2174 (has_addr_references_p): Likewise.
2175 (cgraph_non_local_node_p_1): Likewise.
2176 * is-a.h (varpool_analyze_node): Likewise.
2177 * lto-cgraph.c (input_symtab): Likewise.
2178 (merge_profile_summaries): Likewise.
2179 (input_cgraph_1): Likewise.
2180 (input_edge): Likewise.
2181 (input_varpool_node): Likewise.
2182 (input_node): Likewise.
2183 (input_overwrite_node): Likewise.
2184 (compute_ltrans_boundary): Likewise.
2185 (output_refs): Likewise.
2186 (lto_output_varpool_node): Likewise.
2187 (lto_output_node): Likewise.
2188 (reachable_from_other_partition_p): Likewise.
2189 (referenced_from_other_partition_p): Likewise.
2190 (lto_output_edge): Likewise.
2191 (output_node_opt_summary): Likewise.
2192 (add_node_to): Likewise.
2193 (reachable_from_this_partition_p): Likewise.
2194 (lto_set_symtab_encoder_in_partition): Likewise.
2195 (lto_symtab_encoder_in_partition_p): Likewise.
2196 (lto_set_symtab_encoder_encode_initializer): Likewise.
2197 (lto_symtab_encoder_encode_initializer_p): Likewise.
2198 (lto_set_symtab_encoder_encode_body): Likewise.
2199 (lto_symtab_encoder_encode_body_p): Likewise.
2200 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2202 * lto-streamer-in.c (lto_read_body): Likewise.
2203 (fixup_call_stmt_edges): Likewise.
2204 (fixup_call_stmt_edges_1): Likewise.
2205 * lto-streamer-out.c (produce_symtab): Likewise.
2206 (output_symbol_p): Likewise.
2207 (write_symbol): Likewise.
2208 (lto_output): Likewise.
2209 (copy_function): Likewise.
2210 (output_function): Likewise.
2211 * passes.c (function_called_by_processed_nodes_p): Likewise.
2212 (ipa_write_optimization_summaries): Likewise.
2213 (ipa_write_summaries): Likewise.
2214 (do_per_function_toporder): Likewise.
2215 (do_per_function): Likewise.
2216 (dump_passes): Likewise.
2217 * symtab.c (symtab_semantically_equivalent_p): Likewise.
2218 (symtab_nonoverwritable_alias): Likewise.
2219 (symtab_nonoverwritable_alias_1): Likewise.
2220 (symtab_for_node_and_aliases): Likewise.
2221 (symtab_resolve_alias): Likewise.
2222 (fixup_same_cpp_alias_visibility): Likewise.
2223 (symtab_alias_ultimate_target): Likewise.
2224 (symtab_used_from_object_file_p): Likewise.
2225 (verify_symtab_base): Likewise.
2226 (dump_symtab_base): Likewise.
2227 (symtab_node_name): Likewise.
2228 (symtab_node_asm_name): Likewise.
2229 (symtab_dissolve_same_comdat_group_list): Likewise.
2230 (symtab_add_to_same_comdat_group): Likewise.
2231 (symtab_unregister_node): Likewise.
2232 (symtab_insert_node_to_hashtable): Likewise.
2233 (symtab_register_node): Likewise.
2234 (unlink_from_assembler_name_hash): Likewise.
2235 (insert_to_assembler_name_hash): Likewise.
2236 (eq_assembler_name): Likewise.
2237 (hash_node_by_assembler_name): Likewise.
2238 (eq_node): Likewise.
2239 (hash_node): Likewise.
2240 * toplev.c (wrapup_global_declaration_2): Likewise.
2241 * trans-mem.c (ipa_tm_execute): Likewise.
2242 (ipa_tm_transform_clone): Likewise.
2243 (ipa_tm_transform_transaction): Likewise.
2244 (ipa_tm_transform_calls_redirect): Likewise.
2245 (ipa_tm_insert_gettmclone_call): Likewise.
2246 (ipa_tm_insert_irr_call): Likewise.
2247 (ipa_tm_create_version): Likewise.
2248 (ipa_tm_create_version_alias): Likewise.
2249 (ipa_tm_mark_forced_by_abi_node): Likewise.
2250 (ipa_tm_mark_force_output_node): Likewise.
2251 (ipa_tm_diagnose_tm_safe): Likewise.
2252 (ipa_tm_mayenterirr_function): Likewise.
2253 (ipa_tm_scan_irr_function): Likewise.
2254 (ipa_tm_note_irrevocable): Likewise.
2255 (ipa_tm_scan_calls_clone): Likewise.
2256 (get_cg_data): Likewise.
2257 * tree-eh.c (tree_could_trap_p): Likewise.
2258 * tree-emutls.c (ipa_lower_emutls): Likewise.
2259 (create_emultls_var): Likewise.
2260 (lower_emutls_function_body): Likewise.
2261 (gen_emutls_addr): Likewise.
2262 (emutls_decl): Likewise.
2263 (new_emutls_decl): Likewise.
2264 * tree-inline.c (tree_function_versioning): Likewise.
2265 (optimize_inline_calls): Likewise.
2266 (expand_call_inline): Likewise.
2267 (estimate_num_insns): Likewise.
2268 (copy_bb): Likewise.
2269 (delete_unreachable_blocks_update_callgraph): Likewise.
2270 * tree-nested.c (gimplify_all_functions): Likewise.
2271 (create_nesting_tree): Likewise.
2272 (check_for_nested_with_variably_modified): Likewise.
2273 * tree-pretty-print.c (dump_function_header): Likewise.
2274 * tree-profile.c (tree_profiling): Likewise.
2275 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
2276 (modify_function): Likewise.
2277 (convert_callers): Likewise.
2278 (convert_callers_for_node): Likewise.
2279 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
2280 (associate_varinfo_to_alias): Likewise.
2281 (create_variable_info_for): Likewise.
2282 (get_constraint_for_ssa_var): Likewise.
2283 * tree-vectorizer.c (increase_alignment): Likewise.
2284 * tree.c (find_decls_types_in_var): Likewise.
2285 (find_decls_types_in_node): Likewise.
2286 (free_lang_data_in_decl): Likewise.
2287 * value-prof.c (gimple_ic_transform): Likewise.
2288 (gimple_ic): Likewise.
2289 (check_ic_target): Likewise.
2290 (init_node_map): Likewise.
2291 * varasm.c (decl_binds_to_current_def_p): Likewise.
2292 (default_binds_local_p_1): Likewise.
2293 (dump_tm_clone_pairs): Likewise.
2294 (assemble_alias): Likewise.
2295 (find_decl): Likewise.
2296 (mark_decl_referenced): Likewise.
2297 * varpool.c (varpool_for_node_and_aliases): Likewise.
2298 (varpool_extra_name_alias): Likewise.
2299 (varpool_create_variable_alias): Likewise.
2300 (add_new_static_var): Likewise.
2301 (varpool_finalize_named_section_flags): Likewise.
2302 (varpool_remove_unreferenced_decls): Likewise.
2303 (enqueue_node): Likewise.
2304 (varpool_assemble_decl): Likewise.
2305 (assemble_aliases): Likewise.
2306 (varpool_analyze_node): Likewise.
2307 (cgraph_variable_initializer_availability): Likewise.
2308 (varpool_add_new_variable): Likewise.
2309 (ctor_for_folding): Likewise.
2310 (dump_varpool_node): Likewise.
2311 (varpool_remove_initializer): Likewise.
2312 (varpool_remove_node): Likewise.
2313 (varpool_node_for_decl): Likewise.
2314 (varpool_create_empty_node): Likewise.
2315 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2316 (ix86_get_function_versions_dispatcher): Likewise.
2318 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2320 * cgraph.h (symtab_node_base): Convert to a class;
2321 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2322 chain_next/prev from symtab_node_def.
2323 (cgraph_node): Inherit from symtab_node; add GTY option
2324 tag ("SYMTAB_FUNCTION").
2325 (varpool_node): Inherit from symtab_node; add GTY option
2326 tag ("SYMTAB_VARIABLE").
2327 (symtab_node_def): Remove.
2328 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2329 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2330 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2331 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2333 * ipa-ref.h (symtab_node_def): Drop.
2334 (symtab_node): Change underlying type from symtab_node_def to
2336 (const_symtab_node): Likwise.
2338 * is-a.h: Update examples in comment.
2340 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2341 (assembler_name_hash): Likewise.
2343 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2345 * doc/tree-ssa.texi (gimple_phi_result): Document.
2346 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2347 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2348 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2349 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2351 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2353 * expr.h: Revert change and include tree-core.h.
2354 * rtl.h: Revert change and don't include tree-core.h.
2356 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2358 * config/darwin.c: Include gimple.h.
2359 * config/i386/winnt.c: Likewise.
2361 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2363 PR tree-optimization/19831
2364 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2366 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2368 * tree-outof-ssa.h: Remove include files.
2369 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2371 * tree-ssa-coalesce.c: Likewise.
2372 * cfgexpand.c: Likewise.
2373 * tree-ssa-ter.c: Likewise.
2374 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2375 * lto-streamer.h: Likewise.
2376 * cgraphbuild.c: Add gimple.h to include list.
2377 * data-streamer-in.c: Likewise.
2378 * ipa-cp.c: Likewise.
2379 * tree-streamer.c: Likewise.
2380 * lto-compress.c: Likewise.
2381 * ipa-reference.c: Likewise.
2382 * data-streamer-out.c: Likewise.
2383 * lto-cgraph.c: Likewise.
2384 * cgraphclones.c: Likewise.
2385 * ipa-utils.c: Likewise.
2386 * data-streamer.c: Likewise.
2387 * ipa-split.c: Likewise.
2388 * lto-section-in.c: Likewise.
2389 * tree-streamer-out.c: Likewise.
2390 * ipa-prop.c: Likewise.
2391 * tree-streamer-in.c: Likewise.
2392 * symtab.c: Likewise.
2393 * opts-global.c: Likewise.
2394 * lto-opts.c: Likewise.
2395 * lto-section-out.c: Likewise.
2396 * lto-streamer.c: Likewise.
2397 * rtl.h: Add tree-core.h to include list.
2398 * expr.h: Remove tree-core.h from include list.
2399 * gimple.h: Likewise.
2400 * ipa-utils.h: Likewise.
2401 * streamer-hooks.h: Likewise.
2402 * streamer-hooks.c: Include input.h.
2404 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2406 * config/arm/arm.c (cortexa7_extra_costs): New table.
2407 (arm_cortex_a7_tune): New.
2408 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2410 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2412 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2414 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2416 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2417 to use autogenerated markers for a class-hierarchy, every class
2418 must have a GTY marker.
2419 * gengtype.h (struct type): Add linked list of subclasses to
2420 the "s" member of the union.
2421 (add_subclass): New decl.
2422 * gengtype-state.c (read_state_struct_type): Set up subclass
2424 * gengtype.c (get_ultimate_base_class): New.
2425 (add_subclass): New.
2426 (new_structure): Set up subclass linked list.
2427 (set_gc_used_type): Propagate usage information to subclasses.
2428 (output_mangled_typename): Use get_ultimate_base_class.
2429 (walk_subclasses): Use the subclass linked list, avoiding an
2430 O(N^2) when writing out all types.
2431 (walk_type): Issue an error if the base class is missing a tag,
2432 rather than generating bogus C code. Add a gcc_unreachable
2433 default case, in case people omit tags from concrete subclasses,
2434 or get the values wrong.
2435 (write_func_for_structure): Issue an error for subclasses for
2436 which the base doesn't have a "desc", since otherwise the
2437 autogenerated routines for the base would silently fail to visit
2438 any subclass fields.
2439 (write_root): Use get_ultimate_base_class, tweaking constness of
2440 tp to match that function's signature.
2442 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2444 * doc/gty.texi (GTY Options): Add note about inheritance to
2445 description of desc and tag.
2446 (Inheritance and GTY): New.
2448 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2450 * gengtype-parse.c (opts_have): Drop "static" so that
2451 we can use this from gengtype.c.
2452 * gengtype.c (set_gc_used_type): Mark any base class as used;
2453 update field traversal to visit inherited fields.
2454 (output_mangled_typename): Convert references to classes within
2455 an inheritance hierarchy to reference the ultimate base class,
2456 since only it will have gt_ functions.
2457 (get_string_option): New.
2458 (walk_subclasses): New.
2459 (walk_type): Treat GTY structs that have a "desc" as being the
2460 root of an inheritance hierarchy. Generate a switch on it
2461 within the marking function which walks all subclasses, adding
2462 cases for them via walk_subclasses. For subclasses, visit all
2463 fields of the type (including inherited ones).
2464 (write_func_for_structure): Don't write fns for subclasses, only
2465 for the ultimate base class within an inheritance hierarchy.
2466 Subclasses-marking will be handled by the base class marking functions.
2467 (write_types): Likewise.
2468 (write_local_func_for_structure): Likewise.
2469 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2470 a "tag" option (and are thus concrete subclasses).
2471 (write_root): Use the marker function for the ultimate base class.
2472 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2473 (opts_have): Add declaration.
2475 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2477 * lra-spills.c (lra_final_code_change): Remove useless move insns
2478 originated from moves of pseudos.
2480 2013-10-28 Jeff Law <law@redhat.com>
2482 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2484 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2486 * df-scan.c (df_collection_rec): Adjust.
2487 (copy_defs): New constant.
2488 (copy_uses): Likewise.
2489 (copy_eq_uses): Likewise.
2490 (copy_mw): Likewise.
2491 (copy_all): Likewise.
2492 (df_insn_rescan): Adjust.
2493 (df_notes_rescan): Likewise.
2494 (df_swap_refs): Likewise.
2495 (df_sort_and_compress_refs): Likewise.
2496 (df_sort_and_compress_mws): Likewise.
2497 (df_install_refs): Likewise.
2498 (df_install_mws): Likewise.
2499 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2501 (df_bb_refs_record): Adjust.
2502 (df_record_entry_block_defs): Likewise.
2503 (df_record_exit_block_defs): Likewise.
2504 (df_refs_verify): Likewise.
2505 (df_mws_verify): Likewise.
2506 (df_insn_refs_verify): Likewise.
2507 (df_bb_verify): Likewise.
2508 * ipa-pure-const.c (finish_state): Remove.
2509 (propagate): Adjust.
2510 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2511 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2512 var-tracking.c: Adjust.
2513 * vec.c (stack_vecs): Remove.
2514 (register_stack_vec): Likewise.
2515 (stack_vec_register_index): Likewise.
2516 (unregister_stack_vec): Likewise.
2517 * vec.h (struct va_stack): Remove.
2518 (struct vec<T, A, vl_ptr>): Specialize as
2519 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2520 allocation strategy compatable with the vl_ptr layout.
2521 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2522 specialization anyway.
2523 (class stack_vec): New class.
2524 (vec_stack_alloc): Remove.
2525 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2527 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2528 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2529 Sergey Lega <sergey.s.lega@intel.com>
2530 Anna Tikhonova <anna.tikhonova@intel.com>
2531 Ilya Tocar <ilya.tocar@intel.com>
2532 Andrey Turetskiy <andrey.turetskiy@intel.com>
2533 Ilya Verbin <ilya.verbin@intel.com>
2534 Kirill Yukhin <kirill.yukhin@intel.com>
2535 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2537 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2538 (*prefetch_avx512pf_<mode>): New.
2539 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2540 (avx512f_maskcmp<mode>3): Ditto.
2541 (vashrv16si3): Ditto.
2543 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2544 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2545 Sergey Lega <sergey.s.lega@intel.com>
2546 Anna Tikhonova <anna.tikhonova@intel.com>
2547 Ilya Tocar <ilya.tocar@intel.com>
2548 Andrey Turetskiy <andrey.turetskiy@intel.com>
2549 Ilya Verbin <ilya.verbin@intel.com>
2550 Kirill Yukhin <kirill.yukhin@intel.com>
2551 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2553 * config/i386/i386.md (any_truncate): New.
2554 (trunsuffix): Ditto.
2555 * config/i386/predicates.md (const_8_to_9_operand): New.
2556 (const_10_to_11_operand): Ditto.
2557 (const_12_to_13_operand): Ditto.
2558 (const_14_to_15_operand): Ditto.
2559 (const_16_to_19_operand): Ditto.
2560 (const_20_to_23_operand): Ditto.
2561 (const_24_to_27_operand): Ditto.
2562 (const_28_to_31_operand): Ditto.
2563 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2564 (cvtusi2<ssescalarmodesuffix>32): New.
2565 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2566 (ufloatv16siv16sf2): Ditto.
2567 (avx512f_fix_notruncv16sfv16si): Ditto.
2568 (avx512f_ufix_notruncv16sfv16si): Ditto.
2569 (avx512f_vcvtss2usi): Ditto.
2570 (avx512f_vcvtss2usiq): Ditto.
2571 (avx512f_vcvttss2usi): Ditto.
2572 (avx512f_vcvttss2usiq): Ditto.
2573 (avx512f_vcvtsd2usi): Ditto.
2574 (avx512f_vcvtsd2usiq): Ditto.
2575 (avx512f_vcvttsd2usi): Ditto.
2576 (avx512f_vcvttsd2usiq): Ditto.
2577 (ufloatv8siv8df): Ditto.
2578 (avx512f_cvtdq2pd512_2): Ditto.
2579 (avx512f_cvtpd2dq512): Ditto.
2580 (avx512f_ufix_notruncv8dfv8si): Ditto.
2581 (avx512f_cvtpd2ps512): Ditto.
2582 (vec_unpacks_lo_v16sf): Ditto.
2583 (vec_unpacks_hi_v16sf): Ditto.
2584 (vec_unpacks_float_hi_v16si): Ditto.
2585 (vec_unpacks_float_lo_v16si): Ditto.
2586 (avx512f_unpckhps512): Ditto.
2587 (avx512f_unpcklps512): Ditto.
2588 (avx512f_movshdup512): Ditto.
2589 (avx512f_movsldup512): Ditto.
2590 (vec_extract_lo_v32hi): Ditto.
2591 (vec_extract_hi_v32hi): Ditto.
2592 (vec_extract_lo_v64qi): Ditto.
2593 (vec_extract_hi_v64qi): Ditto.
2594 (avx512f_unpckhpd512): Ditto.
2595 (avx512f_movddup512): Ditto.
2596 (avx512f_unpcklpd512): Ditto.
2597 (*avx512f_unpcklpd512): Ditto.
2598 (avx512f_shufps512_1): Ditto.
2599 (avx512f_shufpd512_1): Ditto.
2600 (avx512f_interleave_highv8di): Ditto.
2601 (avx512f_interleave_lowv8di): Ditto.
2602 (PMOV_DST_MODE): Ditto.
2603 (pmov_src_mode): Ditto.
2604 (pmov_src_lower): Ditto.
2606 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2607 (*avx512f_<code>v8div16qi2): Ditto.
2608 (*avx512f_<code>v8div16qi2_store): Ditto.
2609 (vec_widen_umult_even_v16si): Ditto.
2610 (*vec_widen_umult_even_v16si): Ditto.
2611 (vec_widen_smult_even_v16si): Ditto.
2612 (*vec_widen_smult_even_v16si): Ditto.
2613 (avx512f_interleave_highv16si): Ditto.
2614 (avx512f_interleave_lowv16si): Ditto.
2615 (avx512f_<code>v16qiv16si2): Ditto.
2616 (avx512f_<code>v16hiv16si2): Ditto.
2617 (avx512f_<code>v8qiv8di2): Ditto.
2618 (avx512f_<code>v8hiv8di2): Ditto.
2619 (avx512f_<code>v8siv8di2): Ditto.
2620 (avx512cd_maskb_vec_dupv8di): Ditto.
2621 (avx512cd_maskw_vec_dupv16si): Ditto.
2622 (avx512f_vcvtph2ps512): Ditto.
2623 (avx512f_vcvtps2ph512): Ditto.
2624 (VEC_EXTRACT_MODE): Extened with wider modes.
2625 (VEC_PERM_AVX2): Ditto.
2626 (VEC_PERM_CONST): Ditto.
2628 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2630 * config/arc/arc.c (arc_ccfsm_post_advance):
2631 Add comment about TYPE_RETURN.
2633 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2635 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2636 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2637 (strip_offset): Convert offset to unsigned number.
2639 2013-10-27 Tom de Vries <tom@codesourcery.com>
2641 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2642 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2644 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2646 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2647 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2648 (TARGET_BIG_ENDIAN): New macro.
2649 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2650 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2651 * config/sh/sh.c: Likewise.
2652 * config/sh/sh.h: Likewise.
2654 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2656 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2657 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2659 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2661 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2663 (cmpstr_t, cmpstrsi): Fix formatting.
2665 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2668 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2669 addressing, do not use general_operand for memory operands.
2671 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2674 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2675 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2677 2013-10-26 Jeff Law <law@redhat.com>
2679 * predict.c (PRED_MUDFLAP): Remove.
2680 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2682 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2683 (OBJS): Remove tree-nomudflap.o
2684 (GTFILES): Remove tree-mudflap.c
2685 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2686 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2687 (mfwrap_spec, mflib_spec): Likewise.
2688 (cpp_unique_options, cc1_options, static_specs): Likewise.
2689 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2690 * passes.def: Likewise.
2691 * toplev.c (compile_file, process_options): Likewise.
2692 * tree-inline.c (copy_tree_r): Likewise.
2693 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2694 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2695 (build_constant_desc, output_constant_def_contents): Likewise.
2696 (categorize_decl_for_section): Likewise.
2697 * tree-mudflap.c: Removed.
2698 * tree-mudflap.h: Removed.
2699 * tree-nomudflap.c: Removed.
2700 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2701 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2702 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2703 * config/sol2.h (MFLIB_SPEC): Likewise.
2704 * doc/install.texi: Remove mudflap references.
2705 * doc/passes.texi: Similarly.
2706 * doc/sourcebuild.texi: Similarly.
2707 * doc/invoke.texi: Remove mudlfap related options.
2709 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2711 PR rtl-optimization/58759
2712 * lra-constraints.c (lra_constraints): Remove wrong condition to
2713 remove insn setting up an equivalent pseudo.
2715 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2717 * config/rs6000/rs6000-protos.h
2718 (rs6000_secondary_memory_needed_mode): New prototype.
2719 * config/rs6000/rs6000.c: Include ira.h.
2720 (TARGET_LRA_P): Redefine.
2721 (rs6000_legitimate_offset_address_p): Call
2722 legitimate_constant_pool_address_p in strict mode for LRA.
2723 (rs6000_legitimate_address_p): Ditto.
2724 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2725 (rs6000_emit_move): Add LRA version of code to generate load/store
2727 (rs6000_secondary_memory_needed_mode): New.
2728 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2729 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2730 constants, memory, and FP registers.
2731 (rs6000_lra_p): New.
2732 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2733 * config/rs6000/rs6000.opt (mlra): New option.
2734 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2736 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2738 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2739 has_single_use () and not do the conversion if has_single_use ()
2740 returns false for the multiplication result.
2742 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2744 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2745 to "NODE", since this works on a "tree", not an
2747 (CONSTANT_CLASS_P): Likewise.
2750 (INDIRECT_REF_P): Likewise.
2751 (REFERENCE_CLASS_P): Likewise.
2752 (COMPARISON_CLASS_P): Likewise.
2753 (UNARY_CLASS_P): Likewise.
2754 (BINARY_CLASS_P): Likewise.
2755 (STATEMENT_CLASS_P): Likewise.
2756 (VL_EXP_CLASS_P): Likewise.
2757 (EXPRESSION_CLASS_P): Likewise.
2758 (IS_TYPE_OR_DECL_P): Likewise.
2760 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2762 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2763 ADDR_EXPR in the defining statement.
2765 2013-10-25 Richard Biener <rguenther@suse.de>
2767 PR tree-optimization/58626
2768 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2769 anti_dd, output_dd and input_dd.
2770 (struct rdg_edge): Remove level and relation members.
2771 (RDGE_LEVEL, RDGE_RELATION): Remove.
2772 (dot_rdg_1): Adjust.
2773 (create_rdg_edge_for_ddr): Remove.
2774 (create_rdg_edges_for_scalar): Adjust.
2775 (create_edge_for_control_dependence): Likewise.
2776 (create_rdg_edges): Split into ...
2777 (create_rdg_flow_edges): ... this
2778 (create_rdg_cd_edges): ... and this.
2780 (build_rdg): Likewise, do not compute data dependences or
2782 (pg_add_dependence_edges): New function.
2784 (distribute_loop): First apply all non-dependence based
2785 partition mergings. Then compute dependences between partitions
2786 and merge and order partitions according to them.
2788 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2790 PR rtl-optimization/58831
2791 * alias.c (init_alias_analysis): At the beginning of each iteration,
2792 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2794 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2796 * recog.c (search_ofs): New static variable moved from...
2797 (peep2_find_free_register): ...here.
2798 (peephole2_optimize): Initialize it.
2800 2013-10-25 Tobias Burnus <burnus@net-b.de>
2802 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2804 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2806 * config/i386/i386.h (TARGET_MPX): New define.
2807 (TARGET_MPX_P): Ditto.
2809 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2811 * config/i386/constraints.md (B): New.
2814 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2815 * config/i386/i386-modes.def (BND32): New.
2817 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2818 * config/i386/i386.c (isa_opts): Add mmpx.
2819 (regclass_map): Add bound registers.
2820 (dbx_register_map): Likewise.
2821 (dbx64_register_map): Likewise.
2822 (svr4_dbx_register_map): Likewise.
2824 (ix86_option_override_internal): Support MPX ISA.
2825 (ix86_conditional_register_usage): Support bound registers.
2826 (print_reg): Likewise.
2827 (ix86_code_end): Add MPX bnd prefix.
2828 (output_set_got): Likewise.
2829 (ix86_output_call_insn): Likewise.
2830 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2831 (ix86_print_operand_punct_valid_p): Likewise.
2832 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2834 (ix86_class_likely_spilled_p): Add bound regs support.
2835 (ix86_hard_regno_mode_ok): Likewise.
2836 (x86_order_regs_for_local_alloc): Likewise.
2837 (ix86_bnd_prefixed_insn_p): New.
2838 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2839 (FIXED_REGISTERS): Add bound registers.
2840 (CALL_USED_REGISTERS): Likewise.
2841 (REG_ALLOC_ORDER): Likewise.
2842 (HARD_REGNO_NREGS): Likewise.
2844 (VALID_BND_REG_MODE): New.
2845 (FIRST_BND_REG): New.
2846 (LAST_BND_REG): New.
2847 (reg_class): Add BND_REGS.
2848 (REG_CLASS_NAMES): Likewise.
2849 (REG_CLASS_CONTENTS): Likewise.
2851 (ANY_BND_REG_P): New.
2853 (HI_REGISTER_NAMES): Add bound registers.
2854 * config/i386/i386.md (UNSPEC_BNDMK): New.
2855 (UNSPEC_BNDMK_ADDR): New.
2856 (UNSPEC_BNDSTX): New.
2857 (UNSPEC_BNDLDX): New.
2858 (UNSPEC_BNDLDX_ADDR): New.
2859 (UNSPEC_BNDCL): New.
2860 (UNSPEC_BNDCU): New.
2861 (UNSPEC_BNDCN): New.
2862 (UNSPEC_MPX_FENCE): New.
2865 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2866 (length_immediate): Likewise.
2867 (prefix_0f): Likewise.
2869 (prefix_rep): Check for bnd prefix.
2870 (length_nobnd): New.
2871 (length): Use length_nobnd if specified.
2876 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2879 (simple_return_internal): Likewise.
2880 (simple_return_pop_internal): Likewise.
2881 (*indirect_jump): Add MPX bnd prefix.
2882 (*tablejump_1): Likewise.
2883 (simple_return_internal_long): Likewise.
2884 (simple_return_indirect_internal): Likewise.
2888 (*mov<mode>_internal_mpx): New.
2889 (<mode>_<bndcheck>): New.
2890 (*<mode>_<bndcheck>): New.
2895 * config/i386/predicates.md (lea_address_operand): Rename to...
2896 (address_no_seg_operand): ... this.
2897 (address_mpx_no_base_operand): New.
2898 (address_mpx_no_index_operand): New.
2899 (bnd_mem_operator): New.
2900 * config/i386/i386.opt (mmpx): New.
2901 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2902 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2904 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2906 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2907 * tree.def (POINTER_BOUNDS_TYPE): New.
2908 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2909 (POINTER_BOUNDS_MODE): New.
2910 (make_pointer_bounds_mode): New.
2911 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2912 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2913 (layout_type): Support POINTER_BOUNDS_TYPE.
2914 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2915 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2916 (type_contains_placeholder_1): Likewise.
2917 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2918 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2919 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2921 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2923 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2924 mode of address computation when calling memory_address_addr_space.
2926 2013-08-24 Richard Henderson <rth@twiddle.net>
2929 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2930 instead of create_convert_operand_to.
2931 (maybe_emit_sync_lock_test_and_set): Likewise.
2932 (expand_atomic_compare_and_swap): Likewise.
2933 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2935 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2937 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2938 Change param_is to use the struct and not the pointer to the struct.
2940 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2942 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2943 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2944 Move to targhooks.c.
2945 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2946 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2947 Relocate from gimplify.c.
2948 * targhooks.h: Add 2 prototypes.
2949 * tree.h: Delete 2 prototypes.
2951 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2953 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2954 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2956 2013-10-24 Cong Hou <congh@google.com>
2958 * convert.c (convert_to_real): Guard those unsafe math function
2959 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2962 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2965 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2967 (cgraph_create_edge): Use the new argument.
2968 (cgraph_create_indirect_edge): Likewise.
2970 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2972 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2974 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2975 changes statep->state.
2977 2013-10-24 Tobias Burnus <burnus@net-b.de>
2980 * tree-cfg.c (replace_loop_annotate): New function.
2981 (execute_build_cfg): Call it.
2982 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2983 * internal-fn.c (expand_ANNOTATE): New function.
2984 * internal-fn.def (ANNOTATE): Define as new internal function.
2985 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2986 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2987 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2988 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2989 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2991 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2992 Marcus Shawcroft <marcus.shawcroft@arm.com>
2994 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2995 Special case reload SP+C into none GENERAL_REGS.
2997 2013-10-24 Michael Matz <matz@suse.de>
2999 * gengtype.c (is_file_equal): Check that files will be same length.
3001 2013-10-25 Christian Bruel <christian.bruel@st.com>
3003 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
3004 * config/sh/t-sh (sh-mem.o): New rule.
3005 * config/sh/sh-mem.cc (expand_block_move): Moved here.
3006 (sh_expand_cmpstr): New function.
3007 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
3008 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
3009 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
3010 (rotlhi3_8): Rename.
3012 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3014 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
3015 * Makefile.in (CXXDEPMODE): Assign and change users.
3016 (CCDEPMODE): Delete.
3017 * configure: Regenerate.
3019 2013-10-23 David Malcolm <dmalcolm@redhat.com>
3021 * gengtype-parse.c (require_without_advance): New.
3022 (type): For GTY-marked types that are not GTY((user)), parse any
3023 base classes, requiring them to be single-inheritance, and not
3024 be templates. For non-GTY-marked types and GTY((user)),
3025 continue to skip over any C++ inheritance specification.
3026 * gengtype-state.c (state_writer::write_state_struct_type):
3027 Write base class of type (if any).
3028 (read_state_struct_type): Read base class of type (if any).
3029 * gengtype.c (new_structure): Add a "base_class" parameter.
3030 (create_optional_field_): Update for new parameter to new_structure.
3031 (adjust_field_rtx_def): Likewise.
3032 (adjust_field_tree_exp): Likewise.
3033 * gengtype.h (struct type): Add "base_class" field to the s
3035 (new_structure): Add "base" parameter.
3037 2013-10-23 Sriraman Tallam <tmsriram@google.com>
3040 * config/i386/i386.c (ix86_option_override_internal):
3041 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
3042 (ix86_valid_target_attribute_tree):
3043 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
3044 Change TARGET_SSE to TARGET_SSE_P (opts->...)
3046 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3048 * tree-ssa-loop.h: Remove include files.
3049 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3050 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
3051 * graphite-clast-to-gimple.c: Likewise.
3052 * graphite-scop-detection.c: Likewise.
3053 * graphite-sese-to-poly.c: Likewise.
3054 * ipa-inline-analysis.c: Likewise.
3055 * ipa-pure-const.c: Likewise.
3056 * loop-init.c: Likewise.
3057 * passes.c: Likewise.
3058 * predict.c: Likewise.
3059 * tree-cfg.c: Likewise.
3060 * tree-cfgcleanup.c: Likewise.
3061 * tree-chrec.c: Likewise.
3062 * tree-data-ref.c: Likewise.
3063 * tree-loop-distribution.c: Likewise.
3064 * tree-parloops.c: Likewise.
3065 * tree-predcom.c: Likewise.
3066 * tree-scalar-evolution.c: Likewise.
3067 * tree-ssa-address.c: Likewise.
3068 * tree-ssa.c: Likewise.
3069 * tree-ssa-dce.c: Likewise.
3070 * tree-ssa-loop.c: Likewise.
3071 * tree-ssa-loop-im.c: Likewise.
3072 * tree-ssa-loop-ivcanon.c: Likewise.
3073 * tree-ssa-loop-ivopts.c: Likewise.
3074 * tree-ssa-loop-manip.c: Likewise.
3075 * tree-ssa-loop-niter.c: Likewise.
3076 * tree-ssa-loop-prefetch.c: Likewise.
3077 * tree-ssa-loop-unswitch.c: Likewise.
3078 * tree-ssa-reassoc.c: Likewise.
3079 * tree-vect-data-refs.c: Likewise.
3080 * tree-vect-loop.c: Likewise.
3081 * tree-vect-loop-manip.c: Likewise.
3082 * tree-vectorizer.c: Likewise.
3083 * tree-vect-stmts.c: Likewise.
3084 * tree-vrp.c: Likewise.
3086 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3088 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
3090 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3092 PR tree-optimization/58775
3093 PR tree-optimization/58791
3094 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
3095 (insert_stmt_after): Rewritten, don't move the stmt, but really
3097 (get_stmt_uid_with_default): Remove.
3098 (build_and_add_sum): Use insert_stmt_after and
3099 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
3100 (update_range_test): Set uid on stmts added by
3101 force_gimple_operand_gsi. Don't immediately modify statements
3102 in inter-bb optimization, just update oe->op values.
3103 (optimize_range_tests): Return bool whether any changed have been made.
3104 (update_ops): New function.
3105 (struct inter_bb_range_test_entry): New type.
3106 (maybe_optimize_range_tests): Perform statement changes here.
3107 (not_dominated_by, appears_later_in_bb, get_def_stmt,
3108 ensure_ops_are_available): Remove.
3109 (find_insert_point): Rewritten.
3110 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
3111 return LHS of the (new resp. old) stmt. Don't call
3112 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
3113 instead of last, move new stmt at the right place.
3114 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
3115 (negate_value): Likewise. Set uids.
3116 (break_up_subtract_bb): Initialize uids.
3117 (reassociate_bb): Adjust rewrite_expr_tree caller.
3118 (do_reassoc): Don't call renumber_gimple_stmt_uids.
3120 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
3123 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
3124 TARGET_32BIT final condition.
3125 (mulsi3_internal2 and splitter): Same.
3127 2013-10-23 Jeff Law <law@redhat.com>
3129 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
3130 through joiner blocks with abnormal outgoing edges.
3132 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
3133 Add parameter JOINERS, to allow/disallow threading through joiner
3135 (thread_block): New. Call thread_block_1.
3136 (mark_threaded_blocks): Remove code to filter out certain cases
3137 of threading through joiner blocks.
3138 (thread_through_all_blocks): Document how we can have a dangling
3139 edge AUX field and clear it.
3141 2013-10-23 Ian Lance Taylor <iant@google.com>
3143 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
3144 (C++ Dialect Options): Likewise.
3145 (Optimize Options): Likewise.
3147 2013-10-23 Tom de Vries <tom@codesourcery.com>
3149 PR tree-optimization/58805
3150 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
3152 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3154 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
3155 sequence based on get_range_info returned range.
3157 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3159 * tree-ssa.h: Remove all #include's
3160 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3161 * alias.c: Move required includes from tree-ssa.h.
3163 * builtins.c: Likewise.
3164 * calls.c: Likewise.
3165 * cfgexpand.c: Likewise.
3166 * cfghooks.c: Likewise.
3167 * cfgloop.c: Likewise.
3168 * cfgloopmanip.c: Likewise.
3169 * cgraph.c: Likewise.
3170 * cgraphbuild.c: Likewise.
3171 * cgraphclones.c: Likewise.
3172 * cgraphunit.c: Likewise.
3174 * except.c: Likewise.
3176 * final.c: Likewise.
3177 * fold-const.c: Likewise.
3178 * ggc-page.c: Likewise.
3179 * gimple-builder.c: Likewise.
3180 * gimple-fold.c: Likewise.
3181 * gimple-iterator.c: Likewise.
3182 * gimple-low.c: Likewise.
3183 * gimple-pretty-print.c: Likewise.
3184 * gimple-ssa-strength-reduction.c: Likewise.
3185 * gimple-streamer-in.c: Likewise.
3186 * gimple-streamer-out.c: Likewise.
3187 * gimplify.c: Likewise.
3188 * graphite-blocking.c: Likewise.
3189 * graphite-clast-to-gimple.c: Likewise.
3190 * graphite-dependences.c: Likewise.
3191 * graphite-interchange.c: Likewise.
3192 * graphite-optimize-isl.c: Likewise.
3193 * graphite-poly.c: Likewise.
3194 * graphite-scop-detection.c: Likewise.
3195 * graphite-sese-to-poly.c: Likewise.
3196 * graphite.c: Likewise.
3197 * ipa-cp.c: Likewise.
3198 * ipa-inline-analysis.c: Likewise.
3199 * ipa-inline-transform.c: Likewise.
3200 * ipa-inline.c: Likewise.
3201 * ipa-prop.c: Likewise.
3202 * ipa-pure-const.c: Likewise.
3203 * ipa-reference.c: Likewise.
3204 * ipa-split.c: Likewise.
3205 * ipa-utils.c: Likewise.
3206 * loop-init.c: Likewise.
3207 * lto-cgraph.c: Likewise.
3208 * lto-section-in.c: Likewise.
3209 * lto-section-out.c: Likewise.
3210 * lto-streamer-in.c: Likewise.
3211 * lto-streamer-out.c: Likewise.
3212 * lto-streamer.c: Likewise.
3213 * omp-low.c: Likewise.
3214 * passes.c: Likewise.
3215 * predict.c: Likewise.
3216 * print-tree.c: Likewise.
3217 * profile.c: Likewise.
3219 * targhooks.c: Likewise.
3220 * tracer.c: Likewise.
3221 * trans-mem.c: Likewise.
3222 * tree-call-cdce.c: Likewise.
3223 * tree-cfg.c: Likewise.
3224 * tree-cfgcleanup.c: Likewise.
3225 * tree-chrec.c: Likewise.
3226 * tree-complex.c: Likewise.
3227 * tree-data-ref.c: Likewise.
3228 * tree-dfa.c: Likewise.
3229 * tree-eh.c: Likewise.
3230 * tree-emutls.c: Likewise.
3231 * tree-if-conv.c: Likewise.
3232 * tree-inline.c: Likewise.
3233 * tree-into-ssa.c: Likewise.
3234 * tree-loop-distribution.c: Likewise.
3235 * tree-mudflap.c: Likewise.
3236 * tree-nested.c: Likewise.
3237 * tree-nrv.c: Likewise.
3238 * tree-object-size.c: Likewise.
3239 * tree-outof-ssa.c: Likewise.
3240 * tree-parloops.c: Likewise.
3241 * tree-phinodes.c: Likewise.
3242 * tree-predcom.c: Likewise.
3243 * tree-pretty-print.c: Likewise.
3244 * tree-profile.c: Likewise.
3245 * tree-scalar-evolution.c: Likewise.
3246 * tree-sra.c: Likewise.
3247 * tree-ssa-address.c: Likewise.
3248 * tree-ssa-alias.c: Likewise.
3249 * tree-ssa-ccp.c: Likewise.
3250 * tree-ssa-coalesce.c: Likewise.
3251 * tree-ssa-copy.c: Likewise.
3252 * tree-ssa-copyrename.c: Likewise.
3253 * tree-ssa-dce.c: Likewise.
3254 * tree-ssa-dom.c: Likewise.
3255 * tree-ssa-dse.c: Likewise.
3256 * tree-ssa-forwprop.c: Likewise.
3257 * tree-ssa-ifcombine.c: Likewise.
3258 * tree-ssa-live.c: Likewise.
3259 * tree-ssa-loop-ch.c: Likewise.
3260 * tree-ssa-loop-im.c: Likewise.
3261 * tree-ssa-loop-ivcanon.c: Likewise.
3262 * tree-ssa-loop-ivopts.c: Likewise.
3263 * tree-ssa-loop-manip.c: Likewise.
3264 * tree-ssa-loop-niter.c: Likewise.
3265 * tree-ssa-loop-prefetch.c: Likewise.
3266 * tree-ssa-loop-unswitch.c: Likewise.
3267 * tree-ssa-loop.c: Likewise.
3268 * tree-ssa-math-opts.c: Likewise.
3269 * tree-ssa-operands.c: Likewise.
3270 * tree-ssa-phiopt.c: Likewise.
3271 * tree-ssa-phiprop.c: Likewise.
3272 * tree-ssa-pre.c: Likewise.
3273 * tree-ssa-propagate.c: Likewise.
3274 * tree-ssa-reassoc.c: Likewise.
3275 * tree-ssa-sccvn.c: Likewise.
3276 * tree-ssa-sink.c: Likewise.
3277 * tree-ssa-strlen.c: Likewise.
3278 * tree-ssa-structalias.c: Likewise.
3279 * tree-ssa-tail-merge.c: Likewise.
3280 * tree-ssa-ter.c: Likewise.
3281 * tree-ssa-threadedge.c: Likewise.
3282 * tree-ssa-threadupdate.c: Likewise.
3283 * tree-ssa-uncprop.c: Likewise.
3284 * tree-ssa-uninit.c: Likewise.
3285 * tree-ssa.c: Likewise.
3286 * tree-ssanames.c: Likewise.
3287 * tree-stdarg.c: Likewise.
3288 * tree-streamer-in.c: Likewise.
3289 * tree-switch-conversion.c: Likewise.
3290 * tree-tailcall.c: Likewise.
3291 * tree-vect-data-refs.c: Likewise.
3292 * tree-vect-generic.c: Likewise.
3293 * tree-vect-loop-manip.c: Likewise.
3294 * tree-vect-loop.c: Likewise.
3295 * tree-vect-patterns.c: Likewise.
3296 * tree-vect-slp.c: Likewise.
3297 * tree-vect-stmts.c: Likewise.
3298 * tree-vectorizer.c: Likewise.
3299 * tree-vrp.c: Likewise.
3302 * value-prof.c: Likewise.
3303 * var-tracking.c: Likewise.
3304 * varpool.c: Likewise.
3305 * vtable-verify.c: Likewise.
3307 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3309 * config/tilegx/tilegx.c: Include "tree.h".
3311 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3313 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3314 the info, not after it.
3315 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3316 dump_ssaname_info call.
3317 (pp_gimple_stmt_1): Adjust caller.
3318 (dump_phi_nodes): Likewise. Don't print "# " here.
3320 2013-10-22 Jan Hubicka <jh@suse.cz>
3322 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3324 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3325 * i386.c (expand_small_movmem_or_setmem): New function.
3326 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3328 (alg_usable_p): Add support for value ranges; cleanup.
3329 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3331 2013-10-22 Sterling Augustine <saugustine@google.com>
3333 * doc/invoke.texi: Document -ggnu-pubnames.
3334 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3336 * dwarf2out.c: Include gdb/gdb-index.h.
3337 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3338 debug_generate_pub_sections.
3339 (is_java, output_pubtables, output_pubname): New functions.
3340 (include_pubname_in_output): Handle debug_generate_pub_sections at
3342 (size_of_pubnames): Use new local space_for_flags based on
3343 debug_generate_pub_sections.
3344 (output_pubnames): Unify pubnames and pubtypes output logic.
3345 Genericize comments. Call output_pubname.
3346 (dwarf2out_finish): Move logic to output_pubnames and call it.
3348 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3351 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3352 Remove CCCmode handling.
3353 <case LTU>: Return 'c' suffix for CCCmode.
3354 <case GEU>: Return 'nc' suffix for CCCmode.
3355 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3356 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3357 (*sub<mode>3_cc_overflow): Ditto.
3358 (*subsi3_zext_cc_overflow): Ditto.
3360 2013-10-22 Steve Ellcey <sellcey@mips.com>
3362 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3363 (AND (NOT OP1) (NOT OP2)).
3365 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3367 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3368 meaning of merge-high and merge-low masks for little endian; avoid
3369 use of vector-pack masks for little endian for mismatched modes.
3371 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3373 * config/tilepro/tilepro.c: Include "tree.h".
3375 2013-10-22 Andreas Schwab <schwab@suse.de>
3377 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3380 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3382 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3383 [Ed Smith-Rowland]): New entries.
3384 ([Stephen M. Webb]): Update.
3386 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3388 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3390 2013-10-21 Marek Polacek <polacek@redhat.com>
3393 * fold-const.c (fold_range_test): Return 0 if the type is not
3396 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3398 * system.h: Move hwint.h include further down.
3399 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3400 gcc_checking_asserts.
3401 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3403 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3405 Fix volatile issues in optimize_bit_field_compare.
3406 * fold-const.c (optimize_bit_field_compare): Bail out if
3407 lvolatilep or rvolatilep.
3409 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3411 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3412 and get_inner_reference returning different pmode for non-volatile
3413 bit-field members dependent on flag_strict_volatile_bitfields.
3414 * stor-layout.c (layout_decl): Remove special handling of
3415 flag_strict_volatile_bitfields.
3416 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3417 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3419 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3421 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3424 2013-10-21 Jeff Law <law@redhat.com>
3426 * tree-ssa-threadedge.c (thread_through_normal_block): New
3427 argument VISITED. Remove VISISTED as a local variable. When we
3428 have a threadable jump, verify the destination of the jump has not
3430 (thread_across_edge): Allocate VISITED bitmap once at function
3431 scope and use it throughout. Make sure to set appropriate bits in
3432 VISITED for E (start of jump thread path).
3433 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3434 through a joiner if any edge on the path has a recorded jump thread.
3436 2013-10-21 Ian Lance Taylor <iant@google.com>
3438 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3439 don't imply that attributes can solve all problems.
3440 (Directory Options): Fix typo.
3442 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3444 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3445 extend and extend_add.
3447 2013-10-21 Richard Biener <rguenther@suse.de>
3449 PR tree-optimization/58794
3450 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3451 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3453 2013-10-21 Richard Biener <rguenther@suse.de>
3456 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3457 to (T) X for sign-changing conversions (or no conversion).
3459 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3461 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3463 2013-10-20 Jan Hubicka <jh@suse.cz>
3465 * config/i386/i386-tune.def: Add comment; organize into categories
3467 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3469 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3470 argument. Update function comment.
3471 (expand_set_or_movmem_via_rep): New function combining
3472 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3473 (expand_movmem_via_rep_mov): Remove.
3474 expand_setmem_via_rep_stos): Remove.
3475 (expand_movmem_epilogue): Update calls correspondingly.
3476 (expand_setmem_epilogue_via_loop): Likewise.
3478 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3479 (expand_set_or_movmem_prologue): New function combining
3480 expand_movmem_prologue and expand_setmem_prologue.
3481 (expand_movmem_prologue): Remove.
3482 (expand_setmem_prologue): Remove.
3483 (expand_set_or_movmem_constant_prologue): New function combining
3484 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3485 (expand_constant_movmem_prologue): Remove.
3486 (expand_constant_setmem_prologue): Remove.
3487 (promote_duplicated_reg): Allow vector-const0 value.
3488 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3489 and ix86_expand_setmem.
3490 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3491 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3493 2013-10-21 Diego Novillo <dnovillo@google.com>
3495 * asan.c: Include tree.h
3496 * bb-reorder.c: Likewise.
3497 * cfgcleanup.c: Likewise.
3498 * cfgloopmanip.c: Likewise.
3499 * data-streamer-in.c: Likewise.
3500 * data-streamer-out.c: Likewise.
3501 * data-streamer.c: Likewise.
3502 * dwarf2cfi.c: Likewise.
3503 * graphite-blocking.c: Likewise.
3504 * graphite-clast-to-gimple.c: Likewise.
3505 * graphite-dependences.c: Likewise.
3506 * graphite-interchange.c: Likewise.
3507 * graphite-optimize-isl.c: Likewise.
3508 * graphite-poly.c: Likewise.
3509 * graphite-scop-detection.c: Likewise.
3510 * graphite-sese-to-poly.c: Likewise.
3511 * graphite.c: Likewise.
3512 * ipa-devirt.c: Likewise.
3513 * ipa-profile.c: Likewise.
3516 * loop-init.c: Likewise.
3517 * loop-unroll.c: Likewise.
3518 * lower-subreg.c: Likewise.
3519 * lto/lto-object.c: Likewise.
3520 * recog.c: Likewise.
3521 * reginfo.c: Likewise.
3522 * tree-loop-distribution.c: Likewise.
3523 * tree-parloops.c: Likewise.
3524 * tree-ssa-strlen.c: Likewise.
3525 * tree-streamer.c: Likewise.
3526 * value-prof.c: Likewise.
3527 * target-globals.c: Likewise.
3528 * expr.h: Include tree-core.h instead of tree.h.
3529 * gimple.h: Likewise.
3530 * ipa-prop.h: Likewise.
3531 * ipa-utils.h: Likewise.
3532 * lto-streamer.h: Likewise.
3533 * streamer-hooks.h: Likewise.
3534 * ipa-reference.h: Include cgraph.h instead of tree.h.
3535 * cgraph.h: Include basic-block.h instead of tree.h.
3536 * tree-streamer.h: Do not include tree.h.
3537 * genattrtab.c (write_header): Generate inclusion of
3539 * genautomata.c (main): Likewise.
3540 * genemit.c: Likewise.
3541 * genopinit.c: Likewise.
3542 * genoutput.c (output_prologue): Likewise.
3543 * genpeep.c: Likewise.
3545 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3547 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3549 (vec_unpacku_hi_v8hi): Likewise.
3550 (vec_unpacku_lo_v16qi): Likewise.
3551 (vec_unpacku_lo_v8hi): Likewise.
3553 2013-10-20 Jan Hubicka <jh@suse.cz>
3555 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3556 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3557 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3559 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3561 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3562 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3564 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3566 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3569 * system.h: Undef m_slot.
3571 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3573 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3574 all elements for both endian flavors.
3576 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3579 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3580 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3581 and SI_REG for 64bit SYSV ABI targets.
3583 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3585 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3586 to multi_reg_return. Clarify that we are skipping USEs of multiple
3587 return registers. Use bool type where appropriate.
3589 2013-10-18 Jan Hubicka <jh@suse.cz>
3591 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3593 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3594 (X86_TUNE_PUSH_MEMORY): Likewise.
3595 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3596 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3597 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3599 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3600 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3602 (ix86_option_override_internal): Update to use tune features instead
3605 2013-10-18 Cong Hou <congh@google.com>
3607 PR tree-optimization/58508
3608 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3609 statement that contains data refs with zero-step.
3611 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3613 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3614 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3615 * gimple-low.c (gimple_check_call_arg,
3616 gimple_check_call_matching_types): Move to cgraph.c.
3617 * gimple-low.h: Remove prototype.
3618 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3619 Relocate from gimple-low.c.
3620 * cgraph.h: Add prototype. Don't include basic-block.h.
3621 * gimplify.c: Add gimple-low to include list.
3622 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3623 * tree-eh.c: Add gimple-low to include list.
3624 * tree-nested.c: Likewise.
3625 * cfgexpand.c: Add tree-ssa-address.h to include list.
3627 * gimple-fold.c: Likewise.
3628 * gimple-ssa-strength-reduction.c: Likewise.
3629 * trans-mem.c: Likewise.
3630 * tree-mudflap.c: Likewise.
3631 * tree-ssa-loop-ivopts.c: Likewise.
3632 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3633 (degenerate_phi_result): Move to tree-phinodes.c.
3634 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3635 * tree-ssa-threadedge.c: Likewise.
3636 * tree-vrp.c: Likewise.
3637 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3638 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3639 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3640 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3641 * tree-ssa-forwprop.c: Likewise.
3642 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3643 pass_data_cleanup_cfg_post_optimizing,
3644 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3645 * tree-optimize.c: Delete File.
3646 * graphite.c: Include tree-cfgcleanup.h.
3647 * passes.c: Likewise.
3648 * tree-cfg.c: Likewise.
3649 * tree-profile.c: Likewise.
3650 * tree-ssa-dse.c: Likewise.
3651 * tree-ssa-loop-ivcanon.c: Likewise.
3652 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3653 * tree-outof-ssa.c: Include sbitmap.h.
3654 * tree-ssa-live.c: Likewise.
3655 * tree-ssa-propagate.c: Likewise.
3656 * tree-ssa-structalias.c: Likewise.
3657 * tree-stdarg.c: Likewise.
3658 * Makefile.in (OBJS): Delete tree-optimize.o.
3659 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3660 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3661 * varasm.c: Include basic-block.h.
3662 * cfgloop.h: Include function.h instead of basic-block.h
3663 (bb_loop_depth): Move to cfgloop.c.
3664 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3666 2013-10-18 Teresa Johnson <tejohnson@google.com>
3668 * predict.c (probably_never_executed): Compare frequency-based
3669 count to number of training runs.
3670 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3672 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3674 * config/arm/arm.c (cortexa9_extra_costs): New table.
3675 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3677 2013-10-18 Jeff Law <law@redhat.com>
3679 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3681 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3682 (register_jump_thread): Add "registered_jump_thread" debug
3684 * dbgcnt.def (registered_jump_thread): New debug counter.
3686 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3688 * config/rs6000/rs6000.c: Include cgraph.h.
3690 2013-10-18 Teresa Johnson <tejohnson@google.com>
3692 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3693 weights during merging.
3695 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3697 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3698 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3699 * tree-ssa-operands.h: Remove prototype.
3700 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3701 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3702 * gimple.h: Add prototype.
3703 * gimple-ssa.h: Include tree-ssa-operands.h.
3704 * tree-dump.c: Add tree-cfg.h to include list.
3705 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3706 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3707 * config/i386/i386.c: Don't include tree-flow.h.
3708 * config/rs6000/rs6000.c: Likewise.
3710 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3712 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3714 2013-10-18 Richard Biener <rguenther@suse.de>
3716 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3717 or TYPE_UNSIGNED of integral types.
3718 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3719 NULL_TREE for zero-precision integral types.
3721 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3723 * config/aarch64/arm_neon.h
3724 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3726 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3729 * opth-gen.awk: Define target_flags_explicit.
3731 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3733 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3734 fields to the reg_addr array that describes the valid addressing
3735 mode for any register, general purpose registers, floating point
3736 registers, and Altivec registers.
3737 (FIRST_RELOAD_REG_CLASS): Likewise.
3738 (LAST_RELOAD_REG_CLASS): Likewise.
3739 (struct reload_reg_map_type): Likewise.
3740 (reload_reg_map_type): Likewise.
3741 (RELOAD_REG_VALID): Likewise.
3742 (RELOAD_REG_MULTIPLE): Likewise.
3743 (RELOAD_REG_INDEXED): Likewise.
3744 (RELOAD_REG_OFFSET): Likewise.
3745 (RELOAD_REG_PRE_INCDEC): Likewise.
3746 (RELOAD_REG_PRE_MODIFY): Likewise.
3747 (reg_addr): Likewise.
3748 (mode_supports_pre_incdec_p): New helper functions to say whether
3749 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3750 (mode_supports_pre_modify_p): Likewise.
3751 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3752 print the valid address mode bits for each mode.
3753 (rs6000_debug_print_mode): Likewise.
3754 (rs6000_debug_reg_global): Likewise.
3755 (rs6000_setup_reg_addr_masks): New function to set up the address
3756 mask bits for each type.
3757 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3758 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3759 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3760 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3761 PRE_MODIFY are supported.
3762 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3763 registers, instead of {src,dest}_av_p.
3764 (rs6000_print_options_internal): Tweak the debug output slightly.
3766 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3768 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3771 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3773 * tree-flow.h (struct omp_region): Move to omp-low.c.
3774 Remove omp_ prototypes and variables.
3775 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3776 (copy_var_decl): Relocate prototype from tree-flow.h.
3777 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3778 * tree.h: Move prototype to omp-low.h.
3779 * omp-low.h: New File. Relocate prototypes here.
3780 * omp-low.c (struct omp_region): Make local here.
3781 (root_omp_region): Make static.
3782 (copy_var_decl) Move to gimple.c.
3783 (new_omp_region): Make static.
3784 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3785 * tree-cfg.c: Include omp-low.h.
3786 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3787 * gimplify.c: Include omp-low.h.
3788 * tree-parloops.c: Likewise.
3790 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3792 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3793 operands are in memory, prefer to force non-matched operand 1 to
3796 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3799 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3800 restrict TImode addresses to single indirect registers if both
3801 -mquad-memory and -mvsx-timode are used.
3802 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3803 we should emit load/store quad. Remove using %y for quad memory
3806 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3807 constraints to allow load/store quad on machines where TImode is
3808 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3809 for TImode to load integer constants.
3811 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3813 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3815 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3817 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3818 handling of STACK_REG.
3820 2013-10-17 Richard Biener <rguenther@suse.de>
3822 PR tree-optimization/58143
3823 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3825 (rewrite_to_defined_overflow): Likewise.
3826 (move_computations_dom_walker::before_dom): Rewrite stmts
3827 with undefined signed overflow that are not always executed
3828 into unsigned arithmetic.
3830 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3833 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3834 explicit isa flag to be an options variable, instead of using
3835 global_options_set. Remove define from rs6000.h.
3836 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3838 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3839 Initialize rs6000_isa_flags_explicit.
3840 (rs6000_function_specific_save): Add gcc_options* parameter, so
3841 that the powerpc builds after the 2013-10-15 changes.
3842 (rs6000_function_specific_restore): Likewise.
3844 2013-10-16 DJ Delorie <dj@redhat.com>
3846 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3847 op is a REG before checking REGNO.
3848 (rl78_alloc_physical_registers): Verify pattern is a SET before
3851 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3853 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3855 (vec_unpacks_lo_v4sf): Likewise.
3856 (vec_unpacks_float_hi_v4si): Likewise.
3857 (vec_unpacks_float_lo_v4si): Likewise.
3858 (vec_unpacku_float_hi_v4si): Likewise.
3859 (vec_unpacku_float_lo_v4si): Likewise.
3861 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3863 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3864 (vsx_concat_v2sf): Likewise.
3866 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3868 * config/aarch64/aarch64.md
3869 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3871 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3873 PR tree-optimization/58697
3874 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3875 estimated_loop_iterations_int.
3876 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3877 (get_max_loop_iterations_int): New. HWINT version of
3878 get_max_loop_iterations.
3879 * cfgloop.h: Add prototypes.
3880 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3881 * loop-unroll.c (decide_peel_once_rolling): Call
3882 get_estimated_loop_iterations_int.
3883 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3884 * tree-ssa-loop-niter.h: Tweak prototypes.
3886 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3888 * gengtype-parse.c (struct_field_seq): Ignore access-control
3889 keywords ("public:" etc).
3891 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3893 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3894 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3896 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3898 * config/arm/arm.opt (mlra): New option.
3899 * config/arm/arm.c (arm_lra_p): New function.
3900 (TARGET_LRA_P): Define.
3902 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3904 * tree-core.h (tree_code_name): Remove.
3905 * tree.h (get_tree_code_name): New prototype.
3906 * tree.c (tree_code_name): Make static.
3907 (get_tree_code_name): New function.
3908 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3909 tree_class_check_failed, tree_range_check_failed,
3910 tree_not_class_check_failed, omp_clause_check_failed,
3911 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3912 wrapper get_tree_code_name instead of calling tree_code_name directly.
3913 * tree-vrp.c (dump_asserts_for): Likewise.
3914 * tree-dump.c (dequeue_and_dump): Likewise.
3915 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3916 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3917 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3918 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3919 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3920 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3921 dump_gimple_omp_for): Likewise.
3922 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3923 * tree-ssa-pre.c (print_pre_expr): Likewise.
3924 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3925 * print-tree.c (print_node_brief, print_node): Likewise.
3926 * gimple.c (gimple_check_failed): Likewise.
3927 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3928 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3929 * config/mep/mep.c (mep_validate_vliw): Likewise.
3930 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3931 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3933 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3935 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3938 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3940 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3941 (MULTILIB_MATCHES): Add multilib for -march=v8.
3943 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3946 * optc-save-gen.awk: Add extra parameter to the save and restore
3948 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3949 * tree.c (build_optimization_node): New parameter. Add extra parameter
3950 to call to cl_optimization_save.
3951 (build_target_option_node): New parameter. Add extra parameter
3952 to call to cl_target_option_save.
3953 * tree.h (build_optimization_node): New parameter.
3954 (build_target_option_node): New parameter.
3955 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3956 build_optimization_node and build_target_option_node.
3957 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3958 (handle_pragma_push_options): Ditto.
3959 * toplev.c (process_options): Ditto.
3960 * opts.c (init_options_struct): Check for opts_set non-null.
3961 * target.def (target_option.save): New parameter.
3962 (target_option.restore): New parameter.
3963 * tm.texi: Generate.
3964 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3965 (ix86_pragma_target_parse): Ditto.
3966 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3968 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3969 to build_optimization_node and build_target_option_node.
3970 (rs6000_valid_attribute_p): Ditto.
3971 (rs6000_pragma_target_parse): Ditto.
3972 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3974 * config/i386/i386.h:
3975 TARGET_64BIT_P: New Macro
3976 TARGET_MMX_P: New Macro.
3977 TARGET_3DNOW_P: New Macro.
3978 TARGET_3DNOW_A_P: New Macro.
3979 TARGET_SSE_P: New Macro.
3980 TARGET_SSE2_P: New Macro.
3981 TARGET_SSE3_P: New Macro.
3982 TARGET_SSSE3_P: New Macro.
3983 TARGET_SSE4_1_P: New Macro.
3984 TARGET_SSE4_2_P: New Macro.
3985 TARGET_AVX_P: New Macro.
3986 TARGET_AVX2_P: New Macro.
3987 TARGET_AVX512F_P: New Macro.
3988 TARGET_AVX512PF_P: New Macro.
3989 TARGET_AVX512ER_P: New Macro.
3990 TARGET_AVX512CD_P: New Macro.
3991 TARGET_FMA_P: New Macro.
3992 TARGET_SSE4A_P: New Macro.
3993 TARGET_FMA4_P: New Macro.
3994 TARGET_XOP_P: New Macro.
3995 TARGET_LWP_P: New Macro.
3996 TARGET_ABM_P: New Macro.
3997 TARGET_BMI_P: New Macro.
3998 TARGET_BMI2_P: New Macro.
3999 TARGET_LZCNT_P: New Macro.
4000 TARGET_TBM_P: New Macro.
4001 TARGET_POPCNT_P: New Macro.
4002 TARGET_SAHF_P: New Macro.
4003 TARGET_MOVBE_P: New Macro.
4004 TARGET_CRC32_P: New Macro.
4005 TARGET_AES_P: New Macro.
4006 TARGET_PCLMUL_P: New Macro.
4007 TARGET_CMPXCHG16B_P: New Macro.
4008 TARGET_FSGSBASE_P: New Macro.
4009 TARGET_RDRND_P: New Macro.
4010 TARGET_F16C_P: New Macro.
4011 TARGET_RTM_P: New Macro.
4012 TARGET_HLE_P: New Macro.
4013 TARGET_RDSEED_P: New Macro.
4014 TARGET_PRFCHW_P: New Macro.
4015 TARGET_ADX_P: New Macro.
4016 TARGET_FXSR_P: New Macro.
4017 TARGET_XSAVE_P: New Macro.
4018 TARGET_XSAVEOPT_P: New Macro.
4019 TARGET_LP64_P: New Macro.
4020 TARGET_X32_P: New Macro.
4021 TARGET_FPMATH_DEFAULT_P: New Macro.
4022 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
4023 * config/i386/i386.c (ix86_option_override_internal): New parameters.
4025 Change ix86_tune_string to access opts->x_ix86_tune_string.
4026 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
4027 Change ix86_arch_string to access opts->x_ix86_arch_string.
4028 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
4029 Change ix86_pmode to access opts->x_ix86_pmode.
4030 Change ix86_abi to access opts->x_ix86_abi.
4031 Change ix86_cmodel to access opts->x_ix86_cmodel.
4032 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
4033 Change ix86_isa_flags_explicit to access
4034 opts->x_ix86_isa_flags_explicit.
4035 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
4036 Change ix86_regparm to access opts->x_ix86_regparm.
4037 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
4038 Change ix86_preferred_stack_boundary_arg to access
4039 opts->x_ix86_preferred_stack_boundary_arg.
4040 Change ix86_force_align_arg_pointer to access
4041 opts->x_ix86_force_align_arg_pointer.
4042 Change ix86_incoming_stack_boundar_arg to access
4043 opts->x_ix86_incoming_stack_boundar_arg.
4044 Change ix86_fpmath to access opts->x_ix86_fpmath.
4045 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
4046 Change ix86_recip_name to access opts->x_ix86_recip_name.
4047 Change ix86_stack_protector_guard to access
4048 opts->x_ix86_stack_protector_guard.
4049 Change ix86_tune_memcpy_strategy to access
4050 opts->x_ix86_tune_memcpy_strategy.
4051 Change ix86_tune_memset_strategy to access
4052 opts->x_ix86_tune_memset_strategy.
4053 Change global_options to access opts.
4054 Change global_options_set to access opts_set.
4055 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
4056 Change TARGET_MMX to TARGET_MMX_P (opts->...).
4057 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
4058 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
4059 Change TARGET_SSE to TARGET_SSE_P (opts->...).
4060 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
4061 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
4062 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
4063 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
4064 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
4065 Change TARGET_AVX to TARGET_AVX_P (opts->...).
4066 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
4067 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
4068 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
4069 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
4070 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
4071 Change TARGET_FMA to TARGET_FMA_P (opts->...).
4072 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
4073 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
4074 Change TARGET_XOP to TARGET_XOP_P (opts->...).
4075 Change TARGET_LWP to TARGET_LWP_P (opts->...).
4076 Change TARGET_ABM to TARGET_ABM_P (opts->...).
4077 Change TARGET_BMI to TARGET_BMI_P (opts->...).
4078 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
4079 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
4080 Change TARGET_TBM to TARGET_TBM_P (opts->...).
4081 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
4082 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
4083 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
4084 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
4085 Change TARGET_AES to TARGET_AES_P (opts->...).
4086 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
4087 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
4088 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
4089 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
4090 Change TARGET_F16C to TARGET_F16C_P (opts->...).
4091 Change TARGET_RTM to TARGET_RTM_P (opts->...).
4092 Change TARGET_HLE to TARGET_HLE_P (opts->...).
4093 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
4094 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
4095 Change TARGET_ADX to TARGET_ADX_P (opts->...).
4096 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
4097 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
4098 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
4099 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
4100 Change TARGET_X32 to TARGET_X32_P (opts->...).
4101 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
4102 Change TARGET_FLOAT_RETURNS_IN_80387 to
4103 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
4104 (ix86_function_specific_save): New parameter. Use opts-> fields
4105 to replace global fields.
4106 (ix86_function_specific_restore): Ditto.
4107 (ix86_valid_target_attribute_inner_p): New parameters.
4109 Fix call to ix86_handle_option and set_option.
4110 (ix86_valid_target_attribute_tree): New parameters.
4111 Change global_options to access opts.
4112 Change global_options_set to access opts_set.
4113 Fix call to ix86_valid_target_attribute_inner_p.
4114 Change ix86_tune_string to access opts->x_ix86_tune_string.
4115 Change ix86_arch_string to access opts->x_ix86_arch_string.
4116 Change ix86_fpmath to access opts->x_ix86_fpmath
4117 Fix call to ix86_option_override_internal.
4118 Fix call to ix86_add_new_builtins.
4119 Fix calls to build_optimization_node and build_target_option_node.
4120 (ix86_valid_target_attribute_p): Remove access to global_options.
4121 Use new gcc_options structure func_options.
4122 Fix call to ix86_valid_target_attribute_tree.
4123 Fix call to build_optimization_node.
4124 (get_builtin_code_for_version): Fix call to
4125 ix86_valid_target_attribute_tree.
4127 2013-10-15 David Malcolm <dmalcolm@redhat.com>
4129 * Makefile.in (PICFLAG): New.
4130 (enable_host_shared): New.
4131 (INTERNAL_CFLAGS): Use PICFLAG.
4132 (LIBIBERTY): Use pic build of libiberty.a if configured with
4133 --enable-host-shared.
4134 * configure.ac: Add --enable-host-shared, setting up new
4136 * configure: Regenerate.
4137 * doc/install.texi (--enable-shared): Add note contrasting it with ...
4138 (--enable-host-shared): New option.
4140 2013-10-15 Richard Biener <rguenther@suse.de>
4142 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
4143 for built-in functions.
4145 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
4147 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
4148 (optimize_range_tests_1): New function,
4149 extracted from optimize_range_tests.
4150 (optimize_range_tests_xor): Similarly.
4151 (optimize_range_tests_diff): New function.
4152 (optimize_range_tests): Use optimize_range_tests_1.
4154 2013-10-15 Cong Hou <congh@google.com>
4156 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
4157 requirement of the reduction pattern so that one operand of the
4158 reduction operation can come from outside of the loop.
4160 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4162 * config/arm/neon-schedgen.ml: Remove.
4163 * config/arm/cortex-a9-neon.md: Remove comment regarding
4166 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4168 * config/arm/types: Remove old neon types.
4170 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4172 * config/arm/cortex-a7.md
4173 (cortex_a7_neon_type): New.
4174 (cortex_a7_neon_mul): Update for new types.
4175 (cortex_a7_neon_mla): Likewise.
4176 (cortex_a7_neon): Likewise.
4178 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4180 * config/arm/cortex-a15-neon.md
4181 (cortex_a15_neon_type): New,
4183 (cortex_a15_neon_int_1): Remove.
4184 (cortex_a15_neon_int_2): Likewise.
4185 (cortex_a15_neon_int_3): Likewise.
4186 (cortex_a15_neon_int_4): Likewise.
4187 (cortex_a15_neon_int_5): Likewise.
4188 (cortex_a15_neon_vqneg_vqabs): Likewise.
4189 (cortex_a15_neon_vmov): Likewise.
4190 (cortex_a15_neon_vaba): Likewise.
4191 (cortex_a15_neon_vaba_qqq): Likewise.
4192 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4193 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4194 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4196 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4197 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4198 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
4199 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4200 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4201 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4202 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4203 (cortex_a15_neon_shift_1): Likewise.
4204 (cortex_a15_neon_shift_2): Likewise.
4205 (cortex_a15_neon_shift_3): Likewise.
4206 (cortex_a15_neon_vshl_ddd): Likewise.
4207 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4208 (cortex_a15_neon_vsra_vrsra): Likewise.
4209 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4210 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4211 (cortex_a15_neon_bp_3cycle): Likewise.
4212 (cortex_a15_neon_ldm_2): Likewise.
4213 (cortex_a15_neon_stm_2): Likewise.
4214 (cortex_a15_neon_mcr): Likewise.
4215 (cortex_a15_neon_mrc): Likewise.
4216 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4217 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4218 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4219 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4220 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4221 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4222 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4223 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4224 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4225 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4226 (cortex_a15_neon_bp_simple): Likewise.
4227 (cortex_a15_neon_bp_2cycle): Likewise.
4228 (cortex_a15_neon_bp_3cycle): Likewise.
4229 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4230 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4231 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4232 (cortex_a15_neon_vld2_4_regs): Likewise.
4233 (cortex_a15_neon_vld3_vld4): Likewise.
4234 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4235 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4236 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
4237 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
4238 (cortex_a15_neon_vst3_vst4): Rename to...
4239 (cortex_a15_neon_vst4): This, update for new attributes.
4240 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
4241 (cortex_a15_neon_vld3_vld4_lane): Likewise.
4242 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4243 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4244 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4245 (cortex_a15_neon_ldm_2): Likewise.
4246 (cortex_a15_neon_stm_2): Likewise.
4247 (cortex_a15_neon_mcr): Likewise.
4248 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4249 (cortex_a15_neon_mrc): Likewise.
4250 (cortex_a15_neon_mrrc): Likewise.
4252 (cortex_a15_neon_abd): New.
4253 (cortex_a15_neon_abd_q): Likewise.
4254 (cortex_a15_neon_aba): Likewise.
4255 (cortex_a15_neon_aba_q): Likewise.
4256 (cortex_a15_neon_acc): Likewise.
4257 (cortex_a15_neon_acc_q): Likewise.
4258 (cortex_a15_neon_arith_basic): Likewise.
4259 (cortex_a15_neon_arith_complex): Likewise.
4260 (cortex_a15_neon_multiply): Likewise.
4261 (cortex_a15_neon_multiply_q): Likewise.
4262 (cortex_a15_neon_mla): Likewise.
4263 (cortex_a15_neon_mla_q): Likewise.
4264 (cortex_a15_neon_sat_mla_long): Likewise.
4265 (cortex_a15_neon_shift_acc): Likewise.
4266 (cortex_a15_neon_shift_imm_basic): Likewise.
4267 (cortex_a15_neon_shift_imm_complex): Likewise.
4268 (cortex_a15_neon_shift_reg_basic): Likewise.
4269 (cortex_a15_neon_shift_reg_basic_q): Likewise.
4270 (cortex_a15_neon_shift_reg_complex): Likewise.
4271 (cortex_a15_neon_shift_reg_complex_q): Likewise.
4272 (cortex_a15_neon_fp_negabs): Likewise
4273 (cortex_a15_neon_fp_arith): Likewise
4274 (cortex_a15_neon_fp_arith_q): Likewise
4275 (cortex_a15_neon_fp_cvt_int): Likewise
4276 (cortex_a15_neon_fp_cvt_int_q): Likewise
4277 (cortex_a15_neon_fp_cvt_16): Likewise
4278 (cortex_a15_neon_fp_mul): Likewise
4279 (cortex_a15_neon_fp_mul_q): Likewise
4280 (cortex_a15_neon_fp_mla): Likewise
4281 (cortex_a15_neon_fp_mla_q): Likewise
4282 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
4283 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
4284 (cortex_a15_neon_bitops): Likewise.
4285 (cortex_a15_neon_bitops_q): Likewise.
4286 (cortex_a15_neon_from_gp): Likewise.
4287 (cortex_a15_neon_from_gp_q): Likewise.
4288 (cortex_a15_neon_tbl3_tbl4): Likewise.
4289 (cortex_a15_neon_zip_q): Likewise.
4290 (cortex_a15_neon_to_gp): Likewise.
4291 (cortex_a15_neon_load_a): Likewise.
4292 (cortex_a15_neon_load_b): Likewise.
4293 (cortex_a15_neon_load_c): Likewise.
4294 (cortex_a15_neon_load_d): Likewise.
4295 (cortex_a15_neon_load_e): Likewise.
4296 (cortex_a15_neon_load_f): Likewise.
4297 (cortex_a15_neon_store_a): Likewise.
4298 (cortex_a15_neon_store_b): Likewise.
4299 (cortex_a15_neon_store_c): Likewise.
4300 (cortex_a15_neon_store_d): Likewise.
4301 (cortex_a15_neon_store_e): Likewise.
4302 (cortex_a15_neon_store_f): Likewise.
4303 (cortex_a15_neon_store_g): Likewise.
4304 (cortex_a15_neon_store_h): Likewise.
4305 (cortex_a15_vfp_to_from_gp): Likewise.
4307 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4309 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4311 (cortex_a9_neon_vshl_ddd): Remove.
4312 (cortex_a9_neon_vst3_vst4): Likewise.
4313 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4315 (cortex_a9_neon_bit_ops_q): New.
4317 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4318 (cortex_a9_neon_int_2): Likewise.
4319 (cortex_a9_neon_int_3): Likewise.
4320 (cortex_a9_neon_int_4): Likewise.
4321 (cortex_a9_neon_int_5): Likewise.
4322 (cortex_a9_neon_vqneg_vqabs): Likewise.
4323 (cortex_a9_neon_vmov): Likewise.
4324 (cortex_a9_neon_vaba): Likewise.
4325 (cortex_a9_neon_vaba_qqq): Likewise.
4326 (cortex_a9_neon_shift_1): Likewise.
4327 (cortex_a9_neon_shift_2): Likewise.
4328 (cortex_a9_neon_shift_3): Likewise.
4329 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4330 (cortex_a9_neon_vsra_vrsra): Likewise.
4331 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4332 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4333 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4335 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4336 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4337 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4339 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4340 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4341 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4342 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4343 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4344 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4345 (cortex_a9_neon_fp_vsum): Likewise.
4346 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4347 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4348 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4349 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4350 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4351 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4352 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4353 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4354 (cortex_a9_neon_bp_simple): Likewise.
4355 (cortex_a9_neon_bp_2cycle): Likewise.
4356 (cortex_a9_neon_bp_3cycle): Likewise.
4357 (cortex_a9_neon_ldr): Likewise.
4358 (cortex_a9_neon_str): Likewise.
4359 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4360 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4361 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4362 (cortex_a9_neon_vld2_4_regs): Likewise.
4363 (cortex_a9_neon_vld3_vld4): Likewise.
4364 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4365 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4366 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4367 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4368 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4369 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4370 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4371 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4372 (cortex_a9_neon_mcr): Likewise.
4373 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4374 (cortex_a9_neon_mrc): Likewise.
4375 (cortex_a9_neon_mrrc): Likewise.
4377 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4379 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4381 (cortex_a8_neon_vshl_ddd): Remove.
4382 (cortex_a8_neon_vst3_vst4): Likewise.
4383 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4385 (cortex_a8_neon_bit_ops_q): New.
4387 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4388 (cortex_a8_neon_int_2): Likewise..
4389 (cortex_a8_neon_int_3): Likewise.
4390 (cortex_a8_neon_int_5): Likewise.
4391 (cortex_a8_neon_vqneg_vqabs): Likewise.
4392 (cortex_a8_neon_int_4): Likewise.
4393 (cortex_a8_neon_vaba): Likewise.
4394 (cortex_a8_neon_vaba_qqq): Likewise.
4395 (cortex_a8_neon_shift_1): Likewise.
4396 (cortex_a8_neon_shift_2): Likewise.
4397 (cortex_a8_neon_shift_3): Likewise.
4398 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4399 (cortex_a8_neon_vsra_vrsra): Likewise.
4400 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4401 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4402 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4404 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4405 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4406 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4408 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4409 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4410 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4411 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4412 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4413 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4414 (cortex_a8_neon_fp_vsum): Likewise.
4415 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4416 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4417 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4418 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4419 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4420 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4421 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4422 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4423 (cortex_a8_neon_bp_simple): Likewise.
4424 (cortex_a8_neon_bp_2cycle): Likewise.
4425 (cortex_a8_neon_bp_3cycle): Likewise.
4426 (cortex_a8_neon_ldr): Likewise.
4427 (cortex_a8_neon_str): Likewise.
4428 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4429 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4430 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4431 (cortex_a8_neon_vld2_4_regs): Likewise.
4432 (cortex_a8_neon_vld3_vld4): Likewise.
4433 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4434 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4435 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4436 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4437 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4438 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4439 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4440 (cortex_a8_neon_mcr): Likewise.
4441 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4442 (cortex_a8_neon_mrc): Likewise.
4443 (cortex_a8_neon_mrrc): Likewise.
4445 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4447 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4449 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4450 (aarch64_simd_dup<mode>): Add "type" attribute.
4451 (aarch64_dup_lane<mode>): Likewise.
4452 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4453 (*aarch64_simd_mov<mode>): Likewise.
4454 (aarch64_simd_mov_from_<mode>low): Likewise.
4455 (aarch64_simd_mov_from_<mode>high): Likewise.
4456 (orn<mode>3): Likewise.
4457 (bic<mode>3): Likewise.
4458 (add<mode>3): Likewise.
4459 (sub<mode>3): Likewise.
4460 (mul<mode>3): Likewise.
4461 (*aarch64_mul3_elt<mode>): Likewise.
4462 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4463 (*aarch64_mul3_elt_to_128df): Likewise.
4464 (*aarch64_mul3_elt_to_64v2df): Likewise.
4465 (neg<mode>2): Likewise.
4466 (abs<mode>2): Likewise.
4467 (abd<mode>_3): Likewise.
4468 (aba<mode>_3): Likewise.
4469 (fabd<mode>_3): Likewise.
4470 (*fabd_scalar<mode>3): Likewise.
4471 (and<mode>3): Likewise.
4472 (ior<mode>3): Likewise.
4473 (xor<mode>3): Likewise.
4474 (one_cmpl<mode>2): Likewise.
4475 (aarch64_simd_vec_set<mode>): Likewise.
4476 (aarch64_simd_lshr<mode>): Likewise.
4477 (aarch64_simd_ashr<mode>): Likewise.
4478 (aarch64_simd_imm_shl<mode>): Likewise.
4479 (aarch64_simd_reg_sshl<mode): Likewise.
4480 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4481 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4482 (aarch64_simd_vec_setv2di): Likewise.
4483 (aarch64_simd_vec_set<mode>): Likewise.
4484 (aarch64_mla<mode>): Likewise.
4485 (*aarch64_mla_elt<mode>): Likewise.
4486 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4487 (aarch64_mls<mode>): Likewise.
4488 (*aarch64_mls_elt<mode>): Likewise.
4489 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4490 (<su><maxmin><mode>3): Likewise.
4491 (move_lo_quad_<mode>): Likewise.
4492 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4493 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4494 (vec_pack_trunc_<mode>): Likewise.
4495 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4496 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4497 (*aarch64_<su>mlal_lo<mode>): Likewise.
4498 (*aarch64_<su>mlal_hi<mode>): Likewise.
4499 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4500 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4501 (*aarch64_<su>mlal<mode>): Likewise.
4502 (*aarch64_<su>mlsl<mode>): Likewise.
4503 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4504 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4505 (add<mode>3): Likewise.
4506 (sub<mode>3): Likewise.
4507 (mul<mode>3): Likewise.
4508 (div<mode>3): Likewise.
4509 (neg<mode>2): Likewise.
4510 (abs<mode>2): Likewise.
4511 (fma<mode>4): Likewise.
4512 (*aarch64_fma4_elt<mode>): Likewise.
4513 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4514 (*aarch64_fma4_elt_to_128df): Likewise.
4515 (*aarch64_fma4_elt_to_64v2df): Likewise.
4516 (fnma<mode>4): Likewise.
4517 (*aarch64_fnma4_elt<mode>): Likewise.
4518 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4519 (*aarch64_fnma4_elt_to_128df): Likewise.
4520 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4521 (<frint_pattern><mode>2): Likewise.
4522 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4523 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4524 (vec_unpacks_lo_v4sf): Likewise.
4525 (aarch64_float_extend_lo_v2df): Likewise.
4526 (vec_unpacks_hi_v4sf): Likewise.
4527 (aarch64_float_truncate_lo_v2sf): Likewise.
4528 (aarch64_float_truncate_hi_v4sf): Likewise.
4529 (aarch64_vmls<mode>): Likewise.
4530 (<su><maxmin><mode>3): Likewise.
4531 (<maxmin_uns><mode>3): Likewise.
4532 (reduc_<sur>plus_<mode>): Likewise.
4533 (reduc_<sur>plus_v2di): Likewise.
4534 (reduc_<sur>plus_v2si): Likewise.
4535 (reduc_<sur>plus_<mode>): Likewise.
4536 (aarch64_addpv4sf): Likewise.
4537 (clz<mode>2): Likewise.
4538 (reduc_<maxmin_uns>_<mode>): Likewise.
4539 (reduc_<maxmin_uns>_v2di): Likewise.
4540 (reduc_<maxmin_uns>_v2si): Likewise.
4541 (reduc_<maxmin_uns>_<mode>): Likewise.
4542 (reduc_<maxmin_uns>_v4sf): Likewise.
4543 (aarch64_simd_bsl<mode>_internal): Likewise.
4544 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4545 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4546 (aarch64_get_lane<mode>): Likewise.
4547 (*aarch64_combinez<mode>): Likewise.
4548 (aarch64_combine<mode>): Likewise.
4549 (aarch64_simd_combine<mode>): Likewise.
4550 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4551 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4552 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4553 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4554 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4555 (aarch64_<sur>h<addsub><mode>): Likewise.
4556 (aarch64_<sur><addsub>hn<mode>): Likewise.
4557 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4558 (aarch64_pmul<mode>): Likewise.
4559 (aarch64_<su_optab><optab><mode>): Likewise.
4560 (aarch64_<sur>qadd<mode>): Likewise.
4561 (aarch64_sqmovun<mode>): Likewise.
4562 (aarch64_<sur>qmovn<mode>): Likewise.
4563 (aarch64_s<optab><mode>): Likewise.
4564 (aarch64_sq<r>dmulh<mode>): Likewise.
4565 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4566 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4567 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4568 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4569 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4570 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4571 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4572 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4573 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4574 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4575 (aarch64_sqdmull<mode>): Likewise.
4576 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4577 (aarch64_sqdmull_n<mode>): Likewise.
4578 (aarch64_sqdmull2<mode>_internal): Likewise.
4579 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4580 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4581 (aarch64_<sur>shl<mode>): Likewise.
4582 (aarch64_<sur>q<r>shl<mode>
4583 (aarch64_<sur>shll_n<mode>): Likewise.
4584 (aarch64_<sur>shll2_n<mode>): Likewise.
4585 (aarch64_<sur>shr_n<mode>): Likewise.
4586 (aarch64_<sur>sra_n<mode>): Likewise.
4587 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4588 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4589 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4590 (aarch64_cm<optab><mode>): Likewise.
4591 (aarch64_cm<optab>di): Likewise.
4592 (aarch64_cm<optab><mode>): Likewise.
4593 (aarch64_cm<optab>di): Likewise.
4594 (aarch64_cmtst<mode>): Likewise.
4595 (aarch64_cmtstdi): Likewise.
4596 (aarch64_cm<optab><mode>): Likewise.
4597 (*aarch64_fac<optab><mode>): Likewise.
4598 (aarch64_addp<mode>): Likewise.
4599 (aarch64_addpdi): Likewise.
4600 (sqrt<mode>2): Likewise.
4601 (vec_load_lanesoi<mode>): Likewise.
4602 (vec_store_lanesoi<mode>): Likewise.
4603 (vec_load_lanesci<mode>): Likewise.
4604 (vec_store_lanesci<mode>): Likewise.
4605 (vec_load_lanesxi<mode>): Likewise.
4606 (vec_store_lanesxi<mode>): Likewise.
4607 (*aarch64_mov<mode>): Likewise.
4608 (aarch64_ld2<mode>_dreg): Likewise.
4609 (aarch64_ld2<mode>_dreg): Likewise.
4610 (aarch64_ld3<mode>_dreg): Likewise.
4611 (aarch64_ld3<mode>_dreg): Likewise.
4612 (aarch64_ld4<mode>_dreg): Likewise.
4613 (aarch64_ld4<mode>_dreg): Likewise.
4614 (aarch64_tbl1<mode>): Likewise.
4615 (aarch64_tbl2v16qi): Likewise.
4616 (aarch64_combinev16qi): Likewise.
4617 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4618 (aarch64_st2<mode>_dreg): Likewise.
4619 (aarch64_st2<mode>_dreg): Likewise.
4620 (aarch64_st3<mode>_dreg): Likewise.
4621 (aarch64_st3<mode>_dreg): Likewise.
4622 (aarch64_st4<mode>_dreg): Likewise.
4623 (aarch64_st4<mode>_dreg): Likewise.
4624 (*aarch64_simd_ld1r<mode>): Likewise.
4625 (aarch64_frecpe<mode>): Likewise.
4626 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4627 (aarch64_frecps<mode>): Likewise.
4629 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4631 * config/arm/iterators.md (V_elem_ch): New.
4633 (VQH_type): Likewise.
4634 * config/arm/arm.md (is_neon_type): New.
4635 (conds): Use is_neon_type.
4636 (anddi3_insn): Update type attribute.
4637 (xordi3_insn): Likewise.
4638 (one_cmpldi2): Likewise.
4639 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4640 * config/arm/neon.md (neon_mov): Update type attribute.
4641 (*movmisalign<mode>_neon_store): Likewise.
4642 (*movmisalign<mode>_neon_load): Likewise.
4643 (vec_set<mode>_internal): Likewise.
4644 (vec_set<mode>_internal): Likewise.
4645 (vec_setv2di_internal): Likewise.
4646 (vec_extract<mode>): Likewise.
4647 (vec_extract<mode>): Likewise.
4648 (vec_extractv2di): Likewise.
4649 (*add<mode>3_neon): Likewise.
4650 (adddi3_neon): Likewise.
4651 (*sub<mode>3_neon): Likewise.
4652 (subdi3_neon): Likewise.
4653 (fma<VCVTF:mode>4): Likewise.
4654 (fma<VCVTF:mode>4_intrinsic): Likewise.
4655 (*fmsub<VCVTF:mode>4): Likewise.
4656 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4657 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4658 (ior<mode>3): Likewise.
4659 (and<mode>3): Likewise.
4660 (orn<mode>3_neon): Likewise.
4661 (orndi3_neon): Likewise.
4662 (bic<mode>3_neon): Likewise.
4663 (bicdi3_neon): Likewise.
4664 (xor<mode>3): Likewise.
4665 (one_cmpl<mode>2): Likewise.
4666 (abs<mode>2): Likewise.
4667 (neg<mode>2): Likewise.
4668 (negdi2_neon): Likewise.
4669 (*umin<mode>3_neon): Likewise.
4670 (*umax<mode>3_neon): Likewise.
4671 (*smin<mode>3_neon): Likewise.
4672 (*smax<mode>3_neon): Likewise.
4673 (vashl<mode>3): Likewise.
4674 (vashr<mode>3_imm): Likewise.
4675 (vlshr<mode>3_imm): Likewise.
4676 (ashl<mode>3_signed): Likewise.
4677 (ashl<mode>3_unsigned): Likewise.
4678 (neon_load_count): Likewise.
4679 (ashldi3_neon_noclobber): Likewise.
4680 (ashldi3_neon): Likewise.
4681 (signed_shift_di3_neon): Likewise.
4682 (unsigned_shift_di3_neon): Likewise.
4683 (ashrdi3_neon_imm_noclobber): Likewise.
4684 (lshrdi3_neon_imm_noclobber): Likewise.
4685 (<shift>di3_neon): Likewise.
4686 (widen_ssum<mode>3): Likewise.
4687 (widen_usum<mode>3): Likewise.
4688 (quad_halves_<code>v4si): Likewise.
4689 (quad_halves_<code>v4sf): Likewise.
4690 (quad_halves_<code>v8hi): Likewise.
4691 (quad_halves_<code>v16qi): Likewise.
4692 (reduc_splus_v2di): Likewise.
4693 (neon_vpadd_internal<mode>): Likewise.
4694 (neon_vpsmin<mode>): Likewise.
4695 (neon_vpsmax<mode>): Likewise.
4696 (neon_vpumin<mode>): Likewise.
4697 (neon_vpumax<mode>): Likewise.
4698 (*ss_add<mode>_neon): Likewise.
4699 (*us_add<mode>_neon): Likewise.
4700 (*ss_sub<mode>_neon): Likewise.
4701 (*us_sub<mode>_neon): Likewise.
4702 (neon_vadd<mode>_unspec): Likewise.
4703 (neon_vaddl<mode>): Likewise.
4704 (neon_vaddw<mode>): Likewise.
4705 (neon_vhadd<mode>): Likewise.
4706 (neon_vqadd<mode>): Likewise.
4707 (neon_vaddhn<mode>): Likewise.
4708 (neon_vmul<mode>): Likewise.
4709 (neon_vfms<VCVTF:mode>): Likewise.
4710 (neon_vmlal<mode>): Likewise.
4711 (neon_vmls<mode>): Likewise.
4712 (neon_vmlsl<mode>): Likewise.
4713 (neon_vqdmulh<mode>): Likewise.
4714 (neon_vqdmlal<mode>): Likewise.
4715 (neon_vqdmlsl<mode>): Likewise.
4716 (neon_vmull<mode>): Likewise.
4717 (neon_vqdmull<mode>): Likewise.
4718 (neon_vsub<mode>_unspec): Likewise.
4719 (neon_vsubl<mode>): Likewise.
4720 (neon_vsubw<mode>): Likewise.
4721 (neon_vqsub<mode>): Likewise.
4722 (neon_vhsub<mode>): Likewise.
4723 (neon_vsubhn<mode>): Likewise.
4724 (neon_vceq<mode>): Likewise.
4725 (neon_vcge<mode>): Likewise.
4726 (neon_vcgeu<mode>): Likewise.
4727 (neon_vcgt<mode>): Likewise.
4728 (neon_vcgtu<mode>): Likewise.
4729 (neon_vcle<mode>): Likewise.
4730 (neon_vclt<mode>): Likewise.
4731 (neon_vcage<mode>): Likewise.
4732 (neon_vcagt<mode>): Likewise.
4733 (neon_vtst<mode>): Likewise.
4734 (neon_vabd<mode>): Likewise.
4735 (neon_vabdl<mode>): Likewise.
4736 (neon_vaba<mode>): Likewise.
4737 (neon_vabal<mode>): Likewise.
4738 (neon_vmax<mode>): Likewise.
4739 (neon_vmin<mode>): Likewise.
4740 (neon_vpaddl<mode>): Likewise.
4741 (neon_vpadal<mode>): Likewise.
4742 (neon_vpmax<mode>): Likewise.
4743 (neon_vpmin<mode>): Likewise.
4744 (neon_vrecps<mode>): Likewise.
4745 (neon_vrsqrts<mode>): Likewise.
4746 (neon_vqabs<mode>): Likewise.
4747 (neon_vqneg<mode>): Likewise.
4748 (neon_vcls<mode>): Likewise.
4749 (clz<mode>2): Likewise.
4750 (popcount<mode>2): Likewise.
4751 (neon_vrecpe<mode>): Likewise.
4752 (neon_vrsqrte<mode>): Likewise.
4753 (neon_vget_lane<mode>_sext_internal): Likewise.
4754 (neon_vget_lane<mode>_zext_internal): Likewise.
4755 (neon_vdup_n<mode>): Likewise.
4756 (neon_vdup_n<mode>): Likewise.
4757 (neon_vdup_nv2di): Likewise.
4758 (neon_vdup_lane<mode>_interal): Likewise.
4759 (*neon_vswp<mode>): Likewise.
4760 (neon_vcombine<mode>): Likewise.
4761 (float<mode><V_cvtto>2): Likewise.
4762 (floatuns<mode><V_cvtto>2): Likewise.
4763 (fix_trunc<mode><V_cvtto>2): Likewise.
4764 (fixuns_trunc<mode><V_cvtto>2
4765 (neon_vcvt<mode>): Likewise.
4766 (neon_vcvt<mode>): Likewise.
4767 (neon_vcvtv4sfv4hf): Likewise.
4768 (neon_vcvtv4hfv4sf): Likewise.
4769 (neon_vcvt_n<mode>): Likewise.
4770 (neon_vcvt_n<mode>): Likewise.
4771 (neon_vmovn<mode>): Likewise.
4772 (neon_vqmovn<mode>): Likewise.
4773 (neon_vqmovun<mode>): Likewise.
4774 (neon_vmovl<mode>): Likewise.
4775 (neon_vmul_lane<mode>): Likewise.
4776 (neon_vmul_lane<mode>): Likewise.
4777 (neon_vmull_lane<mode>): Likewise.
4778 (neon_vqdmull_lane<mode>): Likewise.
4779 (neon_vqdmulh_lane<mode>): Likewise.
4780 (neon_vqdmulh_lane<mode>): Likewise.
4781 (neon_vmla_lane<mode>): Likewise.
4782 (neon_vmla_lane<mode>): Likewise.
4783 (neon_vmlal_lane<mode>): Likewise.
4784 (neon_vqdmlal_lane<mode>): Likewise.
4785 (neon_vmls_lane<mode>): Likewise.
4786 (neon_vmls_lane<mode>): Likewise.
4787 (neon_vmlsl_lane<mode>): Likewise.
4788 (neon_vqdmlsl_lane<mode>): Likewise.
4789 (neon_vext<mode>): Likewise.
4790 (neon_vrev64<mode>): Likewise.
4791 (neon_vrev32<mode>): Likewise.
4792 (neon_vrev16<mode>): Likewise.
4793 (neon_vbsl<mode>_internal): Likewise.
4794 (neon_vshl<mode>): Likewise.
4795 (neon_vqshl<mode>): Likewise.
4796 (neon_vshr_n<mode>): Likewise.
4797 (neon_vshrn_n<mode>): Likewise.
4798 (neon_vqshrn_n<mode>): Likewise.
4799 (neon_vqshrun_n<mode>): Likewise.
4800 (neon_vshl_n<mode>): Likewise.
4801 (neon_vqshl_n<mode>): Likewise.
4802 (neon_vqshlu_n<mode>): Likewise.
4803 (neon_vshll_n<mode>): Likewise.
4804 (neon_vsra_n<mode>): Likewise.
4805 (neon_vsri_n<mode>): Likewise.
4806 (neon_vsli_n<mode>): Likewise.
4807 (neon_vtbl1v8qi): Likewise.
4808 (neon_vtbl2v8qi): Likewise.
4809 (neon_vtbl3v8qi): Likewise.
4810 (neon_vtbl4v8qi): Likewise.
4811 (neon_vtbl1v16qi): Likewise.
4812 (neon_vtbl2v16qi): Likewise.
4813 (neon_vcombinev16qi): Likewise.
4814 (neon_vtbx1v8qi): Likewise.
4815 (neon_vtbx2v8qi): Likewise.
4816 (neon_vtbx3v8qi): Likewise.
4817 (neon_vtbx4v8qi): Likewise.
4818 (*neon_vtrn<mode>_insn): Likewise.
4819 (*neon_vzip<mode>_insn): Likewise.
4820 (*neon_vuzp<mode>_insn): Likewise.
4821 (neon_vld1<mode>): Likewise.
4822 (neon_vld1_lane<mode>): Likewise.
4823 (neon_vld1_lane<mode>): Likewise.
4824 (neon_vld1_dup<mode>): Likewise.
4825 (neon_vld1_dup<mode>): Likewise.
4826 (neon_vld1_dupv2di): Likewise.
4827 (neon_vst1<mode>): Likewise.
4828 (neon_vst1_lane<mode>): Likewise.
4829 (neon_vst1_lane<mode>): Likewise.
4830 (neon_vld2<mode>): Likewise.
4831 (neon_vld2<mode>): Likewise.
4832 (neon_vld2_lane<mode>): Likewise.
4833 (neon_vld2_lane<mode>): Likewise.
4834 (neon_vld2_dup<mode>): Likewise.
4835 (neon_vst2<mode>): Likewise.
4836 (neon_vst2<mode>): Likewise.
4837 (neon_vst2_lane<mode>): Likewise.
4838 (neon_vst2_lane<mode>): Likewise.
4839 (neon_vld3<mode>): Likewise.
4840 (neon_vld3qa<mode>): Likewise.
4841 (neon_vld3qb<mode>): Likewise.
4842 (neon_vld3_lane<mode>): Likewise.
4843 (neon_vld3_lane<mode>): Likewise.
4844 (neon_vld3_dup<mode>): Likewise.
4845 (neon_vst3<mode>): Likewise.
4846 (neon_vst3qa<mode>): Likewise.
4847 (neon_vst3qb<mode>): Likewise.
4848 (neon_vst3_lane<mode>): Likewise.
4849 (neon_vst3_lane<mode>): Likewise.
4850 (neon_vld4<mode>): Likewise.
4851 (neon_vld4qa<mode>): Likewise.
4852 (neon_vld4qb<mode>): Likewise.
4853 (neon_vld4_lane<mode>): Likewise.
4854 (neon_vld4_lane<mode>): Likewise.
4855 (neon_vld4_dup<mode>): Likewise.
4856 (neon_vst4<mode>): Likewise.
4857 (neon_vst4qa<mode>): Likewise.
4858 (neon_vst4qb<mode>): Likewise.
4859 (neon_vst4_lane<mode>): Likewise.
4860 (neon_vst4_lane<mode>): Likewise.
4861 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4862 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4863 (neon_vec_<US>mult_lo_<mode>): Likewise.
4864 (neon_vec_<US>mult_hi_<mode>): Likewise.
4865 (neon_vec_<US>shiftl_<mode>): Likewise.
4866 (neon_unpack<US>_<mode>): Likewise.
4867 (neon_vec_<US>mult_<mode>): Likewise.
4868 (vec_pack_trunc_<mode>): Likewise.
4869 (neon_vec_pack_trunc_<mode>): Likewise.
4870 (neon_vabd<mode>_2): Likewise.
4871 (neon_vabd<mode>_3): Likewise.
4873 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4875 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4876 (load_pair): Update type attribute.
4877 (store_pair): Update type attribute.
4878 * config/aarch64/iterators.md (q): New.
4880 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4882 * config/arm/types.md: Add new types for Neon insns.
4884 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4885 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4886 Sergey Lega <sergey.s.lega@intel.com>
4887 Anna Tikhonova <anna.tikhonova@intel.com>
4888 Ilya Tocar <ilya.tocar@intel.com>
4889 Andrey Turetskiy <andrey.turetskiy@intel.com>
4890 Ilya Verbin <ilya.verbin@intel.com>
4891 Kirill Yukhin <kirill.yukhin@intel.com>
4892 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4894 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4895 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4896 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4898 (srcp14<mode>): Ditto.
4899 (rsqrt14<mode>): Ditto.
4900 (rsqrt14<mode>): Ditto.
4901 (avx512f_vmscalef<mode>): Ditto.
4902 (avx512f_scalef<mode>): Ditto.
4903 (avx512f_getexp<mode>): Ditto.
4904 (avx512f_sgetexp<mode>): Ditto.
4905 (avx512f_fixupimm<mode>): Ditto.
4906 (avx512f_sfixupimm<mode>): Ditto.
4907 (avx512f_rndscale<mode>): Ditto.
4908 (*avx512er_exp2<mode>): Ditto.
4909 (*avx512er_rcp28<mode>): Ditto.
4910 (avx512er_rsqrt28<mode>): Ditto.
4911 (avx512f_getmant<mode>): Ditto.
4912 (avx512f_getmant<mode>): Ditto.
4913 (avx512f_rndscale<mode>): Fix formatting.
4915 2013-10-15 Martin Jambor <mjambor@suse.cz>
4917 * ipa-utils.h (ipa_edge_within_scc): Declare.
4918 * ipa-cp.c (edge_within_scc): Moved...
4919 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4921 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4922 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4923 Sergey Lega <sergey.s.lega@intel.com>
4924 Anna Tikhonova <anna.tikhonova@intel.com>
4925 Ilya Tocar <ilya.tocar@intel.com>
4926 Andrey Turetskiy <andrey.turetskiy@intel.com>
4927 Ilya Verbin <ilya.verbin@intel.com>
4928 Kirill Yukhin <kirill.yukhin@intel.com>
4929 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4931 * config/i386/predicates.md (const_8_to_15_operand): New.
4932 (const_16_to_31_operand): Ditto.
4933 * config/i386/sse.md (V8FI): New.
4935 (reduc_splus_v8df): Ditto.
4936 (reduc_splus_v16sf): Ditto.
4937 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4938 (vec_extract_hi_<mode>): Ditto.
4939 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4940 (vec_set_lo_<mode>): Ditto.
4941 (vec_set_hi_<mode>): Ditto.
4942 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4943 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4944 (avx512f_pshufd_1): Ditto.
4945 (avx512f_broadcast<mode>): Ditto.
4946 (avx512f_broadcast<mode>): Ditto.
4947 (define_split): Split vec_extract_lo into move.
4948 (ssequartermode): Ditto.
4949 (ssedoublemode): Extened with wider modes.
4950 (vec_extract_lo_<mode>): Ditto.
4952 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4953 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4954 Sergey Lega <sergey.s.lega@intel.com>
4955 Anna Tikhonova <anna.tikhonova@intel.com>
4956 Ilya Tocar <ilya.tocar@intel.com>
4957 Andrey Turetskiy <andrey.turetskiy@intel.com>
4958 Ilya Verbin <ilya.verbin@intel.com>
4959 Kirill Yukhin <kirill.yukhin@intel.com>
4960 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4962 * config/i386/predicates.md (register_or_constm1_operand): New.
4963 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4964 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4965 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4966 UNSPEC_SCATTER_PREFETCH
4968 (avx512f_ucmp<mode>3): Ditto.
4969 (avx512f_vternlog<mode>): Ditto.
4970 (avx512f_align<mode>): Ditto.
4971 (<shift_insn><mode>3): Ditto.
4972 (avx512f_<rotate>v<mode>): Ditto.
4973 (avx512f_<rotate><mode>): Ditto.
4974 (avx512f_eq<mode>3): Ditto.
4975 (avx512f_eq<mode>3_1): Ditto.
4976 (avx512f_gt<mode>3): Ditto.
4977 (avx512f_testm<mode>3): Ditto.
4978 (avx512f_testnm<mode>3): Ditto.
4979 (avx512pf_gatherpf<mode>): Ditto.
4980 (*avx512pf_gatherpf<mode>_mask): Ditto.
4981 (*avx512pf_gatherpf<mode>): Ditto.
4982 (avx512pf_scatterpf<mode>): Ditto.
4983 (*avx512pf_scatterpf<mode>_mask): Ditto.
4984 (*avx512pf_scatterpf<mode>): Ditto.
4985 (avx512f_vec_dup_gpr<mode>): Ditto.
4986 (clz<mode>2): Ditto.
4987 (conflict<mode>): Ditto.
4988 (REDUC_SMINMAX_MODE): Extened with wider modes.
4989 (reduc_<code>_<mode>): Ditto.
4990 (vlshr<mode>3): Ditto.
4991 (vashl<mode>3): Ditto.
4993 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4994 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4995 Sergey Lega <sergey.s.lega@intel.com>
4996 Anna Tikhonova <anna.tikhonova@intel.com>
4997 Ilya Tocar <ilya.tocar@intel.com>
4998 Andrey Turetskiy <andrey.turetskiy@intel.com>
4999 Ilya Verbin <ilya.verbin@intel.com>
5000 Kirill Yukhin <kirill.yukhin@intel.com>
5001 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5003 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
5006 (avx512fmaskmode): Ditto.
5007 (bcstscalarsuff): Ditto.
5008 (avx512f_blendm<mode>): Ditto.
5009 (cmp_imm_predicate): Ditto.
5010 (avx512f_cmp<mode>3): Ditto.
5011 (avx512f_vec_dup<mode>): Ditto.
5012 (avx512f_vec_dup_mem<mode>): Ditto.
5013 (avx512f_vpermi2var<mode>3): Ditto.
5014 (avx512f_vpermt2var<mode>3): Ditto.
5015 (vec_init<mode>): Ditto.
5016 (avx512f_gathersi<mode>): Ditto.
5017 (*avx512f_gathersi<mode>): Ditto.
5018 (*avx512f_gathersi<mode>_2): Ditto.
5019 (avx512f_gatherdi<mode>): Ditto.
5020 (*avx512f_gatherdi<mode>): Ditto.
5021 (*avx512f_gatherdi<mode>_2): Ditto.
5022 (avx512f_scattersi<mode>): Ditto.
5023 (*avx512f_scattersi<mode>): Ditto.
5024 (avx512f_scatterdi<mode>): Ditto.
5025 (*avx512f_scatterdi<mode>): Ditto.
5026 (sseintprefix): Extened with wider modes.
5027 (VEC_GATHER_IDXSI): Ditto.
5028 (VEC_GATHER_IDXDI): Ditto.
5029 (VEC_GATHER_SRCDI): Ditto.
5031 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
5032 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5034 * config/arm/t-aprofile: New file.
5035 * config.gcc: Handle --with-multilib-list option.
5037 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
5039 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
5042 2013-10-15 Richard Biener <rguenther@suse.de>
5044 * tree-loop-distribution.c (build_empty_rdg): Inline into
5046 (rdg_flag_vertex): Inline into single user.
5047 (rdg_flag_vertex_and_dependent): Likewise.
5048 (build_rdg_partition_for_vertex): Remove processed bitmap.
5049 (rdg_build_partitions): Simplify.
5051 2013-10-15 Richard Biener <rguenther@suse.de>
5053 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
5054 Restructure forwarding through conversions and copies to
5055 avoid performing copy-propagation the wrong way. Adjust
5056 recursion invocations.
5057 (forward_propagate_addr_expr): Add argument stating if we
5058 are recursing from a single-use.
5059 (ssa_forward_propagate_and_combine): Adjust.
5061 2013-10-14 David Malcolm <dmalcolm@redhat.com>
5063 * dumpfile.h (gcc::dump_manager): New class, to hold state
5064 relating to dumpfile management.
5065 (get_dump_file_name): Remove in favor of method of dump_manager.
5066 (dump_initialized_p): Likewise.
5067 (dump_start): Likewise.
5068 (dump_finish): Likewise.
5069 (dump_switch_p): Likewise.
5070 (dump_register): Likewise.
5071 (get_dump_file_info): Likewise.
5072 * context.c (gcc::context::context): Construct the dump_manager
5074 * context.h (gcc::context::get_dumps): New.
5075 (gcc::context::m_dumps): New.
5076 * coverage.c (coverage_init): Port to dump_manager API.
5077 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
5078 (extra_dump_files_in_use): Likewise.
5079 (extra_dump_files_alloced): Likewise.
5080 (gcc::dump_manager::dump_manager): New.
5081 (dump_register): Convert to...
5082 (gcc::dump_manager::dump_register): ...method, replacing
5083 function-static next_dump with m_next_dump field.
5084 (get_dump_file_info): Convert to...
5085 (gcc::dump_manager::get_dump_file_info): ...method.
5086 (get_dump_file_name): Convert to...
5087 (gcc::dump_manager::get_dump_file_name): ...method.
5088 (dump_start): Convert to...
5089 (gcc::dump_manager::dump_start): ...method.
5090 (dump_finish): Convert to...
5091 (gcc::dump_manager::dump_finish): ...method.
5092 (dump_begin): Replace body with...
5093 (gcc::dump_manager::dump_begin): ...new method.
5094 (dump_phase_enabled_p): Convert to...
5095 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5096 (dump_phase_enabled_p): Convert to...
5097 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5098 (dump_initialized_p): Convert to...
5099 (gcc::dump_manager::dump_initialized_p): ...method.
5100 (dump_flag_name): Replace body with...
5101 (gcc::dump_manager::dump_flag_name): ...new method.
5102 (dump_enable_all): Convert to...
5103 (gcc::dump_manager::dump_enable_all): ...new method.
5104 (opt_info_enable_passes): Convert to...
5105 (gcc::dump_manager::opt_info_enable_passes): ...new method.
5106 (dump_switch_p_1): Convert to...
5107 (gcc::dump_manager::dump_switch_p_1): ...new method.
5108 (dump_switch_p): Convert to...
5109 (gcc::dump_manager::dump_switch_p): ...new method.
5110 (opt_info_switch_p): Port to dump_manager API.
5111 (enable_rtl_dump_file): Likewise.
5112 * opts-global.c (handle_common_deferred_options): Port to new
5114 * passes.c (pass_manager::finish_optimization_passes): Likewise.
5115 (pass_manager::register_one_dump_file): Likewise.
5116 (pass_manager::register_pass): Likewise.
5117 (pass_init_dump_file): Likewise.
5118 (pass_fini_dump_file): Likewise.
5119 * statistics.c (statistics_early_init): Likewise.
5121 2013-10-14 Richard Biener <rguenther@suse.de>
5123 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
5124 iterative_hash_canonical_type, gimple_canonical_type_hash,
5125 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
5126 gimple_register_canonical_type, print_gimple_types_stats,
5127 free_gimple_type_tables): Move to lto/lto.c
5128 (gt-gimple.h): Do not include.
5129 * gimple.h (gimple_register_canonical_type,
5130 print_gimple_types_stats, free_gimple_type_tables): Remove.
5131 * Makefile.in (GTFILES): Remove gimple.c.
5133 2013-10-14 Travis Snoozy <quandary@remstate.com>
5136 * config/msp430/msp430.c (msp430_option_override): Correct thinko
5137 scanning for msp430x targets.
5139 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
5142 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
5143 (registers_ok_for_ldd_peep): Move around.
5144 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
5145 * config/sparc/sparc.md (widening peepholes): Use it.
5147 2013-10-14 Richard Biener <rguenther@suse.de>
5151 * gimple.c (iterative_hash_canonical_type): Make sure to
5152 record the hash into the correct hashtable slot.
5154 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
5156 PR rtl-optimization/58662
5157 * combine.c (try_combine): Take into account death nodes on I2 when
5158 splitting a PARALLEL of two independent SETs. Fix dump message.
5160 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
5163 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
5164 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
5165 * config/sh/t-sh (sh_treg_combine.o): New entry.
5166 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
5167 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
5168 (register_sh_passes): New function. Register sh_treg_combine pass.
5169 (sh_option_override): Invoke it.
5170 (sh_canonicalize_comparison): Handle op0_preserve_value.
5171 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
5172 opportunities. Canonicalize branch condition.
5173 (nott): Allow only if pseudos can be created for non-SH2A.
5175 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
5178 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
5179 (ix86_expand_movmem): Replace copy_addr_to_reg with
5180 ix86_copy_addr_to_reg.
5181 (ix86_expand_setmem): Likewise.
5183 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
5185 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
5186 provided by ix86_fp_compare_mode instead of CCFPUmode.
5188 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
5190 * config/aarch64/arm_neon.h
5191 (vtbx<1,3>_<psu>8): Fix register constriants.
5193 2013-10-11 Jeff Law <law@redhat.com>
5195 PR tree-optimization/58640
5196 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
5197 threading paths that cross over two loop entry points.
5199 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5201 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
5202 handle vector float as well.
5203 (*vsx_le_perm_load_v4si): Likewise.
5204 (*vsx_le_perm_store_v2di): Likewise.
5205 (*vsx_le_perm_store_v4si): Likewise.
5207 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5209 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
5210 directly to circumvent subtract from splat{31} workaround.
5211 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
5213 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
5214 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
5215 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
5216 (altivec_vperm_<mode>): Convert to define_insn_and_split to
5217 separate big and little endian logic.
5218 (*altivec_vperm_<mode>_internal): New define_insn.
5219 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
5220 separate big and little endian logic.
5221 (*altivec_vperm_<mode>_uns_internal): New define_insn.
5222 (vec_permv16qi): Add little endian logic.
5224 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
5226 * doc/extend.texi (returns_nonnull): Remove arguments.
5228 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5229 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5230 Sergey Lega <sergey.s.lega@intel.com>
5231 Anna Tikhonova <anna.tikhonova@intel.com>
5232 Ilya Tocar <ilya.tocar@intel.com>
5233 Andrey Turetskiy <andrey.turetskiy@intel.com>
5234 Ilya Verbin <ilya.verbin@intel.com>
5235 Kirill Yukhin <kirill.yukhin@intel.com>
5236 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5238 * config/i386/sse.md (VI48F_256_512): New.
5239 (avx2_permvar<mode>): Change to ...
5240 (<avx2_avx512f>_permvar<mode>): This.
5242 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5243 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5244 Sergey Lega <sergey.s.lega@intel.com>
5245 Anna Tikhonova <anna.tikhonova@intel.com>
5246 Ilya Tocar <ilya.tocar@intel.com>
5247 Andrey Turetskiy <andrey.turetskiy@intel.com>
5248 Ilya Verbin <ilya.verbin@intel.com>
5249 Kirill Yukhin <kirill.yukhin@intel.com>
5250 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5252 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
5253 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
5254 * config/i386/sse.md (VI4_AVX): New.
5255 (sf2simodelower): Ditto.
5256 (sse2_cvtps2dq): Change to ...
5257 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
5259 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5260 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5261 Sergey Lega <sergey.s.lega@intel.com>
5262 Anna Tikhonova <anna.tikhonova@intel.com>
5263 Ilya Tocar <ilya.tocar@intel.com>
5264 Andrey Turetskiy <andrey.turetskiy@intel.com>
5265 Ilya Verbin <ilya.verbin@intel.com>
5266 Kirill Yukhin <kirill.yukhin@intel.com>
5267 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5269 * config/i386/sse.md (V_512): New.
5271 (vcond<V_512:mode><VF_512:mode>): Ditto.
5272 (vcond<V_512:mode><VI_512:mode>): Ditto.
5273 (vcondu<V_512:mode><VI_512:mode>): Ditto.
5275 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5276 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5277 Sergey Lega <sergey.s.lega@intel.com>
5278 Anna Tikhonova <anna.tikhonova@intel.com>
5279 Ilya Tocar <ilya.tocar@intel.com>
5280 Andrey Turetskiy <andrey.turetskiy@intel.com>
5281 Ilya Verbin <ilya.verbin@intel.com>
5282 Kirill Yukhin <kirill.yukhin@intel.com>
5283 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5285 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
5286 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
5288 (fma<mode>4): Removed condition.
5289 (fms<mode>4): Ditto.
5290 (fnma<mode>4): Ditto.
5291 (fnms<mode>4): Ditto.
5292 (fma4i_fmadd_<mode>): Ditto.
5293 (*fma_fmadd_<mode>): Ditto.
5294 (*fma_fmsub_<mode>): Ditto.
5295 (*fma_fnmadd_<mode>): Ditto.
5296 (*fma_fnmsub_<mode>): Ditto.
5297 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
5298 (*fma_fmaddsub_<mode>): Ditto.
5299 (*fma_fmsubadd_<mode>): Ditto.
5300 (*fmai_fmadd_<mode>): Ditto.
5301 (*fmai_fmsub_<mode>): Ditto.
5302 (*fmai_fnmadd_<mode>): Ditto.
5303 (*fmai_fnmsub_<mode>): Ditto.
5305 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5306 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5307 Sergey Lega <sergey.s.lega@intel.com>
5308 Anna Tikhonova <anna.tikhonova@intel.com>
5309 Ilya Tocar <ilya.tocar@intel.com>
5310 Andrey Turetskiy <andrey.turetskiy@intel.com>
5311 Ilya Verbin <ilya.verbin@intel.com>
5312 Kirill Yukhin <kirill.yukhin@intel.com>
5313 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5315 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5316 (VI124_256): Changed to ...
5317 (VI124_256_48_512): This.
5318 (ssepackmode): Extended with wider modes.
5319 (<code><mode>3): Changed iterator.
5320 (*avx2_<code><mode>3): Ditto.
5321 (vec_pack_trunc_<mode>): Ditto.
5323 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5324 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5325 Sergey Lega <sergey.s.lega@intel.com>
5326 Anna Tikhonova <anna.tikhonova@intel.com>
5327 Ilya Tocar <ilya.tocar@intel.com>
5328 Andrey Turetskiy <andrey.turetskiy@intel.com>
5329 Ilya Verbin <ilya.verbin@intel.com>
5330 Kirill Yukhin <kirill.yukhin@intel.com>
5331 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5333 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5334 (VI8F_256_512): Ditto.
5335 (abs<mode>2): Changed iterator.
5336 (avx2_perm<mode>): Changed to ...
5337 (<avx2_avx512f>_perm<mode>): This.
5338 (avx2_perm<mode>_1): Changed to ...
5339 (<avx2_avx512f>_perm<mode>_1): This.
5341 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5342 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5343 Sergey Lega <sergey.s.lega@intel.com>
5344 Anna Tikhonova <anna.tikhonova@intel.com>
5345 Ilya Tocar <ilya.tocar@intel.com>
5346 Andrey Turetskiy <andrey.turetskiy@intel.com>
5347 Ilya Verbin <ilya.verbin@intel.com>
5348 Kirill Yukhin <kirill.yukhin@intel.com>
5349 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5351 * config/i386/sse.md (VI48_AVX512F): New.
5352 (VI48_AVX2): Changed to ...
5353 (VI48_AVX2_48_AVX512F): This.
5354 (avx2_ashrv<mode>): Changed to ...
5355 (<avx2_avx512f>_ashrv<mode>): This.
5356 (avx2_<shift_insn>v<mode>): Changed to ...
5357 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5359 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5360 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5361 Sergey Lega <sergey.s.lega@intel.com>
5362 Anna Tikhonova <anna.tikhonova@intel.com>
5363 Ilya Tocar <ilya.tocar@intel.com>
5364 Andrey Turetskiy <andrey.turetskiy@intel.com>
5365 Ilya Verbin <ilya.verbin@intel.com>
5366 Kirill Yukhin <kirill.yukhin@intel.com>
5367 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5369 * config/i386/sse.md (VI4_AVX512F): New.
5370 (VI8_AVX2_AVX512F): Ditto.
5371 (mul<mode>3): Extended with wider modes.
5372 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5373 (mul<mode>3): Ditto.
5374 (vec_widen_<s>mult_odd_<mode>): Ditto.
5376 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5377 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5378 Sergey Lega <sergey.s.lega@intel.com>
5379 Anna Tikhonova <anna.tikhonova@intel.com>
5380 Ilya Tocar <ilya.tocar@intel.com>
5381 Andrey Turetskiy <andrey.turetskiy@intel.com>
5382 Ilya Verbin <ilya.verbin@intel.com>
5383 Kirill Yukhin <kirill.yukhin@intel.com>
5384 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5386 * config/i386/sse.md (VI2_AVX512F): New.
5387 (VI124_AVX512F): Ditto.
5388 (sseunpackmode): Extended with wider modes.
5389 (sseunpackfltmode): Ditto.
5390 (vec_unpacks_float_hi_<mode>): Ditto.
5391 (vec_unpacks_float_lo_<mode>): Ditto.
5392 (vec_unpacku_float_hi_<mode>): Ditto.
5393 (vec_unpacku_float_lo_<mode>): Ditto.
5394 (vec_unpacks_lo_<mode>): Ditto.
5395 (vec_unpacks_hi_<mode>): Ditto.
5396 (vec_unpacku_lo_<mode>): Ditto.
5397 (vec_unpacku_hi_<mode>): Ditto.
5399 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5400 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5401 Sergey Lega <sergey.s.lega@intel.com>
5402 Anna Tikhonova <anna.tikhonova@intel.com>
5403 Ilya Tocar <ilya.tocar@intel.com>
5404 Andrey Turetskiy <andrey.turetskiy@intel.com>
5405 Ilya Verbin <ilya.verbin@intel.com>
5406 Kirill Yukhin <kirill.yukhin@intel.com>
5407 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5409 * config/i386/i386.md (multdiv): New.
5410 (multdiv_mnemonic): Ditto.
5411 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5412 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5413 (<sse>_vmdiv<mode>3): Removed.
5415 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5416 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5417 Sergey Lega <sergey.s.lega@intel.com>
5418 Anna Tikhonova <anna.tikhonova@intel.com>
5419 Ilya Tocar <ilya.tocar@intel.com>
5420 Andrey Turetskiy <andrey.turetskiy@intel.com>
5421 Ilya Verbin <ilya.verbin@intel.com>
5422 Kirill Yukhin <kirill.yukhin@intel.com>
5423 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5425 * config/i386/sse.md (V): Extended with wider modes.
5427 (ssehalfvecmode): Ditto.
5429 (ssepackfltmode): Ditto.
5430 (avx_vec_concat<mode>): Ditto.
5431 (V_256_512): New iterator.
5432 (VF2_512_256): Ditto.
5433 (si2dfmode): New attribute.
5434 (si2dfmodelower): Ditto.
5436 (concat_tg_mode): Ditto.
5437 (floatv4siv4df2): Changed to ...
5438 (float<si2dfmodelower><mode>2): This.
5439 (avx_cvtps2pd256): Changed to ...
5440 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5441 (vec_pack_trunc_v4df): Changed to ...
5442 (vec_pack_trunc_<mode>): This.
5443 (avx_vpermil<mode>): Changed to ...
5444 (<sse2_avx_avx512f>_vpermil<mode>): This.
5445 (<fixsuffix>fix_truncv8dfv8si2): New.
5446 (vec_pack_sfix_trunc_v8df): Ditto.
5447 (avx512f_rndscale<mode>): Ditto.
5448 (avx512f_roundpd512): Ditto.
5449 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5451 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5452 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5453 Sergey Lega <sergey.s.lega@intel.com>
5454 Anna Tikhonova <anna.tikhonova@intel.com>
5455 Ilya Tocar <ilya.tocar@intel.com>
5456 Andrey Turetskiy <andrey.turetskiy@intel.com>
5457 Ilya Verbin <ilya.verbin@intel.com>
5458 Kirill Yukhin <kirill.yukhin@intel.com>
5459 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5461 * config/i386/i386.md (any_fix): New iterator.
5462 (fixsuffix): New attribute.
5463 * config/i386/sse.md (VF1): Extened with wider modes.
5467 (sseintvecmodelower): Ditto.
5468 (ssescalarmode): Ditto.
5469 (ssescalarnum): Ditto.
5471 (ssexmmmode): Ditto.
5472 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5473 (<sse>_rcp<mode>2): Change iterator.
5474 (rsqrt<mode>2): Ditto.
5475 (<sse>_rsqrt<mode>2): Ditto.
5476 (avx2_vec_dup<mode>): Ditto.
5477 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5478 (round<mode>2_sfix): Ditto.
5479 (avx2_pbroadcast<mode>): Ditto.
5480 (*andnot<mode>3): Handle XI mode.
5481 (*<code><mode>3): Ditto.
5482 (AVXTOSSEMODE): Removed.
5483 (avx_vpermil<mode>): Changed to ...
5484 (<sse2_avx_avx512f>_vpermil<mode>): This.
5486 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5487 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5488 Sergey Lega <sergey.s.lega@intel.com>
5489 Anna Tikhonova <anna.tikhonova@intel.com>
5490 Ilya Tocar <ilya.tocar@intel.com>
5491 Andrey Turetskiy <andrey.turetskiy@intel.com>
5492 Ilya Verbin <ilya.verbin@intel.com>
5493 Kirill Yukhin <kirill.yukhin@intel.com>
5494 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5496 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5497 (<sse>_comi): Ditto.
5498 (<sse>_ucomi): Ditto.
5499 (sse_cvtss2siq_2): Ditto.
5500 (sse2_cvtsd2si): Ditto.
5501 (sse2_cvtsd2siq): Ditto.
5502 (sse2_cvttsd2si): Ditto.
5503 (sse2_cvttsd2siq): Ditto.
5504 (<shift_insn><mode>3): Ditto.
5505 (sse2_cvtsi2sdq): Update constraint and prefix.
5506 (sse_cvtsi2ss): Update prefix.
5507 (sse_cvtsi2ssq): Ditto.
5509 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5511 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5512 ignore internal calls.
5514 2013-10-11 Richard Biener <rguenther@suse.de>
5516 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5517 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5518 TREE_OVERFLOW is set.
5520 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5522 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5524 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5525 * gimple.h: Likewise.
5527 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5529 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5532 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5534 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5535 * configure: Regenerate.
5537 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5539 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5540 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5541 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5542 similarly pp_right_paren.
5543 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5544 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5545 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5546 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5548 (omp_declare_simd_clauses_equal,
5549 omp_remove_redundant_declare_simd_attrs): New functions.
5550 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5551 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5552 * tree.h (OMP_LOOP_CHECK): Define.
5553 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5554 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5555 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5556 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5557 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5558 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5559 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5560 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5561 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5562 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5563 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5564 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5565 (omp_remove_redundant_declare_simd_attrs): New prototype.
5566 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5567 GIMPLE_OMP_TEAMS): New codes.
5568 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5569 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5571 (target_nesting_level): New variable.
5572 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5573 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5574 for collapse > 1 static schedule unless ordered.
5575 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5576 (determine_parallel_type): Adjust get_ws_args_for caller.
5577 (install_var_field): Handle mask & 4 for double indirection.
5578 (scan_sharing_clauses): Ignore shared clause on teams construct.
5579 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5580 (create_omp_child_function): If inside target or declare target
5581 constructs, set "omp declare target" attribute on the child function.
5582 (find_combined_for): New function.
5583 (scan_omp_parallel): Handle combined loops.
5584 (scan_omp_target, scan_omp_teams): New functions.
5585 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5586 restrictions and set ctx->cancellable for cancellable constructs.
5587 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5588 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5589 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5590 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5591 (omp_clause_aligned_alignment): New function.
5592 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5593 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5594 on teams constructs. Handle user defined reductions and new
5596 (lower_reduction_clauses): Don't set placeholder to address of ref
5597 if it has already the right type.
5598 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5599 (expand_parallel_call): Use the new non-_start suffixed builtins,
5600 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5601 and GOMP_parallel_end after the call.
5602 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5603 (expand_omp_for_init_counts): Handle combined loops.
5604 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5606 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5607 end of cancellable loops.
5608 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5609 Likewise. Handle collapse > 1 loops.
5610 (expand_omp_simd): Handle combined loops.
5611 (expand_omp_for): Add inner_stmt argument, adjust callers of
5612 expand_omp_for* functions, use expand_omp_for_static*chunk even
5613 for collapse > 1 unless ordered.
5614 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5615 of cancellable sections.
5616 (expand_omp_single): Remove need_barrier variable, just rely on
5617 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5618 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5619 (expand_omp_atomic_load, expand_omp_atomic_store,
5620 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5621 (expand_omp_target): New function.
5622 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5623 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5624 (build_omp_regions_1): Immediately close region for
5625 GF_OMP_TARGET_KIND_UPDATE.
5626 (maybe_add_implicit_barrier_cancel): New function.
5627 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5629 (lower_omp_single): Likewise. Add clobber after the barrier.
5630 (lower_omp_taskgroup): New function.
5631 (lower_omp_for): Handle combined loops. Adjust
5632 lower_rec_input_clauses caller. Handle cancellation.
5633 (lower_depend_clauses): New function.
5634 (lower_omp_taskreg): Lower depend clauses. Adjust
5635 lower_rec_input_clauses caller. Add clobber after the call. Handle
5637 (lower_omp_target, lower_omp_teams): New functions.
5638 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5639 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5640 and GOMP_cancellation_point calls.
5641 (lower_omp): Fold stmts inside of target region.
5642 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5643 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5644 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5645 (BT_FN_VOID_OMPFN_PTR_UINT,
5646 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5647 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5648 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5649 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5650 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5651 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5652 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5653 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5654 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5655 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5656 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5657 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5658 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5659 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5660 BUILT_IN_GOMP_PARALLEL_END.
5661 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5662 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5663 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5664 GF_OMP_FOR_KIND_DISTRIBUTE.
5665 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5666 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5667 (dump_gimple_omp_return): Print lhs if it has any.
5668 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5669 gimple_omp_atomic_seq_cst_p.
5670 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5671 and GIMPLE_OMP_TEAMS.
5672 * langhooks.c (lhd_omp_mappable_type): New function.
5673 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5674 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5676 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5677 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5678 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5679 (struct gimplify_omp_ctx): Add combined_loop field.
5680 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5681 on stmts inside of target region.
5682 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5683 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5684 ORT_TARGET and ORT_TARGET_DATA.
5685 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5687 (omp_notice_threadprivate_variable): Complain about threadprivate
5688 variables in target region.
5689 (omp_notice_variable): Complain about vars with non-mappable type
5690 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5691 (omp_check_private): Ignore ORT_TARGET* regions.
5692 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5693 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5694 (find_combined_omp_for): New function.
5695 (gimplify_omp_for): Handle gimplification of combined loops.
5696 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5698 (gimplify_omp_target_update): New function.
5699 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5700 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5701 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5702 (gimplify_body): If fndecl has "omp declare target" attribute, add
5703 implicit ORT_TARGET context around it.
5704 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5705 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5706 * tree-nested.c (convert_nonlocal_reference_stmt,
5707 convert_local_reference_stmt, convert_gimple_call): Handle
5708 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5709 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5710 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5711 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5712 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5713 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5714 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5715 BUILT_IN_GOMP_LOOP_END_CANCEL,
5716 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5717 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5718 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5719 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5720 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5721 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5722 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5723 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5724 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5725 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5726 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5727 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5728 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5729 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5730 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5731 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5732 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5733 gimple_build_omp_teams): New functions.
5734 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5735 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5736 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5737 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5738 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5739 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5740 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5741 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5742 GF_OMP_ATOMIC_SEQ_CST): New.
5743 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5744 gimple_build_omp_teams): New prototypes.
5745 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5746 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5747 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5748 GIMPLE_OMP_SINGLE as end of range.
5749 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5750 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5751 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5752 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5753 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5754 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5755 gimple_omp_target_kind, gimple_omp_target_set_kind,
5756 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5757 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5758 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5759 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5760 gimple_omp_teams_set_clauses): New inlines.
5761 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5762 and GIMPLE_OMP_TASKGROUP.
5763 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5764 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5765 enum omp_clause_proc_bind_kind): New.
5766 (union omp_clause_subcode): Add depend_kind, map_kind and
5767 proc_bind_kind fields.
5768 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5769 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5770 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5771 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5772 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5774 2013-10-10 Teresa Johnson <tejohnson@google.com>
5776 * predict.c (tree_estimate_probability): Add new parameter
5777 for estimate_bb_frequencies.
5778 (estimate_bb_frequencies): Add new parameter to force estimation.
5779 (rebuild_frequencies): When max frequency in function is small,
5780 recompute counts from frequencies.
5781 * predict.h (estimate_bb_frequencies): New parameter.
5783 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5785 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5786 optimizations are disabled.
5788 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5790 * coverage.c (coverage_finish): Fix leak of da_file_name.
5792 2013-10-10 Jan Hubicka <jh@suse.cz>
5794 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5795 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5798 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5801 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5802 if any labels are in FALLTHRU_BB, use a special label emitted
5803 immediately after the asm goto insn rather than label_rtx
5805 (expand_asm_stmt): Adjust caller.
5806 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5807 edge if the last insn in predecessor is a jump with single successor,
5808 but it isn't simplejump_p.
5810 2013-10-10 Richard Biener <rguenther@suse.de>
5812 PR tree-optimization/58656
5813 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5815 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5817 * gimplify.c: Include expr.h and tm_p.h for targets with special
5818 va-arg padding requirements.
5820 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5822 * tree-flow.h: Move some prototypes to gimple.h.
5823 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5824 * gimple.h: Relocate some prototypes from tree-flow.h
5825 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5827 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5828 (build_va_arg_indirect_ref): Relocate and make static.
5829 (std_gimplify_va_arg_expr): Relocate here.
5830 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5831 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5833 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5835 * doc/md.texi: Document the mnemonic attribute.
5837 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5840 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5841 (set_attr_alternative x ...) when searching for user defined
5844 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5846 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5848 2013-10-09 Easwaran Raman <eraman@google.com>
5850 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5851 * cfgexpand.c (defer_stack_allocation): ...use here
5852 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5854 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5856 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5857 (operand_equal_for_value_replacement): New function, extracted from
5858 value_replacement and enhanced to catch more cases.
5859 (value_replacement): Use operand_equal_for_value_replacement.
5861 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5863 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5864 get_max_loop_iterations.
5866 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5868 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5869 Place comment above function.
5871 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5873 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5874 are not related to tree-cfg.c.
5875 * tree-ssa-address.h: New file. Relocate prototypes.
5876 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5877 (addr_for_mem_ref): New. Combine call to get_address_description and
5878 return addr_for_mem_ref.
5879 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5880 * tree-ssa-live.h: Adjust prototypes.
5881 * passes.c: Include tree-ssa-live.h.
5882 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5883 * graphite.c (graphite_transform_loops): Make static.
5884 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5885 make_pass_graphite, pass_data_graphite_transforms,
5886 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5887 * ipa-pure-const.c (warn_function_noreturn): Make static.
5888 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5889 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5890 Relocate from tree-cfg.c
5891 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5893 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5894 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5895 Move to ipa-pure-const.c.
5896 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5897 Relocate from tree-optimize.c.
5898 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5899 make_pass_fixup_cfg): Move to tree-cfg.c.
5900 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5901 Relocate some prototypes.
5902 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5903 * tree-dump.c (dump_function_to_file): Remove prototype.
5904 Add tree-flow.h to the include file.
5905 * tree-dump.h: Remove prototype.
5906 * tree-parloops.h: New File. Add prototypes.
5907 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5908 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5909 from tree-ssa-loop.c.
5910 * tree-predcom.c (run_tree_predictive_commoning,
5911 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5912 Relocate here from tree-ssa-loop.c.
5913 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5914 ssa_name_values.release ().
5915 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5916 (ssa_name_values): Relocate from tree-flow.h.
5917 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5918 * tree-ssa-loop.c (run_tree_predictive_commoning,
5919 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5920 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5921 make_pass_graphite, pass_data_graphite_transforms,
5922 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5923 tree_parallelize_loops, pass_data_parallelize_loops,
5924 make_pass_parallelize_loops): Move to other files.
5925 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5927 * tree.h: Remove prototypes from tree-address.c.
5929 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5931 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5932 (struct int_tree_map): Move to tree-hasher.h
5933 (SCALE, LABEL, PERCENT): Move to gimple.h
5934 * tree-flow-inline.h: Delete. Move functions to other files.
5935 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5936 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5937 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5938 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5939 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5940 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5941 * tree-hasher.h: Don't include tree-flow.h.
5942 (struct int_tree_map): Relocate from tree-flow.h.
5943 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5944 tree-flow-inline.h and make static.
5945 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5947 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5948 tree-flow-inline.h and make static.
5949 * tree.h (is_global_var, may_be_aliased): Relocate from
5951 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5952 * value-prof.c: No longer include tree-flow-inline.h.
5953 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5955 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5957 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5958 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5959 (enum move_pos): Move to tree-ssa-loop-im.h
5960 * cfgloop.h: Move some prototypes.
5961 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5962 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5963 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5964 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5965 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5966 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5967 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5968 (enum move_pos): Relocate here.
5969 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5971 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5972 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5973 make_pass_lim): Relocate here from tree-ssa-loop.c.
5974 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5976 (loop_edge_to_cancel, unloop_loops): Make static.
5977 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5978 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5979 (tree_complete_unroll, gate_tree_complete_unroll,
5980 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5981 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5982 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5983 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5984 (stmt_invariant_in_loop_p): Remove unused function.
5985 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5986 * tree-ssa-loop-manip.h: New file. Add prototypes.
5987 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5988 (gcov_type_to_double_int): Move to cfgloop.h.
5989 (double_int_cmp, bound_index,
5990 estimate_numbers_of_iterations_loop): Make static.
5991 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5992 (max_loop_iterations): Factor out get_max_loop_iterations.
5993 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5995 * tree-ssa-loop-niter.h: New file. Add prototypes.
5996 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5997 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5998 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5999 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
6000 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
6001 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
6002 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
6003 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
6004 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
6005 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
6006 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6007 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
6008 pass_data_complete_unroll, make_pass_complete_unroll,
6009 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6010 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
6011 tree-ssa-loop-ivcanon.c.
6012 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
6013 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
6014 tree-ssa-loop-prefetch.c.
6015 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
6017 (get_lsm_tmp_name): Relocate and add suffix parameter.
6018 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
6019 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
6020 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
6021 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
6022 (get_estimated_loop_iterations): Factor out accessor from
6023 estimated_loop_iterations in tree-ssa-loop-niter.c.
6024 (get_max_loop_iterations): Factor out accessor from
6025 _max_loop_iterations in tree-ssa-niter.c.
6026 * loop-unroll.c (decide_unroll_constant_iterations,
6027 decide_unroll_runtime_iterations, decide_peel_simple,
6028 decide_unroll_stupid): Use new get_* accessors.
6030 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
6032 PR tree-optimization/20318
6033 * doc/extend.texi (returns_nonnull): New function attribute.
6034 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
6036 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
6037 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
6039 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
6042 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
6043 false if both components are bitfields.
6045 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6047 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
6048 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
6049 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
6050 * config/aarch64/aarch64.h
6051 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
6052 * config/aarch64/aarch64-simd-builtins.def
6053 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
6055 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6057 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
6059 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6061 * config/aarch64/arm_neon.h (vdiv_f64): Added.
6063 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6065 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
6066 (vneg_f64): New intrinsic.
6067 (vneg_s8): Asm replaced with C.
6068 (vneg_s16): Likewise.
6069 (vneg_s32): Likewise.
6070 (vneg_s64): New intrinsic.
6071 (vnegq_f32): Asm replaced with C.
6072 (vnegq_f64): Likewise.
6073 (vnegq_s8): Likewise.
6074 (vnegq_s16): Likewise.
6075 (vnegq_s32): Likewise.
6076 (vnegq_s64): Likewise.
6078 2013-10-09 Renlin Li <Renlin.Li@arm.com>
6080 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
6082 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6084 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
6085 packed stack special handling.
6086 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
6087 back to fixed stack slots for FPRs saved due to stdarg.
6089 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6091 * config/s390/s390.c (s390_frame_info): Restructure function.
6093 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6095 * config/s390/s390.c (struct s390_frame_layout): New field
6097 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
6098 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
6099 regs_ever_clobbered to char*.
6100 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
6101 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6102 (s390_register_info): Enable FPR save slots. Move/Copy some
6103 functionality into ...
6104 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
6105 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
6107 (s390_frame_info): Do gpr slot allocation here now. stdarg does
6108 not imply a stack frame.
6109 (s390_init_frame_layout): Remove variable clobbered_regs.
6110 (s390_update_register_info): Remove function.
6111 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
6112 cannot be used for register renaming.
6113 (s390_hard_regno_scratch_ok): New function.
6114 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
6115 (s390_initial_elimination_offset): Change offset calculation of
6116 the return address pointer.
6117 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6118 (restore_gprs): Set frame related flag.
6119 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6120 (s390_emit_prologue): Call s390_register_info instead of
6121 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
6122 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
6123 (s390_optimize_prologue): Call s390_optimize_register_info.
6124 Try to remove also FPR slot save/restore INSNs. Remove frame
6125 related flags from restore INSNs.
6127 2013-10-08 DJ Delorie <dj@redhat.com>
6129 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
6132 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
6133 avoid conflict with the MI use of %c.
6134 * config/rl78/rl78-real.md: change %c to %C throughout.
6135 * config/rl78/rl78-virt.md: Likewise.
6137 2013-10-08 Jan Hubicka <jh@suse.cz>
6139 * config/i386/i386.c (ix86_option_override_internal): Switch
6140 to SSE math for -ffast-math when target ISA supports SSE2.
6142 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
6144 * tree-flow.h: Remove some prototypes.
6145 * tree.h: Remove some protypes, add a couple.
6146 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
6147 using_eh_for_cleanups_p): Add interface routines for front ends.
6148 * tree-eh.h: New file. Add protoptyes.
6149 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
6150 (add_stmt_to_eh_lp_fn): Make static.
6151 (lower_try_finally): Use new using_eh_for_cleanups_p.
6152 * emit-rtl.c: Include tree-eh.h.
6153 * gimple.h: Include tree-eh.h.
6155 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
6157 PR tree-optimization/58480
6158 * tree-vrp.c (infer_nonnull_range): New function.
6159 (infer_value_range): Call infer_nonnull_range.
6161 2013-10-08 Dehao Chen <dehao@google.com>
6163 PR tree-optimization/58619
6164 * tree-inline.c (copy_phis_for_bb): Combine location data
6167 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6170 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
6171 RTX_FRAME_RELATED_P on INSN.
6173 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6175 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
6176 (altivec_expand_vec_perm_const): Call it.
6178 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6180 * config/rs6000/vector.md (mov<mode>): Emit permuted move
6181 sequences for LE VSX loads and stores at expand time.
6182 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
6184 * config/rs6000/rs6000.c (rs6000_const_vec): New.
6185 (rs6000_gen_le_vsx_permute): New.
6186 (rs6000_gen_le_vsx_load): New.
6187 (rs6000_gen_le_vsx_store): New.
6188 (rs6000_gen_le_vsx_move): New.
6189 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
6190 (*vsx_le_perm_load_v4si): New.
6191 (*vsx_le_perm_load_v8hi): New.
6192 (*vsx_le_perm_load_v16qi): New.
6193 (*vsx_le_perm_store_v2di): New.
6194 (*vsx_le_perm_store_v4si): New.
6195 (*vsx_le_perm_store_v8hi): New.
6196 (*vsx_le_perm_store_v16qi): New.
6197 (*vsx_xxpermdi2_le_<mode>): New.
6198 (*vsx_xxpermdi4_le_<mode>): New.
6199 (*vsx_xxpermdi8_le_V8HI): New.
6200 (*vsx_xxpermdi16_le_V16QI): New.
6201 (*vsx_lxvd2x2_le_<mode>): New.
6202 (*vsx_lxvd2x4_le_<mode>): New.
6203 (*vsx_lxvd2x8_le_V8HI): New.
6204 (*vsx_lxvd2x16_le_V16QI): New.
6205 (*vsx_stxvd2x2_le_<mode>): New.
6206 (*vsx_stxvd2x4_le_<mode>): New.
6207 (*vsx_stxvd2x8_le_V8HI): New.
6208 (*vsx_stxvd2x16_le_V16QI): New.
6210 2013-10-07 Renlin Li <Renlin.Li@arm.com>
6212 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
6214 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6216 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
6217 loop to work also for 31bit ABI.
6218 Save the stack pointer for frame_size > 0.
6220 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6222 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
6223 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
6224 constraint letters from expanders.
6225 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
6226 retry count to general_operand.
6227 ("tabort"): Give operand 0 a mode.
6228 ("tabort_1"): Add mode and constraint letter for operand 0.
6229 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
6231 2013-10-04 Jeff Law <law@redhat.com>
6233 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
6235 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
6237 (thread_across_edge): Here. Call it.
6239 2013-10-04 Cary Coutant <ccoutant@google.com>
6241 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
6242 discarding a location list expression (or a piece of one).
6244 2013-10-03 Jan Hubicka <jh@suse.cz>
6246 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
6247 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
6248 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6250 2013-10-03 Jan Hubicka <jh@suse.cz>
6252 * config/i386/i386.c (ix86_option_override_internal): Do not enable
6253 accumulate-outgoing-args when producing unwind info.
6255 2013-10-03 Wei Mi <wmi@google.com>
6257 * lra-constraints.c (insert_move_for_subreg): New function
6258 extracted from simplify_operand_subreg.
6259 (simplify_operand_subreg): Add reload for paradoxical subreg.
6261 2013-10-03 Rong Xu <xur@google.com>
6263 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
6264 the candidate of builtin_expect such that we should fix the
6265 size/time estimation.
6266 (estimate_function_body_sizes): Do the acutally size/time fix-up
6269 2013-10-03 Rong Xu <xur@google.com>
6271 * predict.c (tree_predict_by_opcode): Get the probability
6272 for builtin_expect from param builtin_expect_probability.
6273 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
6274 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
6275 * doc/invoke.texi: Add documentation for builtin-expect-probability.
6277 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
6280 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
6281 * calls.c (alloca_call_p): Use get_callee_fndecl.
6282 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
6283 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
6285 (vrp_visit_stmt): Remove duplicated code.
6287 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
6289 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
6290 ceildf2, btruncdf2, instead of vsx_* name.
6292 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
6293 iterators to only do V2DF and V4SF here. Move the DF code to
6294 rs6000.md where it is combined with SF mode. Replace <VSv> with
6295 just 'v' since only vector operations are handled with these insns
6296 after moving the DF support to rs6000.md.
6297 (vsx_sub<mode>3): Likewise.
6298 (vsx_mul<mode>3): Likewise.
6299 (vsx_div<mode>3): Likewise.
6300 (vsx_fre<mode>2): Likewise.
6301 (vsx_neg<mode>2): Likewise.
6302 (vsx_abs<mode>2): Likewise.
6303 (vsx_nabs<mode>2): Likewise.
6304 (vsx_smax<mode>3): Likewise.
6305 (vsx_smin<mode>3): Likewise.
6306 (vsx_sqrt<mode>2): Likewise.
6307 (vsx_rsqrte<mode>2): Likewise.
6308 (vsx_fms<mode>4): Likewise.
6309 (vsx_nfma<mode>4): Likewise.
6310 (vsx_copysign<mode>3): Likewise.
6311 (vsx_btrunc<mode>2): Likewise.
6312 (vsx_floor<mode>2): Likewise.
6313 (vsx_ceil<mode>2): Likewise.
6314 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6315 (vsx_sminsf3): Likewise.
6316 (vsx_fmadf4): Likewise.
6317 (vsx_fmsdf4): Likewise.
6318 (vsx_nfmadf4): Likewise.
6319 (vsx_nfmsdf4): Likewise.
6320 (vsx_cmpdf_internal1): Likewise.
6322 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6323 simpler to select whether a target has SPE or traditional floating
6324 point support in iterators.
6325 (TARGET_DF_SPE): Likewise.
6326 (TARGET_SF_FPR): Likewise.
6327 (TARGET_DF_FPR): Likewise.
6328 (TARGET_SF_INSN): Macros to say whether floating point support
6329 exists for a given operation for expanders.
6330 (TARGET_DF_INSN): Likewise.
6332 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6333 combining of SF/DF mode operations, using both traditional and VSX
6341 (abs<mode>2): Combine SF/DF modes using traditional floating point
6342 instructions. Add support for using the upper DF registers with
6343 VSX support, and SF registers with power8-vector support. Update
6344 expanders for operations supported by both the SPE and traditional
6345 floating point units.
6346 (abs<mode>2_fpr): Likewise.
6347 (nabs<mode>2): Likewise.
6348 (nabs<mode>2_fpr): Likewise.
6349 (neg<mode>2): Likewise.
6350 (neg<mode>2_fpr): Likewise.
6351 (add<mode>3): Likewise.
6352 (add<mode>3_fpr): Likewise.
6353 (sub<mode>3): Likewise.
6354 (sub<mode>3_fpr): Likewise.
6355 (mul<mode>3): Likewise.
6356 (mul<mode>3_fpr): Likewise.
6357 (div<mode>3): Likewise.
6358 (div<mode>3_fpr): Likewise.
6359 (sqrt<mode>3): Likewise.
6360 (sqrt<mode>3_fpr): Likewise.
6361 (fre<Fs>): Likewise.
6362 (rsqrt<mode>2): Likewise.
6363 (cmp<mode>_fpr): Likewise.
6364 (smax<mode>3): Likewise.
6365 (smin<mode>3): Likewise.
6366 (smax<mode>3_vsx): Likewise.
6367 (smin<mode>3_vsx): Likewise.
6368 (negsf2): Delete SF operations that are merged with DF.
6375 (fmasf4_fpr): Likewise.
6376 (fmssf4_fpr): Likewise.
6377 (nfmasf4_fpr): Likewise.
6378 (nfmssf4_fpr): Likewise.
6379 (sqrtsf2): Likewise.
6380 (rsqrtsf_internal1): Likewise.
6381 (smaxsf3): Likewise.
6382 (sminsf3): Likewise.
6383 (cmpsf_internal1): Likewise.
6384 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6385 (negdf2): Delete DF operations that are merged with SF.
6387 (nabsdf2): Likewise.
6393 (rsqrtdf_internal1): Likewise.
6394 (fmadf4_fpr): Likewise.
6395 (fmsdf4_fpr): Likewise.
6396 (nfmadf4_fpr): Likewise.
6397 (nfmsdf4_fpr): Likewise.
6398 (sqrtdf2): Likewise.
6399 (smaxdf3): Likewise.
6400 (smindf3): Likewise.
6401 (cmpdf_internal1): Likewise.
6402 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6403 (btrunc<mode>2): Delete separate expander, and combine with the
6404 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6405 (btrunc<mode>2_fpr): Likewise.
6406 (ceil<mode>2): Likewise.
6407 (ceil<mode>2_fpr): Likewise.
6408 (floor<mode>2): Likewise.
6409 (floor<mode>2_fpr): Likewise.
6410 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6411 Add support for using the upper registers with VSX and
6412 power8-vector. Move insns to be closer to the define_expands. On
6413 VSX systems, prefer the traditional form of FMA over the VSX
6414 version, since the traditional form allows the target not to
6415 overlap with the inputs.
6416 (fms<mode>4_fpr): Likewise.
6417 (nfma<mode>4_fpr): Likewise.
6418 (nfms<mode>4_fpr): Likewise.
6420 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6421 Richard Earnshaw <richard.earnshaw@arm.com>
6423 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6424 (struct mult_cost_table): Likewise.
6425 (struct mem_cost_table): Likewise.
6426 (struct fp_cost_table): Likewise.
6427 (struct vector_cost_table): Likewise.
6428 (cpu_cost_table): Likewise.
6429 * config/arm/arm.opt (mold-rts-costs): New option.
6430 (mnew-generic-costs): Likewise.
6431 * config/arm/arm.c (generic_extra_costs): New table.
6432 (cortexa15_extra_costs): Likewise.
6433 (arm_slowmul_tune): Use NULL as new costs.
6434 (arm_fastmul_tune): Likewise.
6435 (arm_strongarm_tune): Likewise.
6436 (arm_xscale_tune): Likewise.
6437 (arm_9e_tune): Likewise.
6438 (arm_v6t2_tune): Likewise.
6439 (arm_cortex_a5_tune): Likewise.
6440 (arm_cortex_a9_tune): Likewise.
6441 (arm_v6m_tune): Likewise.
6442 (arm_fa726te_tune): Likewise.
6443 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6444 (arm_cortex_tune): Use generict_extra_costs.
6445 (shifter_op_p): New function.
6446 (arm_unspec_cost): Likewise.
6447 (LIBCALL_COST): Define.
6448 (arm_new_rtx_costs): New function.
6449 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6450 table is available. Use old costs otherwise unless mnew-generic-costs
6452 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6453 (cpu_cost_table): Declare.
6455 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6458 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6459 (*subs_mul_imm_<mode>)
6460 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6461 (*sub_<shift>_<mode>)
6462 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6463 Remove k constraint.
6465 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6467 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6469 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6471 2013-10-02 Teresa Johnson <tejohnson@google.com>
6473 * predict.c (probably_never_executed): New function.
6474 (probably_never_executed_bb_p): Invoke probably_never_executed.
6475 (probably_never_executed_edge_p): Ditto.
6476 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6477 Treat profile insanities conservatively.
6479 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6481 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6483 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6485 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6486 use it. Use smaller increase for scratch. Don't increase reject
6487 for early clobber scratch.
6488 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6489 setting eliminated regs except setting fp from hfp.
6490 (lra_eliminate): Check lra_insn_recog_data on NULL.
6492 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6495 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6496 setting -mvsx-timode by default until the underlying problem is fixed.
6497 (RS6000_CPU, power7 defaults): Likewise.
6499 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6501 * config/x-linux (host-linux.o): Remove header dependencies.
6502 Use $(COMPILE) and $(POSTCOMPILE).
6503 * config/t-linux-android (linux-android.o): Ditto.
6505 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6507 * Makefile.in (expmed.o-warn): Remove.
6509 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6511 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6512 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6514 2013-10-02 Teresa Johnson <tejohnson@google.com>
6516 * dojump.c (do_jump_1): Divide probability between
6517 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6519 2013-10-02 Tom Tromey <tromey@redhat.com>
6521 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6523 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6525 * tree-flow.h: Remove some prototypes.
6526 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6527 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6528 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6529 mark_virtual_phi_result_for_renaming): Relocate here.
6530 * tree-into-ssa.h: Add prototypes.
6531 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6532 single_pred_before_succ_order.
6533 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6534 (nonfreeing_call_p) Move to gimple.c.
6535 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6537 * basic-block.h (single_pred_before_succ_order): Add prototype.
6538 * gimple.c (nonfreeing_call_p): Relocate here.
6539 * gimple.h: Add prototype.
6540 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6541 * tree-ssa-dom.h: New file. Relocate prototypes here.
6542 * tree-ssa.h: Include tree-ssa-dom.h.
6544 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6546 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6547 Use $(COMPILE) and $(POSTCOMPILE).
6549 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6551 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6553 * tree-flow.h: Remove some prototypes.
6554 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6555 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6556 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6557 propagate_tree_value*): Move from here to...
6558 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6559 propagate_tree_value*): Relocate here.
6560 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6561 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6562 * gimple-fold.c: Remove gimple-fold.h from include list.
6563 * tree-vrp.c: Remove gimple-fold.h from include list.
6564 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6565 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6566 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6567 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6568 * sese.c: Add tree-ssa-propagate.h to include list.
6570 2013-10-02 Richard Biener <rguenther@suse.de>
6572 * tree-loop-distribution.c: Include tree-vectorizer.h for
6574 (enum partition_kind): Remove PKIND_REDUCTION.
6575 (struct partition_s): Remove has_writes member, add reduction_p member.
6576 (partition_alloc): Adjust.
6577 (partition_builtin_p): Likewise.
6578 (partition_has_writes): Remove.
6579 (partition_reduction_p): New function.
6580 (partition_merge_into): Likewise.
6581 (generate_code_for_partition): Commonize builtin partition
6583 (rdg_cannot_recompute_vertex_p): Remove.
6584 (already_processed_vertex_p): Likewise.
6585 (rdg_flag_vertex): Do not set has_writes.
6586 (classify_partition): Adjust.
6587 (rdg_build_partitions): Do not set has_writes, treat all
6588 partitions as useful.
6589 (distribute_loop): Record number of library calls generated. Adjust.
6590 (tree_loop_distribution): Report number of loops and library
6591 calls generated as opt-info.
6593 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6595 * tree-flow.h: Include new .h files. Move prototypes.
6596 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6597 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6598 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6599 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6600 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6601 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6602 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6603 * tree.h (get_ref_base_and_extent): Move prototype out.
6604 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6606 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6607 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6608 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6609 * tree-scalar-evolution.c: Include tree.h.
6610 * sese.c: Include tree.h.
6611 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6612 * dwarf2out.c: Include tree-dfa.h.
6613 * tree-chrec.c: Include tree.h.
6614 * tree-data-ref.c: Include tree.h.
6616 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6618 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6621 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6623 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6624 Use $(COMPILE) and $(POSTCOMPILE).
6625 (sol2-cxx.o): Likewise.
6626 (sol2-stubs.o): Likewise.
6628 * config/x-solaris (host-solaris.o): Likewise.
6630 * config/sparc/t-sparc (sparc.o): Remove.
6631 (sparc-c.o): Remove header dependencies.
6632 Use $(COMPILE) and $(POSTCOMPILE).
6633 * config/sparc/x-sparc: Likewise.
6635 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6637 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6638 * config/arc/arc700.md: Likewise.
6639 * config/arc/arc-modes.def: Likewise.
6640 * config/arc/arc-simd.h: Likewise.
6641 * config/arc/t-arc-uClibc: Likewise.
6642 * config/arc/t-arc-newlib: Likewise.
6644 2013-10-02 Renlin Li <renlin.li@arm.com>
6646 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6648 (aarch64_expand_epilogue): Likewise.
6650 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6651 Yufeng Zhang <yufeng.zhang@arm.com>
6653 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6655 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6656 'base_in' represent a conversion and legal_cast_p_1 holds; set
6657 'base_in' with the returned value from get_unwidened.
6659 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6661 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6662 plus_constant is not used.
6664 2013-10-01 Wei Mi <wmi@google.com>
6666 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6667 * config/i386/i386.md: Add define_peephole2 to
6668 break partial reg stall for cvtss2sd/cvtsd2ss.
6670 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6672 * config/arc/arc.c (pass_arc_ifcvt::clone):
6673 Update for ctxt_ -> m_ctxt change.
6675 2013-10-01 Jeff Law <law@redhat.com>
6677 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6678 outgoing_edge and intermediate_edge fields. Instead store the path.
6679 (redirection_data::hash): Hash on the last edge's destination index.
6680 (redirection_data::equal): Check the entire thread path.
6681 (lookup_redirectio_data): Corresponding changes.
6682 (create_edge_and_update_destination_phis): Likewise.
6683 (thread_single_edge): Likewise.
6685 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6686 Diego Novillo <dnovillo@google.com>
6688 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6689 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6690 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6691 versions of these patterns.
6693 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6694 (__builtin_arc_aligned): Likewise.
6696 * config/arc/arc.md: Expand adc_0 comment stating the intended
6697 purpose and why it isn't ready.
6698 Replace commented out call_value_via_label_mixed with a
6699 plain comment about bl_s.
6701 * config/arc/arc.c (stdio.h): Don't include directly.
6702 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6703 Remove if (1) condition.
6704 (arc_encode_section_info): Fix comment.
6706 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6708 * config/arc/arc.c (arc_conditional_register_usage):
6709 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6710 Also set reg_alloc_order for DMA config regs.
6712 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6713 Jeremy Bennett <jeremy.bennett@embecosm.com>
6715 * doc/install.texi (--with-cpu): Mention ARC.
6716 (arc-*-elf32): New paragraph.
6717 (arc-linux-uclibc): Likewise.
6718 * doc/md.texi (Machine Constraints): Add ARC part.
6719 * doc/invoke.texi: (menu): Add ARC Options.
6720 (Machine Dependent Options) <ARC Options>: Add synopsis.
6721 (node ARC Options): Add.
6722 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6723 (ARC Built-in Functions): New section defining
6724 generic ARC built-in functions.
6725 (ARC SIMD Built-in Functions): New section defining SIMD specific
6727 (Declaring Attributes of Functions): Extended
6728 description of short_call and long_call attributes for ARC and
6729 added index entries.
6731 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6732 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6733 Joern Rennecke <joern.rennecke@embecosm.com>
6734 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6735 Brendan Kehoe <brendan@zen.org>
6736 Michael Eager <eager@eagercon.com>
6737 Simon Cook <simon.cook@embecosm.com>
6738 Jeremy Bennett <jeremy.bennett@embecosm.com>
6740 * config/arc, common/config/arc: New directories.
6742 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6743 Brendan Kehoe <brendan@zen.org>
6744 Simon Cook <simon.cook@embecosm.com>
6746 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6748 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6750 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6751 * tree-ssa-coalesce.h: New. Move prototype to here.
6752 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6753 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6754 (gimple_can_coalesce_p): Move to...
6755 * gimple.c (gimple_can_coalesce_p): Here.
6757 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6759 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6760 (dump_decl_set): Move to gimple.c.
6761 * gimple.h: Don't include tree-ssa-operands.h.
6762 (dump_decl_set): Add prototype.
6763 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6764 Move to gimple-ssa.h.
6765 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6766 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6767 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6768 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6769 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6771 (dump_decl_set): Relocate here.
6772 * gimple-ssa.h: New file.
6773 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6774 Relocate from gimple.h.
6775 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6776 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6777 swap_tree_operands and remove non-ssa path.
6778 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6779 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6781 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6782 vect_is_simple_reduction_1): Use swap_ssa_operands.
6783 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6784 (enum need_phi_state): Move to tree-into-ssa.c.
6785 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6786 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6787 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6788 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6789 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6790 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6791 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6792 num_imm_uses): Move to ssa-iterators.h.
6793 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6794 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6796 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6797 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6798 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6799 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6800 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6801 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6802 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6803 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6804 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6805 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6806 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6807 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6808 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6809 (set_phi_nodes): Move to tree-phinodes.h.
6810 * tree-ssa-operands.h (enum ssa_op_iter_type,
6811 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6812 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6813 (dump_decl_set): Remove prototype.
6814 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6815 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6816 (set_phi_nodes): Relocate from tree-flow-inline.h.
6817 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6819 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6820 include list. Temporarily add gimple.h to include list.
6821 * ssa-iterators.h: New file.
6822 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6823 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6824 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6825 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6826 Relocate from tree-ssa-operands.h.
6827 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6828 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6829 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6830 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6831 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6832 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6833 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6834 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6835 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6836 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6837 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6838 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6839 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6840 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6841 Relocate from tree-flow-inline.h.
6842 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6844 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6847 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6848 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6849 Insert '\t' to output template.
6850 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6851 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6852 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6853 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6855 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6857 * doc/install.texi (Host/target specific installation notes for GCC):
6858 Put @anchor before @heading.
6859 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6860 Use @email for email addresses.
6862 2013-10-01 Jeff Law <law@redhat.com>
6864 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6865 a vec. Only delete the path if we create one without successfully
6866 registering a jump thread.
6867 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6869 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6870 (paths): New vector of jump threading paths.
6871 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6872 (THREAD_PATH): New accessor macro for the entire thread path.
6873 (lookup_redirection_data): Get intermediate and final outgoing edge
6874 from the thread path.
6875 (create_edge_and_update_destination_phis): Copy the threading path.
6876 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6877 jump threading path.
6878 (ssa_redirect_edges): Get edges and block types from the jump threading
6879 path. Free the path vector.
6880 (thread_block): Get edges from the jump threading path. Look at the
6881 entire path to see if we thread to a loop exit. If we cancel a jump
6882 thread request, then free the path vector.
6883 (thread_single_edge): Get edges and block types from the jump threading
6884 path. Free the path vector.
6885 (thread_through_loop_header): Get edges and block types from the jump
6886 threading path. Free the path vector.
6887 (mark_threaded_blocks): Iterate over the vector of paths and store
6888 the path on the appropriate edge. Get edges and block types from the
6889 jump threading path.
6890 (mark_threaded_blocks): Get edges and block types from the jump
6891 threading path. Free the path vector.
6892 (thread_through_all_blocks): Use the vector of paths rather than
6893 a vector of 3-edge sets.
6894 (register_jump_thread): Accept pointer to a path vector rather
6895 than the path vector itself. Store the path vector for later use.
6898 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6899 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6902 * config/s390/s390.c (s390_split_branches): Modify check for table
6904 (s390_chunkify_start): Rearrange table jump insn check in order to
6905 deal with compare and branch insns correctly.
6907 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6911 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6912 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6913 define_insn_and_split.
6914 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6915 (shiftsi3_compare): New pattern.
6917 * config/arm/unspecs.md (UNSPEC_RRX): New.
6919 2013-10-01 Alan Modra <amodra@gmail.com>
6921 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6922 casing inout operands.
6924 2013-10-01 Richard Biener <rguenther@suse.de>
6926 PR tree-optimization/58553
6927 * tree-loop-distribution.c (struct partition_s): Add niter member.
6928 (classify_partition): Populate niter member for the partition
6929 and properly identify whether the relevant store happens before
6930 or after the loop exit.
6931 (generate_memset_builtin): Use niter member from the partition.
6932 (generate_memcpy_builtin): Likewise.
6934 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6936 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6937 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6939 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6941 * basic-block.h (edge_list): Prefix member names with "m_".
6942 * context.h (context): Likewise.
6943 * domwalk.h (dom_walker): Likewise.
6944 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6945 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6946 * hash-table.h (hash_table): Likewise.
6947 * machmode.h (bit_field_mode_iterator): Likewise.
6948 * pass_manager.h (pass_list): Likewise.
6949 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6950 * tree-pass.h (pass_data): Likewise.
6951 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6952 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6953 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6954 * asan.c (pass_data_asan): Update accordingly.
6955 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6956 (control_dependences::control_dependences): Likewise.
6957 (control_dependences::~control_dependences): Likewise.
6958 (control_dependences::~control_dependences): Likewise.
6959 (control_dependences::get_edges_dependent_on): Likewise.
6960 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6961 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6962 * context.c (gcc::context::context): Likewise.
6963 * cprop.c (pass_rtl_cprop::clone): Likewise.
6964 * domwalk.c (dom_walker::walk): Likewise.
6965 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6966 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6967 * mode-switching.c (pass_mode_switching::clone): Likewise.
6968 * passes.c (opt_pass::opt_pass): Likewise.
6969 (pass_manager::pass_manager): Likewise.
6970 * predict.c (pass_strip_predict_hints::clone): Likewise.
6971 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6972 (pass_split_all_insns::clone): Likewise.
6973 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6975 (bit_field_mode_iterator::next_mode): Likewise.
6976 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6977 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6978 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6979 * tree-complex.c (pass_lower_complex::clone): Likewise.
6980 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6981 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6982 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6983 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6984 (pass_fold_builtins::clone): Likewise.
6985 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6986 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6987 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6988 (pass_cd_dce::clone): Likewise.
6989 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6990 (pass_phi_only_cprop::clone): Likewise.
6991 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6992 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6993 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6994 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6995 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6996 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6997 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6998 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6999 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
7000 * tree-vrp.c (pass_vrp::clone): Likewise.
7001 * tsan.c (pass_tsan::clone): Likewise.
7003 2013-09-30 Jakub Jelinek <jakub@redhat.com>
7006 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
7007 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
7010 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
7011 optimization, punt if sign_bit_p looked through any zero extension.
7013 2013-09-30 Teresa Johnson <tejohnson@google.com>
7015 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
7016 Update redirected out edge count in joiner case.
7017 (ssa_redirect_edges): Common the joiner and non-joiner cases
7018 so that joiner case gets profile updates.
7020 2013-09-30 Richard Biener <rguenther@suse.de>
7022 PR tree-optimization/58554
7023 * tree-loop-distribution.c (classify_partition): Require
7024 unconditionally executed stores for memcpy and memset recognition.
7025 (tree_loop_distribution): Calculate dominance info.
7027 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
7029 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
7030 (NO_PROFILE_COUNTERS): Likewise.
7031 (PROFILE_HOOK): Likewise.
7032 (FUNCTION_PROFILER): Likewise.
7033 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
7035 2013-09-30 Iain Sandoe <iain@codesourcery.com>
7037 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
7038 calls and defines in TARGET_MACHO conditional.
7039 (load_macho_picbase_di): Likewise.
7040 (reload_macho_picbase): Likewise.
7041 (reload_macho_picbase_si): Likewise.
7042 (reload_macho_picbase_di): Likewise.
7043 (nonlocal_goto_receiver): Likewise.
7045 2013-09-30 Nick Clifton <nickc@redhat.com>
7047 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
7048 that use the MSP430X ISA.
7049 (msp430_option_override): Scan -mmcu command line option for any
7050 MCU name that supports the MSP430X ISA.
7051 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
7052 -mmcu options which enable the MSP430X ISA.
7054 2013-09-30 Richard Biener <rguenther@suse.de>
7057 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
7058 before looking for setjmp-like calls.
7060 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7063 * config/darwin-protos.h (machopic_get_function_picbase): New.
7064 * config/darwin.c (machopic_get_function_picbase): New.
7065 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
7066 label for a new func. (load_macho_picbase_di): Likewise.
7067 (reload_macho_picbase): New expand.
7068 (reload_macho_picbase_si): New insn.
7069 (reload_macho_picbase_di): New insn.
7070 (nonlocal_goto_receiver): New define and split.
7071 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
7072 (unspecv enum): Add UNSPECV_NLGR.
7074 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7076 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
7077 that altivec registers are correctly sized on Darwin.
7079 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7081 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
7082 darwin-driver.o): Use COMPILE and POSTCOMPILE.
7083 * config/x-darwin (host-darwin.o): Likewise.
7084 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
7085 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
7086 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
7088 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
7090 * doc/invoke.texi: Fix usage of @tie{} command.
7092 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
7094 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
7096 2013-09-28 Jan Hubicka <jh@suse.cz>
7098 * config/i386/x86-tune.def: Add documentation for each of the options;
7101 2013-09-28 Jan Hubicka <jh@suse.cz>
7103 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
7105 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
7106 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
7107 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
7108 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
7109 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
7111 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
7113 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
7114 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
7115 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
7116 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
7117 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
7118 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
7119 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
7120 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
7121 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
7122 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
7123 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
7124 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
7125 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
7126 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
7127 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
7128 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
7129 gimple.h, godump.c, graphite-clast-to-gimple.c,
7130 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
7131 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
7132 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
7133 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7134 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
7135 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
7136 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
7137 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
7138 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
7139 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
7140 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
7141 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
7142 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
7143 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
7144 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
7145 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
7146 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
7147 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
7148 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
7149 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
7150 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
7151 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
7152 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
7153 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
7154 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7155 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7156 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
7157 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
7158 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
7159 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7160 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7161 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
7162 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
7163 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7164 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7165 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
7166 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
7167 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
7168 whitespace before "(".
7170 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
7172 * expr.h (extract_bit_field): Remove packedp parameter.
7173 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
7174 from forward declaration.
7175 (store_split_bit_field): Remove packedp arg from calls to
7176 extract_fixed_bit_field.
7177 (extract_bit_field_1): Remove packedp parameter and packedp
7178 argument from recursive calls and calls to extract_fixed_bit_field.
7179 (extract_bit_field): Remove packedp parameter and corresponding
7180 arg to extract_bit_field_1.
7181 (extract_fixed_bit_field): Remove packedp parameter. Remove code
7183 (extract_split_bit_field): Remove packedp arg from call to
7184 extract_fixed_bit_field.
7185 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
7186 (copy_blkmode_from_reg): Likewise.
7187 (copy_blkmode_to_reg): Likewise.
7188 (read_complex_part): Likewise.
7189 (store_field): Likewise.
7190 (expand_expr_real_1): Likewise.
7191 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
7192 to extract_bit_field.
7193 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
7194 call to extract_bit_field.
7195 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
7196 call to extract_bit_field.
7197 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
7198 and special packedp behavior from -fstrict-volatile-bitfields
7201 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7203 * lra-eliminations.c (init_elim_table): Guard value_p.
7205 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7207 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
7208 DFmode, DImode, and SFmode in the upper VSX registers based on the
7209 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
7210 if -mpower8-vector. Combine -mvsx-timode handling with the rest
7211 of the VSX register handling.
7213 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
7215 (zero_extendsidi2_lfiwzx): Add support for loading into the
7216 Altivec registers with -mpower8-vector. Use wu/wv constraints to
7217 only do VSX memory options on Altivec registers.
7218 (extendsidi2_lfiwax): Likewise.
7219 (extendsfdf2_fpr): Likewise.
7220 (mov<mode>_hardfloat, SF/SD modes): Likewise.
7221 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
7222 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
7223 (movdi_internal64): Likewise.
7225 2013-09-27 Xinliang David Li <davidxl@google.com>
7227 * opts.c (finish_options): Adjust parameters
7228 according to vect cost model.
7229 (common_handle_option): Set dynamic vect cost
7231 targhooks.c (default_add_stmt_cost): Compute stmt cost
7233 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
7234 Use helper function.
7235 * tree-vectorizer.h (unlimited_cost_model): New function.
7236 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
7237 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
7239 (vect_enhance_data_refs_alignment): Ditto.
7240 * flag-types.h: New enum.
7241 * common/config/i386/i386-common.c (ix86_option_init_struct):
7242 No need to initialize vect_cost_model flag.
7243 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
7246 2013-09-27 Diego Novillo <dnovillo@google.com>
7248 * gimple.h (enum ssa_mode): Remove.
7250 2013-09-27 Paulo Matos <pmatos@broadcom.com>
7252 * cfgloop.h (number_of_loops): Fix typo in check for null.
7254 2013-09-27 Jakub Jelinek <jakub@redhat.com>
7257 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
7258 are children of outermost saved_cfun's loop, and set it up to
7259 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
7260 if loop != loop0 and SESE region contains bbs that belong to loop0.
7262 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7264 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
7265 (binary_scale_code_p, get_base_term, get_index_term): New functions.
7266 (set_address_segment, set_address_base, set_address_index)
7267 (set_address_disp): Accept the argument unconditionally.
7268 (baseness): Remove must_be_base_p and must_be_index_p checks.
7269 (decompose_normal_address): Classify as much as possible in the
7272 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7274 * cse.c (count_reg_usage): Handle INT_LIST.
7275 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
7276 * reginfo.c (reg_scan_mark_refs): Likewise.
7277 * reload1.c (eliminate_regs_1): Likewise.
7279 2013-09-27 Iain Sandoe <iain@codesourcery.com>
7282 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
7285 2013-09-27 Richard Biener <rguenther@suse.de>
7287 PR tree-optimization/58459
7288 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
7289 restriction not propagating into loops.
7291 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
7293 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
7294 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
7295 * doc/tree-ssa.texi (Walking use-def chains): Delete.
7297 2013-09-26 Richard Biener <rguenther@suse.de>
7299 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7301 2013-09-26 Richard Biener <rguenther@suse.de>
7303 * alias.h (component_uses_parent_alias_set): Rename to ...
7304 (component_uses_parent_alias_set_from): ... this.
7305 * alias.c (component_uses_parent_alias_set): Rename to ...
7306 (component_uses_parent_alias_set_from): ... this and return
7308 (reference_alias_ptr_type_1): Use the result from
7309 component_uses_parent_alias_set_from instead of stripping
7310 components one at a time.
7311 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7313 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7315 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7316 Move prototypes to...
7317 * tree-ssa-ter.h: New File. Move prototypes here.
7318 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7319 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7320 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7321 from is_replaceable_p.
7322 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7323 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7324 (process_replaceable): Use ter_is_replaceable_p.
7325 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7326 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7327 newly refactored ssa_is_replaceable_p.
7328 * cfgexpand.c: Include tree-outof-ssa.h.
7329 * ssaexpand.h: Delete.
7331 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7333 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7334 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7336 (create_gimple_tmp): Delete.
7337 (get_expr_type, build_assign, build_type_cast): Move to...
7338 * gimple-builder.c: New File.
7339 (get_expr_type): Relocate from gimple.c.
7340 (build_assign, build_type_cast): Change to only create ssanames.
7341 * gimple.h: Move prototypes to...
7342 * gimple-builder.h: New File. Here.
7343 * tree-ssa.h: And here.
7344 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7345 count_uses_and_derefs): Relocate from gimple.c.
7346 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7347 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7348 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7349 gimple_replace_ssa_lhs.
7350 * asan.c: Include gimple-builder.h.
7351 * Makefile.in: Add gimple-builder.o.
7353 2013-09-26 Richard Biener <rguenther@suse.de>
7355 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7356 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7357 (loe_visit_block): Use gcc_checking_assert.
7358 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7359 gimple_assign_ssa_name_copy_p.
7360 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7363 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7365 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7366 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7368 2013-09-26 Richard Biener <rguenther@suse.de>
7370 PR tree-optimization/58539
7371 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7372 the fact that debug statements are not taking part in loop-closed
7375 2013-09-26 Nick Clifton <nickc@redhat.com>
7377 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7378 time warning message.
7379 (msp430_print_operand_raw): Delete unused letter parameter.
7380 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7381 (msp430_print_operand_address): New function.
7382 (msp430_print_operand): Move address printing code from here to
7384 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7386 (zero_extendpsisi2): Likewise.
7387 (extendpsisi2): New pattern.
7388 (andneghi3): New pattern.
7390 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7392 * config/aarch64/aarch64.opt (mlra): New option.
7393 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7394 (TARGET_LRA_P): Define.
7396 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7398 * expr.c (expand_assignment): Remove obsolete comment.
7400 2013-09-25 Jeff Law <law@redhat.com>
7402 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7403 tree-ssa-threadupdate.h.
7404 (register_jump_thread): Similarly.
7405 * tree-ssa-threadupdate.h: New header file.
7406 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7407 * tree-vrp.c: Likewise.
7408 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7409 (thread_around_empty_blocks): Change type of path vector argument to
7410 an edge,type pair from just an edge. Initialize both elements when
7411 appending to a jump threading path. Tweak references to elements
7413 (thread_across_edge): Similarly. Release memory for the elements
7415 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7416 (dump_jump_thread_path): New function broken out from
7417 register_jump_thread.
7418 (register_jump_thread): Use dump_jump_thread_path. Change type of
7419 path vector entries. Search the path for NULL edges and dump
7420 the path if one is found. Tweak the conversion of path to 3-edge
7421 form to use the block copy type information embedded in the path.
7423 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7425 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7427 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7428 Vladimir Makarov <vmakarov@redhat.com>
7430 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7431 from the least significant bit.
7432 (strip_address_mutations): Add bitfield operations handling.
7433 (must_be_index_p): Add shifting and rotate operations handling.
7434 (set_address_base): Use must_be_base_p predicate.
7435 (set_address_index): Use must_be_index_p predicate.
7437 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7438 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7439 Sergey Lega <sergey.s.lega@intel.com>
7440 Anna Tikhonova <anna.tikhonova@intel.com>
7441 Ilya Tocar <ilya.tocar@intel.com>
7442 Andrey Turetskiy <andrey.turetskiy@intel.com>
7443 Ilya Verbin <ilya.verbin@intel.com>
7444 Kirill Yukhin <kirill.yukhin@intel.com>
7445 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7447 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7449 (ix86_expand_vector_move_misalign): Support new unaligned load and
7450 stores and use new names.
7451 (CODE_FOR_sse2_storedqu): Rename to ...
7452 (CODE_FOR_sse2_storedquv16qi): ... this.
7453 (CODE_FOR_sse2_loaddqu): Rename to ...
7454 (CODE_FOR_sse2_loaddquv16qi): ... this.
7455 (CODE_FOR_avx_loaddqu256): Rename to ...
7456 (CODE_FOR_avx_loaddquv32qi): ... this.
7457 (CODE_FOR_avx_storedqu256): Rename to ...
7458 (CODE_FOR_avx_storedquv32qi): ... this.
7459 * config/i386/i386.md (fpint_logic): New.
7460 * config/i386/sse.md (VMOVE): Extend for AVX512.
7464 (VI_UNALIGNED_LOADSTORE): Ditto.
7465 (sse2_avx_avx512f): Ditto.
7466 (sse2_avx2): Extend for AVX512.
7467 (sse4_1_avx2): Ditto.
7468 (avx2_avx512f): New.
7469 (sse): Extend for AVX512.
7472 (avxsizesuffix): Ditto.
7473 (sseintvecmode): Ditto.
7475 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7476 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7477 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7478 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7479 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7480 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7481 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7482 (STORENT_MODE): Extend for AVX512.
7483 (*absneg<mode>2): Replace constraint "x" with "v".
7484 (*mul<mode>3): Ditto.
7485 (*ieee_smin<mode>3): Ditto.
7486 (*ieee_smax<mode>3): Ditto.
7487 (avx_cmp<mode>3): Replace VF with VF_128_256.
7488 (*<sse>_maskcmp<mode>3_comm): Ditto.
7489 (<sse>_maskcmp<mode>3): Ditto.
7490 (<sse>_andnot<mode>3): Extend for AVX512.
7491 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7492 (<code><mode>3, fpint_logic): New.
7493 (*<code><mode>3): Extend for AVX512.
7494 (avx512flogicsuff): New.
7495 (avx512f_<logic><mode>): Ditto.
7496 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7498 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7499 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7500 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7501 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7502 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7503 (xop_vpermil2<mode>3): Ditto.
7504 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7505 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7506 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7507 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7509 2013-09-25 Tom Tromey <tromey@redhat.com>
7511 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7512 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7513 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7514 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7515 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7516 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7517 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7518 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7519 (GRAPHITE_HTAB_H): Remove.
7521 2013-09-25 Tom Tromey <tromey@redhat.com>
7523 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7525 2013-09-25 Tom Tromey <tromey@redhat.com>
7527 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7529 2013-09-25 Tom Tromey <tromey@redhat.com>
7531 * config/i386/t-i386 (i386.o): Remove.
7532 (i386-c.o): Use COMPILE and POSTCOMPILE.
7534 2013-09-25 Tom Tromey <tromey@redhat.com>
7536 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7538 2013-09-25 Tom Tromey <tromey@redhat.com>
7540 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7541 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7542 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7543 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7544 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7545 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7546 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7547 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7548 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7549 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7550 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7551 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7552 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7553 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7554 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7555 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7556 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7557 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7558 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7559 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7560 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7561 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7562 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7563 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7564 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7565 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7566 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7567 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7568 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7569 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7570 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7571 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7572 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7573 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7574 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7575 (graphite-clast-to-gimple.o, graphite-dependences.o)
7576 (graphite-interchange.o, graphite-poly.o)
7577 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7578 (graphite-optimize-isl.o, tree-vect-loop.o)
7579 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7580 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7581 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7582 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7583 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7584 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7585 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7586 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7587 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7588 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7589 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7590 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7591 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7592 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7593 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7594 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7595 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7596 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7597 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7598 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7599 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7600 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7601 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7602 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7603 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7604 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7605 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7606 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7607 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7608 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7609 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7610 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7611 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7612 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7613 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7614 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7615 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7616 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7617 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7618 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7619 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7620 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7621 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7622 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7623 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7624 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7625 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7626 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7627 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7628 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7629 (gcov-dump.o): Remove.
7630 (default-c.o): Use COMPILE and POSTCOMPILE.
7631 (CFLAGS-gcc.o): New variable.
7632 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7634 2013-09-25 Tom Tromey <tromey@redhat.com>
7636 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7637 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7638 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7639 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7640 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7641 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7642 (c-family/array-notation-common.o, c-family/stub-objc.o)
7643 (c-family/c-ubsan.o): Remove.
7645 2013-09-25 Tom Tromey <tromey@redhat.com>
7647 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7649 2013-09-25 Tom Tromey <tromey@redhat.com>
7651 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7652 to add -DENABLE_SHARED_LIBGCC.
7653 (gcc.o): Don't use subshell.
7655 2013-09-25 Tom Tromey <tromey@redhat.com>
7657 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7658 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7659 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7660 * configure, config.in: Rebuild.
7662 2013-09-25 Tom Tromey <tromey@redhat.com>
7664 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7665 (COMPILE, POSTCOMPILE): New variables.
7666 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7667 (DEPFILES): New variable.
7668 Include ".Po" files.
7669 * configure.ac: Add checks for dependency checking.
7670 * configure, aclocal.m4: Regenerate.
7672 2013-09-25 Tom Tromey <tromey@redhat.com>
7674 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7675 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7677 2013-09-25 Tom Tromey <tromey@redhat.com>
7679 * Makefile.in (generated_files): Add options.h,
7680 target-hooks-def.h, insn-opinit.h,
7681 common/common-target-hooks-def.h, pass-instances.def,
7682 c-family/c-target-hooks-def.h.
7684 2013-09-25 Jeff Law <law@redhat.com>
7686 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7687 than foo[foo.length () - 1] to access last member in a vec.
7688 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7690 2013-09-25 Richard Biener <rguenther@suse.de>
7693 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7695 2013-09-25 Jan Hubicka <jh@suse.cz>
7697 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7700 2013-09-25 Marek Polacek <polacek@redhat.com>
7703 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7704 when determining the type name.
7706 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7708 * config/sh/sh.md: Fix formatting.
7710 2013-09-24 Xinliang David Li <davidxl@google.com>
7712 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7713 max peel iterations parameter.
7714 * param.def: New parameter.
7715 * doc/invoke.texi: Document New parameter.
7717 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7719 * gimple-pretty-print.c: Various whitespace tweaks.
7720 * tree-core.h: Likewise.
7721 * tree-pretty-print.c: Likewise.
7722 * tree-ssa-alias.c: Likewise.
7723 * tree-ssa-copy.c: Likewise.
7724 * tree-ssanames.c: Likewise.
7725 * tree-ssanames.h: Likewise.
7726 * tree-vrp.c: Likewise.
7728 2013-09-24 Alan Modra <amodra@gmail.com>
7732 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7733 for output operands that disallow regs. Don't use EXPAND_WRITE on
7736 2013-09-24 Richard Biener <rguenther@suse.de>
7739 * tree.c (reference_alias_ptr_type): Move ...
7740 * alias.c (reference_alias_ptr_type): ... here and implement
7741 in terms of the new reference_alias_ptr_type_1.
7742 (ref_all_alias_ptr_type_p): New helper.
7743 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7744 use ref_all_alias_ptr_type_p.
7745 (get_deref_alias_set): Add flag_strict_aliasing check here.
7746 (reference_alias_ptr_type_1): New function, split out from ...
7747 (get_alias_set): ... here.
7748 (alias_ptr_types_compatible_p): New function.
7749 * alias.h (reference_alias_ptr_type): Declare.
7750 (alias_ptr_types_compatible_p): Likewise.
7751 * tree.h (reference_alias_ptr_type): Remove.
7752 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7753 to compare MEM_REF alias types.
7755 2013-09-24 Richard Biener <rguenther@suse.de>
7757 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7759 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7761 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7762 reload helper function arrays into a single array reg_addr.
7763 (reload_fpr_gpr): Likewise.
7764 (reload_gpr_vsx): Likewise.
7765 (reload_vsx_gpr): Likewise.
7766 (struct rs6000_reg_addr): Likewise.
7767 (reg_addr): Likewise.
7768 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7769 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7770 (rs6000_init_hard_regno_mode_ok): Likewise.
7771 (rs6000_secondary_reload_direct_move): Likewise.
7772 (rs6000_secondary_reload): Likewise.
7774 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7775 constraints: wu, ww, and wy. Repurpose wv constraint added during
7776 power8 changes. Put wg constraint in alphabetical order.
7778 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7779 for future work to add ISA 2.07 VSX single precision support.
7780 (-mvsx-scalar-double): Change default from -1 to 1, update
7781 documentation comment.
7782 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7783 (-mupper-regs-df): New debug switch to control whether DF values
7784 can go in the traditional Altivec registers.
7785 (-mupper-regs-sf): New debug switch to control whether SF values
7786 can go in the traditional Altivec registers.
7788 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7790 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7791 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7792 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7793 (rs6000_debug_legitimate_address_p): Print if we are running
7794 before, during, or after reload.
7795 (rs6000_secondary_reload): Add a comment.
7796 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7798 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7799 constraints. Update documentation string.
7800 (wd constraint): Likewise.
7801 (wf constraint): Likewise.
7802 (wg constraint): Likewise.
7803 (wn constraint): Likewise.
7804 (ws constraint): Likewise.
7805 (wt constraint): Likewise.
7806 (wx constraint): Likewise.
7807 (wz constraint): Likewise.
7808 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7810 (ww constraint): Likewise.
7811 (wy constraint): Likewise.
7812 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7813 in the previous submissions.
7814 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7816 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7818 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7819 (REG_BR_PROB): Say that the probability is stored in an int_list.
7820 * reg-notes.def: Update commentary to mention INT_LIST.
7821 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7822 (INT_LIST): New rtx.
7823 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7824 * rtlanal.c (int_reg_note_p): New function.
7825 (alloc_reg_note): Assert that the note does not have an int argument.
7826 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7827 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7828 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7829 rather than an INSN_LIST. Handle INT_LIST.
7830 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7831 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7832 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7833 Manipulate them as ints rather than rtxes.
7834 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7835 * regmove.c (copy_src_to_dest): Likewise.
7836 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7838 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7839 into the cases that need it.
7840 * config/arm/arm.c (arm_unwind_emit): Likewise.
7842 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7843 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7844 * loop-doloop.c (add_test, doloop_modify): Likewise.
7845 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7846 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7847 * predict.c (combine_predictions_for_insn): Likewise.
7848 * print-rtl.c (print_rtx): Handle INT_LIST.
7849 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7850 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7851 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7852 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7853 (ix86_split_fp_branch, predict_jump): Likewise.
7854 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7855 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7856 * config/spu/spu.c (ea_load_store_inline): Likewise.
7858 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7859 value of a REG_BR_PROB note.
7860 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7861 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7862 * emit-rtl.c (try_split): Likewise.
7863 * predict.c (br_prob_note_reliable_p): Likewise.
7864 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7865 * reorg.c (mostly_true_jump): Likewise.
7866 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7867 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7868 * config/i386/i386.c (ix86_print_operand): Likewise.
7869 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7870 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7871 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7872 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7873 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7874 * config/sparc/sparc.c (output_cbranch): Likewise.
7875 * config/spu/spu.c (get_branch_target): Likewise.
7876 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7877 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7879 2013-09-23 Jan Hubicka <jh@suse.cz>
7881 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7883 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7884 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7886 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7887 (gimple_fold_call): Dump inconsistent devirtualizations; add
7888 sanity check for type based devirtualizations.
7889 * ipa-prop.c: Include ipa-utils.h
7890 (ipa_intraprocedural_devirtualization): Add sanity check.
7891 (try_make_edge_direct_virtual_call): Likewise.
7893 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7895 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7896 assignment statements.
7898 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7900 * gimple-pretty-print.c (dump_ssaname_info): New function.
7901 (dump_gimple_phi): Call it.
7902 (pp_gimple_stmt_1): Likewise.
7903 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7904 (range_info_def): Declare.
7905 * tree-pretty-print.c (pp_double_int): New function.
7906 (dump_generic_node): Call it.
7907 * tree-pretty-print.h (pp_double_int): Declare.
7908 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7909 * tree-ssanames.h (range_info_def): New structure.
7910 (value_range_type): Move definition here.
7911 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7913 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7915 (set_range_info): New function.
7916 (get_range_info): Likewise.
7917 (duplicate_ssa_name_range_info): Likewise.
7918 (duplicate_ssa_name_fn): Check pointer type and call
7919 duplicate_ssa_name_range_info.
7920 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7921 * tree-vrp.c (value_range_type): Remove definition, now in
7923 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7924 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7925 (SSA_NAME_RANGE_INFO): New macro.
7927 2013-09-23 Richard Biener <rguenther@suse.de>
7929 PR tree-optimization/58464
7930 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7931 (phi_trans_add): Change to add conditionally on being not present.
7932 (phi_translate_1): Remove recursion detection here.
7933 (phi_translate): Pre-seed the cache with NULL to catch
7934 recursion here in a more generic way.
7935 (bitmap_find_leader): Adjust comment.
7936 (get_representative_for): Dump value-numbers.
7937 (create_expression_by_pieces): Likewise.
7938 (insert_into_preds_of_block): Likewise.
7940 2013-09-23 Christian Bruel <christian.bruel@st.com>
7943 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7944 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7946 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7949 2013-09-20 Renlin Li <renlin.li@arm.com>
7951 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7953 (aarch64_expand_epilogue): Likewise.
7954 (aarch64_legitimize_reload_address): Likewise.
7956 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7958 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7959 NULL_TREE before pushing them onto the vector. Likewise for labels.
7961 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7963 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7964 comparison operators when -fno-trapping-math is in effect.
7965 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7966 comparison operators in TFmode and assert that unsupported operators
7968 (ia64_print_operand): Likewise.
7970 2013-09-21 Jan Hubicka <jh@suse.cz>
7972 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7973 (sse_typeless_stores): Enable for core
7974 (sse_load0_by_pxor): Likewise.
7975 (four_jump_limit): Disable for core.
7976 (pad_returns): Likewise.
7977 (avoid_vector_decode): Likewise.
7978 (fuse_cmp_and_branch): Enable for cores.
7979 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7981 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7984 * config/pa/pa.c (pa_option_override): Disable auto increment and
7985 decrement instructions until reload is completed.
7987 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7988 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7989 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7991 2013-09-20 DJ Delorie <dj@redhat.com>
7992 Nick Clifton <nickc@redhat.com>
7994 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7995 (need_to_save): Save bank 0 on interrupts.
7996 (characterize_address): Strip far address wrappers.
7997 (rl78_as_legitimate_address): Likewise.
7998 (transcode_memory_rtx): Likewise.
7999 (rl78_peep_movhi_p): Disable this peephole after devirt.
8000 (rl78_propogate_register_origins): Forget all origins when a
8002 * config/rl78/rl78-virt.md: Various whitespace tweaks.
8003 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
8005 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
8006 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
8007 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
8009 * config/rl78/constraints.md: Various whitespace and paren tweaks.
8011 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8013 * config/pa/pa.md: In "scc" insn patterns, change output template to
8014 handle const0_rtx in reg_or_0_operand operands.
8016 2013-09-20 Martin Husemann <martin@NetBSD.org>
8019 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
8020 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
8022 2013-09-20 Richard Biener <rguenther@suse.de>
8025 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
8026 remembering SSA name version and block index.
8027 (new_scev_info_str): Adjust.
8028 (hash_scev_info): Likewise. Also hash the block index.
8029 (eq_scev_info): Adjust.
8030 (find_var_scev_info): Likewise.
8031 (struct instantiate_cache_entry): Remove.
8032 (struct instantiate_cache_type): Use a htab to map name, block
8034 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8035 (get_instantiated_value_entry): Likewise.
8036 (hash_idx_scev_info, eq_idx_scev_info): New functions.
8037 (instantiate_scev_name): Adjust.
8039 2013-09-20 Jeff Law <law@redhat.com>
8041 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
8043 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
8045 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
8046 Call aarch64_simd_expand_args to update op[argc].
8048 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8050 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
8053 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8055 * gengtype.c (file_rules): Added rule for *.cc files.
8056 (get_output_file_with_visibility): Give fatal message when no
8059 2013-09-20 Renlin Li <renlin.li@arm.com>
8061 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
8062 (aarch64_expand_epilogue): Likewise.
8063 (aarch64_legitimize_reload_address): Likewise.
8065 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
8068 * expr.c (expand_assignment): Remove misalignp code path.
8070 2013-09-20 Marek Polacek <polacek@redhat.com>
8073 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
8074 TYPE_PRECISION. Add asserts.
8076 2013-09-20 Richard Biener <rguenther@suse.de>
8078 PR tree-optimization/58453
8079 * tree-loop-distribution.c (distribute_loop): Apply the cost
8080 model for -ftree-loop-distribute-patterns, too.
8082 2013-09-20 Richard Biener <rguenther@suse.de>
8085 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
8086 make type comparison less strict.
8088 2013-09-20 Alan Modra <amodra@gmail.com>
8090 * configure: Regenerate.
8091 * aclocal.m4: Regenerate.
8093 2013-09-20 Marek Polacek <polacek@redhat.com>
8096 * doc/extend.texi: Document that attribute used is meant to be used
8097 on variables with static storage duration.
8099 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8101 PR tree-optimization/58472
8102 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
8103 simd_lane_access set inv_p = false.
8104 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
8105 the simduid magic VAR_DECL.
8107 2013-09-19 Jan Hubicka <jh@suse.cz>
8109 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
8111 2013-09-17 Jeff Law <law@redhat.com>
8113 * tree-ssa-dom.c (record_temporary_equivalences): New function
8114 split out of dom_opt_dom_walker::after_dom_children.
8115 (dom_opt_dom_walker::thread_across_edge): Move common code
8116 in here from dom_opt_dom_walker::after_dom_children.
8117 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
8119 2013-09-19 Jan Hubicka <jh@suse.cz>
8121 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
8122 (TARGET_GENERIC): Use PROCESOR_GENERIC
8123 (enum processor_type): Unify generic32 and 64.
8124 * i386.md (cpu): Likewise.
8125 * x86-tune.def (use_leave): Enable for generic32.
8126 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
8127 * athlon.md: Change generic64 to generic in all occurences.
8128 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
8129 (ix86_target_macros_internal): Likewise.
8130 * driver-i386.c (host_detect_local_cpu): Likewise.
8131 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
8133 (generic_memcpy, generic_memset, generic_cost): This one.
8134 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
8135 (m_GENERIC32, m_GENERIC64): Remove.
8136 (m_GENERIC): Turn into one flag.
8137 (processor_target): Unify generic tunnings.
8138 (ix86_option_override_internal): Replace generic32/64 by generic.
8139 (ix86_issue_rate): Likewise.
8140 (ix86_adjust_cost): Likewise.
8142 2013-09-19 Jan Hubicka <jh@suse.cz>
8144 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
8145 of speculative flag.
8147 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8149 * omp-low.c (expand_omp_sections): Always pass len - 1 to
8150 GOMP_sections_start, even if !exit_reachable.
8152 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8154 * lra-constraints.c (need_for_all_save_p): Use macro
8155 HARD_REGNO_CALL_PART_CLOBBERED.
8156 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
8157 set up pseudo conflict hard regs.
8159 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8162 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
8165 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8167 PR rtl-optimization/58438
8168 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
8169 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
8170 from previous subpasses.
8172 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
8174 * arm.c (arm_get_frame_offsets): Validate architecture supports
8175 LDRD/STRD before accepting the tuning preference.
8176 (arm_expand_prologue): Likewise.
8177 (arm_expand_epilogue): Likewise.
8179 2013-09-18 Richard Biener <rguenther@suse.de>
8181 PR tree-optimization/58417
8182 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
8183 have chrecs with symbols defined in the loop as operands.
8184 (chrec_fold_multiply): Likewise.
8185 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
8186 parameters before folding binary operations.
8187 (struct instantiate_cache_entry_hasher): Remove.
8188 (struct instantiate_cache_type): Use a pointer-map.
8189 (instantiate_cache_type::instantiate_cache_type): New function.
8190 (instantiate_cache_type::get): Likewise.
8191 (instantiate_cache_type::set): Likewise.
8192 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8193 (get_instantiated_value_entry): Likewise.
8194 (global_cache): New global.
8195 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
8196 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
8197 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
8198 (instantiate_scev_name): Adjust.
8199 (instantiate_scev): Construct global instead of local cache.
8200 (resolve_mixers): Likewise.
8202 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
8203 Paolo Carlini <paolo.carlini@oracle.com>
8206 * doc/implement-cxx.texi: Fix references to the C++ standards.
8208 2013-09-18 Jakub Jelinek <jakub@redhat.com>
8210 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
8211 * tree-vect-data-refs.c (vect_analyze_data_refs): For
8212 simd_lane_access drs, update also DR_ALIGNED_TO.
8214 2013-09-18 Marek Polacek <polacek@redhat.com>
8217 * doc/extend.texi: Document no_sanitize_undefined attribute.
8218 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
8219 no_sanitize_undefined attribute.
8221 2013-09-18 Nick Clifton <nickc@redhat.com>
8223 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
8224 (ASM_DECLARE_FUNCTION_NAME): Define.
8226 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
8228 * compare-elim.c (find_comparison_dom_walker): New class
8229 (find_comparisons_in_bb): Rename to
8230 find_comparison_dom_walker::before_dom_children
8231 (find_comparisons): Adjust
8232 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
8234 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
8235 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
8236 (init_walk_dominator_tree): Remove declaration.
8237 (fini_walk_dominator_tree): Remove declaration.
8238 * fwprop.c (single_def_use_dom_walker): New class
8239 (single_def_use_enter_block): Convert to
8240 single_def_use_dom_walker::before_dom_children.
8241 (single_def_use_leave_block): Convert to
8242 single_def_use_dom_walker::after_dom_children.
8243 (build_single_def_use_links): Adjust.
8244 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
8246 (find_candidates_in_block): Convert to
8247 find_candidates_dom_walker::before_dom_children.
8248 (execute_strength_reduction): Adjust.
8249 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
8250 (sese_dom_walker): New class.
8251 (sese_dom_walker::sese_dom_walker): New constructor.
8252 (sese_dom_walker::~sese_dom_walker): New destructor.
8253 (build_sese_conditions_before): Convert to
8254 sese_dom_walker::before_dom_children.
8255 (build_sese_conditions_after): Convert to
8256 sese_dom_walker::after_dom_children.
8257 (build_poly_scop): Adjust
8258 * tree-into-ssa.c (rewrite_dom_walker): New class
8259 (rewrite_enter_block): Convert to
8260 rewrite_dom_walker::before_dom_children.
8261 (rewrite_leave_block): Convert to
8262 rewrite_dom_walker::after_dom_children.
8263 (rewrite_update_dom_walker): New class.
8264 (rewrite_update_enter_block): Convert to
8265 rewrite_update_dom_walker::before_dom_children.
8266 (rewrite_update_leave_block): Convert to
8267 rewrite_update_dom_walker::after_dom_children.
8268 (rewrite_blocks, rewrite_into_ssa): Adjust.
8269 (mark_def_dom_walker): New class.
8270 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
8271 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
8272 (mark_def_sites_blocks): Convert to
8273 mark_def_dom_walker::before_dom_children.
8274 (mark_def_site_blocks): Remove.
8275 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
8276 (tree_ssa_dominator_optimize): Adjust.
8277 (dom_thread_across_edge): Convert to method
8278 dom_opt_dom_walker::thread_across_edge.
8279 (dom_opt_enter_block): Convert to member function
8280 dom_opt_dom_walker::before_dom_children.
8281 (dom_opt_leave_block): Convert to member function
8282 dom_opt_dom_walker::after_dom_children.
8283 * tree-ssa-dse.c (dse_dom_walker): New class.
8284 (dse_enter_block): Convert to member function
8285 dse_dom_walker::before_dom_children.
8286 (tree_ssa_dse): Adjust.
8287 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
8288 (determine_invariantness_stmt): Convert to method
8289 invariantness_dom_walker::before_dom_children.
8290 (determine_invariantness): Remove
8291 (move_computations_dom_walker): New class.
8292 (move_computations_stmt): Convert to method
8293 move_computations_dom_walker::before_dom_children.
8294 (move_computations, tree_ssa_lim): Adjust.
8295 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
8296 (nt_init_block): Convert to method
8297 notrappping_dom_walker::before_dom_children.
8298 (nt_fini_block): Convert to method
8299 method nontrapping_dom_walker::after_dom_children.
8300 (get_non_trapping): Adjust.
8301 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8302 (eliminate_bb): Convert to method
8303 eliminate_dom_walker::before_dom_children.
8304 (eliminate_leave_block): Convert to method
8305 eliminate_dom_walker::after_dom_children.
8306 (eliminate): Adjust.
8307 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8308 (strlen_enter_block): Convert to method
8309 strlen_dom_walker::before_dom_children.
8310 (strlen_leave_block): Convert to method
8311 method strlen_dom_walker::after_dom_children.
8312 (tree_ssa_strlen): Adjust.
8313 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8314 (tree_ssa_uncprop): Adjust.
8315 (uncprop_leave_block): Convert to method
8316 uncprop_dom_walker::after_dom_children.
8317 (uncprop_leave_block): Convert to method
8318 uncprop_dom_walker::before_dom_children.
8320 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8322 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8323 branch in same basic block. Check both src and dest of the move insn.
8325 2013-09-17 Nick Clifton <nickc@redhat.com>
8327 * config/rl78/rl78-real.md (bf): New pattern.
8329 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8330 (rl78_print_operand): Do not put a # before a %B.
8331 * config/rl78/rl78.opt: Tweak doc strings.
8333 2013-09-17 DJ Delorie <dj@redhat.com>
8335 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8336 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8337 (rl78_option_override): Likewise, if -mallregs.
8338 (is_virtual_register): Likewise.
8339 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8340 (REGNO_OK_FOR_BASE_P): Likewise.
8341 * config/rl78/rl78.opt (-mallregs): New.
8343 2013-09-17 Nick Clifton <nickc@redhat.com>
8345 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8346 For interrupt functions: save all call clobbered registers if the
8347 interrupt handler is not a leaf function.
8348 (rl78_expand_prologue): Always recompute the frame information.
8349 For interrupt functions: only select bank 0 if one of the bank 0
8350 registers is going to be psuhed.
8352 2013-09-17 DJ Delorie <dj@redhat.com>
8354 * config/rl78/constraints.md: For each W* constraint, rename to C*
8355 and create a W* constraint that checks for an optional ES: prefix
8357 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8358 (es_addr): New. Used to wrap far addresses.
8359 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8360 (rl78_es_base): New.
8361 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8362 wrapped far addresses.
8363 (rl78_print_operand_1): Unwrap far addresses before processing.
8364 (rl78_lo16): Wrap far addresses in unspecs.
8365 (rl78_es_addr): New.
8366 (rl78_es_base): New.
8367 (insn_ok_now): Check for not-yet-wrapped far addresses.
8368 (transcode_memory_rtx): Properly re-wrap far addresses.
8370 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8372 * config/sparc/t-rtems: Add leon3 multilibs.
8374 2013-09-17 Cong Hou <congh@google.com>
8376 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8377 when checking the dot production pattern. The type of rhs operand
8378 of multiply is now checked correctly.
8380 2013-09-17 Jeff Law <law@redhat.com>
8382 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8383 edge implied equivalences into successor phis.
8384 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8385 here from tree-ssa-threadedge.c.
8386 (mark_threaded_blocks): When threading through a joiner, if both
8387 successors of the joiner's clone reach the same block, verify the
8388 PHI arguments are equal. If not, cancel the jump threading request.
8389 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8390 tree-ssa-threadupdate.c
8391 (thread_across_edge): Don't check PHI argument equality when
8392 threading through joiner block here.
8394 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8396 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8397 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8398 (warn_uninit, warn_uninitialized_vars,
8399 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8400 Move to tree-ssa-uninit.c.
8401 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8402 (has_undefined_value_p): New. Pass dependant parts of
8403 ssa_undefined_value_p.
8404 (uninit_undefined_value_p): Use has_undefined_value_p.
8405 (warn_uninit, warn_uninitialized_vars,
8406 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8407 Move from tree-ssa.c.
8408 * tree-ssa.h: Adjust prototypes.
8410 2013-09-17 Jan Hubicka <jh@suse.cz>
8413 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8414 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8415 FUNCTION_NOT_OPTIMIZED.
8416 * ipa-inline-analysis.c (compute_inline_parameters): Function
8417 not optimized is not inlinable unless it is alwaysinline.
8418 (inline_analyze_function): Force calls in not optimized
8419 function not inlinable.
8421 2013-09-17 Jan Hubicka <jh@suse.cz>
8424 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8426 * ipa.c (function_and_variable_visibility): Likewise.
8427 * ipa-profile.c (ipa_profile): Likewise.
8429 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8432 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8433 attribute, and don't inline the resolver in this case.
8435 2013-09-17 Teresa Johnson <tejohnson@google.com>
8437 * coverage.c (get_coverage_counts): Add missing newline.
8439 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8441 PR tree-optimization/58088
8442 * fold-const.c (mask_with_trailing_zeros): New function.
8443 (fold_binary_loc): Make sure we don't recurse infinitely
8444 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8445 Use mask_with_trailing_zeros where appropriate.
8447 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8449 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8450 of 'prev' var to get better distance estimation.
8452 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8454 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8455 * tree-inline.c (is_parameter_of): New predicate.
8456 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8457 a parameter has been remapped.
8458 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8459 (optimize_inline_calls): Initialize transform_parameter.
8460 (copy_gimple_seq_and_replace_locals): Likewise.
8461 (tree_function_versioning): Likewise.
8462 (maybe_inline_call_in_expr): Likewise.
8464 2013-09-17 Nick Clifton <nickc@redhat.com>
8466 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8467 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8469 (is_attr_func): New function.
8470 (msp430_is_interrupt_func): New function.
8471 (is_naked_func): New function.
8472 (is_reentrant_func): New function.
8473 (is_critical_func): New function.
8474 (msp430_start_function): Add annotations for function attributes.
8475 (msp430_attr): New function.
8476 (msp430_attribute_table): New.
8477 (msp430_function_section): New function.
8478 (TARGET_ASM_FUNCTION_SECTION): Define.
8479 (msp430_builtin): New enum.
8480 (msp430_init_builtins): New function.
8481 (msp430_builtin_devl): New function.
8482 (msp430_expand_builtin): New function.
8483 (TARGET_INIT_BUILTINS): Define.
8484 (TARGET_EXPAND_BUILTINS): Define.
8485 (TARGET_BUILTIN_DECL): Define.
8486 (msp430_expand_prologue): Add support for naked, interrupt,
8487 critical and reentrant functions.
8488 (msp430_expand_epilogue): Likewise.
8489 (msp430_print_operand): Handle 'O' character.
8490 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8492 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8493 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8494 (pushm): Use a 'n' rather than an 'i' constraint.
8495 (msp_return): Add generation of the interrupt return instruction.
8496 (disable_interrupts): New pattern.
8497 (enable_interrupts): New pattern.
8498 (push_intr_state): New pattern.
8499 (pop_intr_state): New pattern.
8500 (bic_SR): New pattern.
8501 (bis_SR): New pattern.
8502 * doc/extend.texi: Document MSP430 function attributes and builtin
8505 2013-09-17 Richard Biener <rguenther@suse.de>
8507 PR tree-optimization/58432
8508 * tree-loop-distribution.c (tree_loop_distribution): Also
8509 scan PHIs for outside loop uses and seed a partition from them.
8511 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8513 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8514 (restructure_reference): Call backtrace_base_for_ref.
8516 2013-09-17 Alan Modra <amodra@gmail.com>
8519 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8522 2013-09-16 DJ Delorie <dj@redhat.com>
8524 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8525 vregs location for RL78/G10.
8526 (rl78_expand_prologue): Avoid SEL on G10.
8527 (rl78_expand_epilogue): Likewise.
8528 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8529 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8530 __RL78_G10__ when appropriate.
8531 (ASM_SPEC): Pass -mg10 along to the assembler.
8532 * config/rl78/rl78.md (sel_rb): Disable for G10.
8533 * config/rl78/rl78.opt: Add -mg10 option.
8534 * config/rl78/t-rl78: Add -mg10 multilib.
8536 2013-09-16 Xinliang David Li <davidxl@google.com>
8538 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8539 * omp-low.c (omp_max_vf): Ditto.
8540 (expand_omp_simd): Ditto.
8541 * tree-vectorizer.c (vectorize_loops): Ditto.
8542 (gate_vect_slp): Ditto.
8543 (gate_increase_alignment): Ditto.
8544 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8545 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8546 (gate_tree_vectorize): Name change.
8547 (tree_vectorize): Ditto.
8548 (pass_vectorize::gate): Call new function.
8549 (pass_vectorize::execute): Ditto.
8550 * opts.c: O3 default setting change.
8551 (finish_options): Check new flag.
8552 * doc/invoke.texi: Document new flags.
8553 * common.opt: New flags.
8555 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8557 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8558 (Condition Code): Don't reference it.
8559 * doc/tm.texi: Regenerate.
8561 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8564 * lra-constraints.c (undo_optional_reloads): Consider all optional
8565 reload even if it did not get a hard reg.
8567 2013-09-16 Teresa Johnson <tejohnson@google.com>
8569 * dumpfile.c (dump_loc): Remove newline emission.
8570 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8571 emission to dump_printf_loc calls where missing.
8572 (vect_mark_for_runtime_alias_test): Ditto.
8573 (vect_analyze_data_ref_dependence): Ditto.
8574 (vect_analyze_data_ref_dependences): Ditto.
8575 (vect_slp_analyze_data_ref_dependence): Ditto.
8576 (vect_slp_analyze_data_ref_dependences): Ditto.
8577 (vect_compute_data_ref_alignment): Ditto.
8578 (vect_update_misalignment_for_peel): Ditto.
8579 (vect_verify_datarefs_alignment): Ditto.
8580 (vector_alignment_reachable_p): Ditto.
8581 (vect_get_data_access_cost): Ditto.
8582 (vect_enhance_data_refs_alignment): Ditto.
8583 (vect_find_same_alignment_drs): Ditto.
8584 (vect_analyze_data_refs_alignment): Ditto.
8585 (vect_analyze_group_access): Ditto.
8586 (vect_analyze_data_ref_access): Ditto.
8587 (vect_analyze_data_ref_accesses): Ditto.
8588 (vect_prune_runtime_alias_test_list): Ditto.
8589 (vect_analyze_data_refs): Ditto.
8590 (vect_create_addr_base_for_vector_ref): Ditto.
8591 (vect_create_data_ref_ptr): Ditto.
8592 (vect_grouped_store_supported): Ditto.
8593 (vect_grouped_load_supported): Ditto.
8594 * value-prof.c (check_counter): Ditto.
8595 (check_ic_target): Ditto.
8596 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8597 (vect_recog_widen_mult_pattern): Ditto.
8598 (vect_recog_widen_sum_pattern): Ditto.
8599 (vect_recog_over_widening_pattern): Ditto.
8600 (vect_recog_widen_shift_pattern): Ditto.
8601 (vect_recog_rotate_pattern): Ditto.
8602 (vect_recog_vector_vector_shift_pattern): Ditto.
8603 (vect_recog_divmod_pattern): Ditto.
8604 (vect_recog_mixed_size_cond_pattern): Ditto.
8605 (vect_recog_bool_pattern): Ditto.
8606 (vect_pattern_recog_1): Ditto.
8607 (vect_pattern_recog): Ditto.
8608 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8609 (vect_is_simple_iv_evolution): Ditto.
8610 (vect_analyze_scalar_cycles_1): Ditto.
8611 (vect_get_loop_niters): Ditto.
8612 (vect_analyze_loop_1): Ditto.
8613 (vect_analyze_loop_form): Ditto.
8614 (vect_analyze_loop_operations): Ditto.
8615 (vect_analyze_loop_2): Ditto.
8616 (vect_analyze_loop): Ditto.
8617 (report_vect_op): Ditto.
8618 (vect_is_slp_reduction): Ditto.
8619 (vect_is_simple_reduction_1): Ditto.
8620 (vect_get_known_peeling_cost): Ditto.
8621 (vect_estimate_min_profitable_iters): Ditto.
8622 (vect_model_reduction_cost): Ditto.
8623 (vect_model_induction_cost): Ditto.
8624 (get_initial_def_for_induction): Ditto.
8625 (vect_create_epilog_for_reduction): Ditto.
8626 (vectorizable_reduction): Ditto.
8627 (vectorizable_induction): Ditto.
8628 (vectorizable_live_operation): Ditto.
8629 (vect_loop_kill_debug_uses): Ditto.
8630 (vect_transform_loop): Ditto.
8631 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8632 (vect_stmt_relevant_p): Ditto.
8633 (process_use): Ditto.
8634 (vect_mark_stmts_to_be_vectorized): Ditto.
8635 (vect_model_simple_cost): Ditto.
8636 (vect_model_promotion_demotion_cost): Ditto.
8637 (vect_model_store_cost): Ditto.
8638 (vect_get_store_cost): Ditto.
8639 (vect_model_load_cost): Ditto.
8640 (vect_get_load_cost): Ditto.
8641 (vect_init_vector_1): Ditto.
8642 (vect_get_vec_def_for_operand): Ditto.
8643 (vect_finish_stmt_generation): Ditto.
8644 (vectorizable_call): Ditto.
8645 (vectorizable_conversion): Ditto.
8646 (vectorizable_assignment): Ditto.
8647 (vectorizable_shift): Ditto.
8648 (vectorizable_operation): Ditto.
8649 (vectorizable_store): Ditto.
8650 (vectorizable_load): Ditto.
8651 (vectorizable_condition): Ditto.
8652 (vect_analyze_stmt): Ditto.
8653 (vect_transform_stmt): Ditto.
8654 (vect_is_simple_use): Ditto.
8655 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8656 (vect_can_advance_ivs_p): Ditto.
8657 (vect_update_ivs_after_vectorizer): Ditto.
8658 (vect_do_peeling_for_loop_bound): Ditto.
8659 (vect_gen_niters_for_prolog_loop): Ditto.
8660 (vect_update_inits_of_drs): Ditto.
8661 (vect_create_cond_for_alias_checks): Ditto.
8662 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8663 (vect_build_slp_tree_1): Ditto.
8664 (vect_supported_load_permutation_p): Ditto.
8665 (vect_analyze_slp_instance): Ditto.
8666 (vect_analyze_slp): Ditto.
8667 (vect_make_slp_decision): Ditto.
8668 (vect_detect_hybrid_slp): Ditto.
8669 (vect_bb_vectorization_profitable_p): Ditto.
8670 (vect_slp_analyze_bb_1): Ditto.
8671 (vect_update_slp_costs_according_to_vf): Ditto.
8672 (vect_get_mask_element): Ditto.
8673 (vect_transform_slp_perm_load): Ditto.
8674 (vect_schedule_slp_instance): Ditto.
8675 (vect_schedule_slp): Ditto.
8676 (vect_slp_transform_bb): Ditto.
8677 * profile.c (read_profile_edge_counts): Ditto.
8678 (compute_branch_probabilities): Ditto.
8679 * coverage.c (get_coverage_counts): Ditto.
8681 2013-09-16 Diego Novillo <dnovillo@google.com>
8683 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8685 2013-09-16 Jan Hubicka <jh@suse.cz>
8687 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8688 abstract functions; for static functions check the presence of body.
8690 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8692 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8693 * config/aarch64/aarch64-simd.md
8694 (aarch64_mla_elt<mode>): New.
8695 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8696 (aarch64_mls_elt<mode>): Likewise.
8697 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8698 (aarch64_fma4_elt<mode>): Likewise.
8699 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8700 (aarch64_fma4_elt_to_128v2df): Likewise.
8701 (aarch64_fma4_elt_to_64df): Likewise.
8702 (fnma<mode>4): Likewise.
8703 (aarch64_fnma4_elt<mode>): Likewise.
8704 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8705 (aarch64_fnma4_elt_to_128v2df): Likewise.
8706 (aarch64_fnma4_elt_to_64df): Likewise.
8707 * config/aarch64/iterators.md (VDQSF): New.
8708 * config/aarch64/arm_neon.h
8709 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8710 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8712 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8714 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8715 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8716 (aarch64_mul3_elt_to_128df): Likewise.
8717 (aarch64_mul3_elt_to_64v2df): Likewise.
8718 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8720 (VMUL_CHANGE_NLANES) Likewise.
8723 * config/aarch64/arm_neon.h
8724 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8726 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8728 * config/aarch64/arm_neon.h
8729 (vcvtx_high_f32_f64): Fix parameters.
8731 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8732 Uros Bizjak <ubizjak@gmail.com>
8734 * config/alpha.c: Include tree-ssanames.h.
8736 2013-09-16 Richard Biener <rguenther@suse.de>
8738 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8739 (dot_rdg_1): Handle control_dd.
8740 (create_edge_for_control_dependence): New function.
8741 (create_rdg_edges): Add control dependences if asked for.
8742 (build_rdg): Likewise.
8743 (generate_loops_for_partition): If there are not necessary
8744 control stmts remove all their dependencies.
8745 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8746 (distribute_loop): Pass on control dependences.
8747 (tree_loop_distribution): Compute control dependences and remove
8748 restriction on number of loop nodes.
8750 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8752 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8755 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8757 * cse.c (try_const_anchors): Punt on CC modes.
8759 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8761 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8763 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8766 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8769 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8772 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8774 (darwin_objc1_section): Likewise.
8775 (darwin_file_end): Emit Image Info section when required.
8777 2013-09-14 Jan Hubicka <jh@suse.cz>
8779 * tree-into-ssa.c (gate_into_ssa): New.
8780 (pass_data_build_ssa): Use it.
8781 * cgraph.h (expand_thunk): Update prototype.
8782 * cgraphunit.c (analyze_function): Expand thunks early.
8783 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8784 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8785 set lowered flag; do not add new function.
8786 (assemble_thunks_and_aliases): Update.
8787 * tree-ssa.c (gate_init_datastructures): New gate.
8788 (pass_data_init_datastructures): Use it.
8790 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8793 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8794 xmm register set described in the psABI.
8796 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8798 * dwarf2out.c (should_emit_struct_debug): Add check
8799 for type_decl variable is not NULL.
8801 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8803 * config.gcc: Use new winnt-c.c target hooks
8804 * config/t-winnt: New file
8805 * config/winnt-c.c: New file
8806 * doc/tm.texi.in: Document new hook
8807 * doc/tm.texi: Regenerated
8809 2013-09-13 Jan Hubicka <jh@suse.cz>
8812 * ipa-inline.c (check_callers): New function.
8813 (check_caller_edge): Remove.
8814 (want_inline_function_to_all_callers_p): Also permit alises that are
8816 (inline_to_all_callers): Terminate the walk when devirtualization
8817 introduce new calls.
8819 2013-09-13 Jan Hubicka <jh@suse.cz>
8821 * ipa-inline-analysis.c (struct growth_data): Add node.
8822 (do_estimate_growth_1): Fix detection of recursion.
8824 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8826 PR tree-optimization/58392
8827 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8828 to avoid shadowing of outer loop variable. If
8829 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8830 replace_by_duplicate_decl simduid of loops that have it set and
8831 set dest_cfun->has_simduid_loops and/or
8832 dest_cfun->has_force_vect_loops.
8833 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8834 instead of maybe_lookup_decl.
8835 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8836 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8837 call. Copy over force_vect and copy and remap simduid. Set
8838 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8839 (copy_cfg_body): Remove blocks_to_copy argument. Use
8840 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8841 caller. Don't set cfun->has_simduid_loops and/or
8842 cfun->has_force_vect_loops here.
8843 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8845 (expand_call_inline, tree_function_versioning): Adjust copy_body
8848 2013-09-13 Martin Jambor <mjambor@suse.cz>
8851 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8852 the assert if the edge was a speculative one.
8854 2013-09-13 Richard Biener <rguenther@suse.de>
8856 * tree-data-ref.h (known_dependences_p): Move here ...
8857 * tree-loop-distribution.c (known_dependences_p): ... from here.
8858 (dump_rdg_component, debug_rdg_component): Remove.
8860 (generate_loops_for_partition): Use gimple_uid instead of
8861 relying on matching stmt visit order.
8862 (rdg_build_partitions): Take starting stmt vector.
8863 (ldist_gen): Merge into ...
8864 (distribute_loop): ... this function. Do not compute starting
8866 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8868 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8870 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8871 Set type attribute correctly. Set predicable_short_it attribute.
8872 (cmpsi_shiftsi): Remove %? from output template.
8874 2013-09-13 Richard Biener <rguenther@suse.de>
8876 * tree-loop-distribution.c (struct rdg_component,
8877 rdg_defs_used_in_other_loops_p, free_rdg_components,
8878 rdg_build_components): Remove.
8879 (stmts_from_loop): Do not record virtual PHIs.
8880 (generate_loops_for_partition): Skip virtual PHIs.
8881 (build_rdg_partition_for_component): Rename to ...
8882 (build_rdg_partition_for_vertex): ... this and adjust.
8883 (rdg_build_partitions): Take a vector of starting vertices
8884 instead of components. Remove unnecessary leftover handling.
8885 (ldist_gen): Do not build components or record other stores.
8886 (distribute_loop): Do not distribute loops containing stmts
8889 2013-09-13 Christian Bruel <christian.bruel@st.com>
8892 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8894 2013-09-13 Kai Tietz <ktietz@redhat.com>
8896 * config.gcc: Separate cases for mingw and cygwin targets,
8897 and add 64-bit cygwin target case.
8899 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8900 dll-export inline-functions.
8901 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8903 2013-09-13 Jeff Law <law@redhat.com>
8907 2013-09-06 Jeff Law <law@redhat.com>
8909 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8910 edge implied equivalences into successor phis.
8912 2013-09-12 DJ Delorie <dj@redhat.com>
8914 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8917 2013-09-12 Brooks Moses <bmoses@google.com>
8920 * Makefile.in: Do not install driver binaries in $(target)/bin.
8922 2013-09-12 DJ Delorie <dj@redhat.com>
8924 * config/rl78/rl78.opt (mrelax): New.
8925 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8926 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8928 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8930 (rl78_expand_epilogue): Likewise.
8932 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8935 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8936 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8937 pointer elimination with using remove_reg_equal_offset_note.
8939 2013-09-12 DJ Delorie <dj@redhat.com>
8941 * config/msp430/: New port.
8942 * config.gcc (msp430): Added.
8943 * doc/invoke.texi: Document MSP430 options.
8944 * doc/install.texi: Document msp430-elf
8945 * doc/md.texi: Document msp430-elf
8946 * doc/contrib.texi: Document msp430-elf
8948 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8951 2013-09-12 Martin Jambor <mjambor@suse.cz>
8954 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8955 reference descriptor is NULL.
8956 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8959 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8961 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8962 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8963 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8964 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8965 make_temp_ssa_name): move to tree-ssanames.h
8966 * tree-ssa-alias.h: Move prototype.
8967 * tree-ssa.h: Include tree-ssanames.h.
8968 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8969 * tree-ssanames.h: New. Move items from tree-flow*.h
8970 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8972 2013-09-12 Richard Biener <rguenther@suse.de>
8974 PR tree-optimization/58404
8975 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8976 propagate non-invariant addresses into dereferences wrapped
8977 in component references.
8979 2013-09-12 Richard Biener <rguenther@suse.de>
8981 PR tree-optimization/58402
8982 * passes.def: Move pass_late_warn_uninitialized later.
8984 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8986 * tree-ssa.h: New. Move content from tree-flow.h and
8988 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8989 Move prototypes belonging to tree-ssa.c.
8990 * tree-flow-inline.h (redirect_edge_var_map_def,
8991 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8993 * gimple.h: Adjust prototypes.
8994 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8996 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8997 * tree.h: Move prototype to tree-ssa.h.
8998 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8999 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
9000 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
9001 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
9002 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
9003 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
9004 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
9005 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
9006 graphite-blocking.c, graphite-clast-to-gimple.c,
9007 graphite-dependences.c, graphite-interchange.c,
9008 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
9009 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
9010 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
9011 ipa-reference.c, ipa-split.c, ipa-utils.c,
9012 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
9013 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
9014 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
9015 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
9016 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
9017 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
9018 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
9019 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
9020 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
9021 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
9022 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
9023 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9024 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
9025 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
9026 value-prof.c, var-tracking.c,
9027 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
9029 2013-09-12 Richard Biener <rguenther@suse.de>
9031 PR tree-optimization/58396
9032 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
9033 (build_rdg): Take a loop-nest parameter, fix memleaks.
9034 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
9036 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9038 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
9041 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
9043 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
9044 The FMA instruction names should have a 'v' prefix.
9046 2013-09-12 Richard Biener <rguenther@suse.de>
9048 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
9049 (dot_rdg): Use popen instead of system in optional code.
9050 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
9051 (already_processed_vertex_p): Adjust.
9052 (has_anti_or_output_dependence, predecessor_has_mem_write,
9053 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
9054 rdg_flag_uses): Remove.
9055 (rdg_flag_vertex): Simplify.
9056 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
9058 (build_rdg_partition_for_component): Process the first vertex
9059 of a component only.
9060 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
9062 2013-09-12 Alan Modra <amodra@gmail.com>
9064 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
9066 2013-09-11 DJ Delorie <dj@redhat.com>
9067 Nick Clifton <nickc@redhat.com>
9069 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
9070 (rl78_stack_based_mem): New.
9071 * config/rl78/constraints.md (Iv08): New.
9080 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
9082 (movsi): Change from expand to insn-and-split.
9083 (ashrsi3): Clobber AX.
9087 * config/rl78/rl78.md (CC_REG): Fix.
9088 (addsi3): Allow memory and immediate operands.
9089 (addsi3_internal): Split into...
9090 (addsi3_internal_virt): ...new, and ...
9091 (addsi3_internal_real): ...new.
9093 (subsi3_internal_virt): New.
9094 (subsi3_internal_real): New.
9095 (mulsi3): Add memory operand.
9096 (mulsi3_rl78): Likewise.
9097 (mulsi3_g13): Likewise.
9098 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
9099 (cbranchqi4_real): Add more constraint options.
9100 (cbranchhi4_real): Expand pattern.
9101 (cbranchhi4_real_signed): New.
9102 (cbranchhi4_real_inverted): New.
9103 (cbranchsi4_real_lt): New.
9104 (cbranchsi4_real_ge): New.
9105 (cbranchsi4_real_signed): New.
9106 (cbranchsi4_real): New.
9108 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
9110 (lshrsi3_virt): Likewise.
9111 (ashlsi3_virt): Likewise.
9112 (cbranchqi4_virt_signed): New.
9113 (cbranchhi4_virt_signed): New.
9114 (cbranchsi4_virt): New.
9115 * config/rl78/rl78.c: Whitespace fixes throughout.
9116 (move_elim_pass): New.
9117 (pass_data_rl78_move_elim): New.
9118 (pass_rl78_move_elim): New.
9119 (make_pass_rl78_move_elim): New.
9120 (rl78_devirt_info): Run devirt earlier.
9121 (rl78_move_elim_info): New.
9122 (rl78_asm_file_start): Register it.
9123 (rl78_split_movsi): New.
9124 (rl78_as_legitimate_address): Allow virtual base registers when
9126 (rl78_addr_space_convert): Remove spurious debug stuff.
9127 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
9128 (rl78_print_operand): More cases for not printing '#'.
9129 (rl78_expand_compare): Remove most of the logic.
9130 (content_memory): New.
9131 (clear_content_memory): New.
9132 (get_content_index): New.
9133 (get_content_name): New.
9134 (display_content_memory): New.
9135 (update_content): New.
9136 (record_content): New.
9137 (already_contains): New.
9138 (insn_ok_now): Re-recog insns with virtual registers.
9139 (add_postponed_content_update): New.
9140 (process_postponed_content_update): New.
9141 (gen_and_emit_move): New.
9142 (transcode_memory_rtx): Record new location content.
9143 Use gen_and_emit_move.
9144 (force_into_acc): New.
9145 (move_to_acc): Use gen_and_emit_move.
9146 (move_from_acc): Likewise.
9147 (move_acc_to_reg): Likewise.
9148 (move_to_x): Likewise.
9149 (move_to_hl): Likewise.
9150 (move_to_de): Likewise.
9151 (rl78_alloc_physical_registers_op1): Record location content.
9152 (has_constraint): New.
9153 (rl78_alloc_physical_registers_op2): Record location content.
9155 (rl78_alloc_physical_registers_ro1): Likewise.
9156 (rl78_alloc_physical_registers_cmp): Likewise.
9157 (rl78_alloc_physical_registers_umul): Likewise.
9158 (rl78_alloc_address_registers_macax): New.
9159 (rl78_alloc_physical_registers): Initialize and set location
9160 content memory as needed.
9161 (rl78_reorg): Make sure split2 is called.
9162 (rl78_rtx_costs): New.
9164 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
9166 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
9167 for (not (neg ...)) and (neg (not ...)) cases.
9169 2013-09-11 Richard Biener <rguenther@suse.de>
9172 * passes.def: Split critical edges before late uninit warning passes.
9173 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
9175 2013-09-11 Jakub Jelinek <jakub@redhat.com>
9177 PR tree-optimization/58385
9178 * fold-const.c (build_range_check): If both low and high are NULL,
9179 use omit_one_operand_loc to preserve exp side-effects.
9181 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9183 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
9185 2013-09-11 Richard Biener <rguenther@suse.de>
9187 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
9188 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
9189 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
9190 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
9191 stmts_from_loop, known_dependences_p, build_empty_rdg,
9192 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
9193 * tree-loop-distribution.c: ... here.
9194 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
9195 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
9196 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
9197 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
9198 * tree-loop-distribution.c: ... here.
9199 * tree-loop-distribution.c: Include gimple-pretty-print.h.
9200 (struct partition_s): Add loops member.
9201 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
9202 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
9203 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
9205 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9206 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9207 Sergey Lega <sergey.s.lega@intel.com>
9208 Anna Tikhonova <anna.tikhonova@intel.com>
9209 Ilya Tocar <ilya.tocar@intel.com>
9210 Andrey Turetskiy <andrey.turetskiy@intel.com>
9211 Ilya Verbin <ilya.verbin@intel.com>
9212 Kirill Yukhin <kirill.yukhin@intel.com>
9213 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9215 * config/i386/constraints.md (k): New.
9217 * config/i386/i386.c (const regclass_map): Add new mask registers.
9218 (dbx_register_map): Ditto.
9219 (dbx64_register_map): Ditto.
9220 (svr4_dbx_register_map): Ditto.
9221 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
9223 (ix86_preferred_reload_class): Disable constants for mask registers.
9224 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
9225 (ix86_hard_regno_mode_ok): Support new mask registers.
9226 (x86_order_regs_for_local_alloc): Ditto.
9227 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
9228 (FIXED_REGISTERS): Add new mask registers.
9229 (CALL_USED_REGISTERS): Ditto.
9230 (REG_ALLOC_ORDER): Ditto.
9231 (VALID_MASK_REG_MODE): New.
9232 (FIRST_MASK_REG): Ditto.
9233 (LAST_MASK_REG): Ditto.
9234 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
9235 (MAYBE_MASK_CLASS_P): New.
9236 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
9237 (REG_CLASS_CONTENTS): Ditto.
9238 (MASK_REGNO_P): New.
9239 (ANY_MASK_REG_P): Ditto.
9240 (HI_REGISTER_NAMES): Add new mask registers.
9241 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
9242 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
9244 (attribute "type"): Add mskmov, msklog.
9245 (attribute "length_immediate"): Support them.
9246 (attribute "memory"): Ditto.
9247 (attribute "prefix_0f"): Ditto.
9248 (*movhi_internal): Support new mask registers.
9249 (*movqi_internal): Ditto.
9250 (define_split): Split out clobber pattern is a logic
9251 insn on mask registers.
9252 (*k<logic><mode>): New.
9253 (*andhi_1): Extend to support mask regs.
9254 (*andqi_1): Extend to support mask regs.
9256 (define_split): Split and-not to and and not if operands
9258 (*<code><mode>_1): Separate HI mode to new pattern...
9259 (*<code>hi_1): This.
9260 (*<code>qi_1): Extend to support mask regs.
9262 (kortestzhi): Ditto.
9263 (kortestchi): Ditto.
9265 (*one_cmpl<mode>2_1): Remove HImode and handle it...
9266 (*one_cmplhi2_1): ...Here, now with mask registers support.
9267 (*one_cmplqi2_1): Support new mask registers.
9268 (HI/QImode arithmetics splitter): Don't split if mask registers
9270 (HI/QImode not splitter): Ditto.
9271 * config/i386/predicated.md (mask_reg_operand): New.
9272 (general_reg_operand): Ditto.
9274 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9276 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
9277 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
9279 2013-09-10 Jeff Law <law@redhat.com>
9281 PR tree-optimization/58380
9282 * tree-ssa-threadupdate.c (thread_block): Recognize another case
9283 of threading through a buried loop header.
9285 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
9286 return value for single successor case.
9288 2013-09-10 Jan Hubicka <jh@suse.cz>
9290 * ipa-devirt.c (ipa_devirt): Enable with LTO.
9292 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
9295 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
9296 support conditional execution.
9297 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
9299 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9301 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9303 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9304 Check destination too to check move insn.
9305 (undo_optional_reloads): Add check that the original peudo did not
9306 changed its allocation and the optional reload was inherited on last
9307 inheritance pass. Break loop after deciding to keep optional reload.
9308 (lra_undo_inheritance): Add check that inherited pseudo still in
9311 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9313 * config/aarch64/aarch64.md (generic_sched): New.
9314 * config/aarch64/aarch64-generic.md (load): Make conditional
9315 on generic_sched attribute.
9316 (nonload): Likewise.
9318 2013-09-10 Jan Hubicka <jh@suse.cz>
9320 * lto-cgraph.c: Include ipa-utils.h.
9321 (compute_ltrans_boundary): Also add possible targets into the boundary.
9323 2013-09-10 Jan Hubicka <jh@suse.cz>
9325 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9326 VAR_DECL of vtable rather than full expression.
9328 2013-09-10 Jan Hubicka <jh@suse.cz>
9329 Paolo Carlini <paolo.carlini@oracle.com>
9331 * cgraphunit.c (analyze_functions): Save input_location, set it
9332 to UNKNOWN_LOCATION and restore it at the end.
9334 2013-09-10 Martin Jambor <mjambor@suse.cz>
9336 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9337 represented by a thunk.
9339 2013-09-10 Jeff Law <law@redhat.com>
9341 PR tree-optimization/58343
9342 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9343 jump threading requests through loop headers buried in the middle
9344 of a jump threading path.
9346 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9347 in return value/type.
9349 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9351 PR rtl-optimization/58365
9352 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9353 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9356 2013-09-10 Richard Biener <rguenther@suse.de>
9358 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9359 * tree-data-ref.c (create_rdg_vertices): Collect all data
9360 references, signal failure to the caller, use data-ref API.
9361 (build_rdg): Compute data references only once. Maintain lifetime
9362 of data references and data dependences from within RDG.
9363 (free_rdg): Free dependence relations.
9364 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9365 inventing extra dependences.
9366 (distribute_loop): Update for RDG API changes.
9368 2013-09-10 Kai Tietz <ktietz@redhat.com>
9370 * doc/invoke.texi (fms-extensions): Document changed
9371 behavior for ms-abi targets.
9372 * config/i386/i386.c (ix86_option_override_internal):
9373 Set default value of option -fms-extension for ms-abi targets.
9375 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9377 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9379 2013-09-10 Alan Modra <amodra@gmail.com>
9382 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9384 2013-09-10 Alan Modra <amodra@gmail.com>
9386 * config/rs6000/predicates.md (add_cint_operand): New.
9387 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9388 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9389 using add_cint_operand.
9390 (largetoc_high_plus_aix): Likewise.
9392 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9394 PR tree-optimization/58364
9395 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9396 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9397 the current range can't be an unconditional true or false.
9399 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9401 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9403 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9405 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9407 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9410 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9412 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9414 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9415 NO_REGS for immediate that can't be moved directly into FP_REGS.
9417 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9419 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9420 comparison with negated operand.
9421 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9424 2013-09-09 Richard Biener <rguenther@suse.de>
9427 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9428 of a subloop record all its block as affecting loop-closed SSA form.
9430 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9432 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9433 of an rtx/bitpos pair.
9434 (store_fixed_bit_field): Update accordingly.
9436 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9438 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9440 * builtins.c (expand_errno_check): Likewise.
9441 * dwarf2cfi.c (init_return_column_size): Likewise.
9442 * except.c (sjlj_mark_call_sites): Likewise.
9443 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9444 * lra-constraints.c (emit_inc): Likewise.
9445 * ree.c (combine_set_extension): Likewise.
9446 * regmove.c (fixup_match_2): Likewise.
9447 * reload1.c (inc_for_reload): Likewise.
9449 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9451 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9452 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9453 Use gen_int_mode with the mode of the associated simplify_* call.
9454 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9455 * expmed.c (expand_shift_1): Likewise.
9456 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9457 * loop-iv.c (iv_number_of_iterations): Likewise.
9458 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9459 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9461 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9463 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9464 of the associated expand_* call.
9465 (asan_emit_stack_protection): Likewise.
9466 * builtins.c (round_trampoline_addr): Likewise.
9467 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9468 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9469 (emit_store_flag): Likewise.
9470 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9471 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9473 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9474 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9475 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9476 * modulo-sched.c (generate_prolog_epilog): Likewise.
9477 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9478 (expand_ctz, expand_ffs, expand_unop): Likewise.
9480 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9482 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9483 of the associated gen_rtx_* call.
9484 * caller-save.c (init_caller_save): Likewise.
9485 * combine.c (find_split_point, make_extraction): Likewise.
9486 (make_compound_operation): Likewise.
9487 * dwarf2out.c (mem_loc_descriptor): Likewise.
9488 * explow.c (plus_constant, probe_stack_range): Likewise.
9489 * expmed.c (expand_mult_const): Likewise.
9490 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9491 * reload1.c (init_reload): Likewise.
9492 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9493 * var-tracking.c (adjust_mems): Likewise.
9494 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9495 rather than gen_rtx_fmt_ee.
9497 2013-09-09 Jan Hubicka <jh@suse.cz>
9500 * value-prof.c (gimple_ic): Copy also abnormal edges.
9502 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9504 * asan.c (asan_shadow_cst): Use gen_int_mode.
9506 2013-09-08 Jan Hubicka <jh@suse.cz>
9508 * ipa-profile.c: Add toplevel comment.
9509 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9510 (contains_hot_call_p): New function.
9511 (ipa_propagate_frequency): Set frequencies based on counts when
9513 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9514 profile; do not tamper with profile after inlining if it is read.
9516 2013-09-08 Jan Hubicka <jh@suse.cz>
9518 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9521 2013-09-08 Jan Hubicka <jh@suse.cz>
9523 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9526 2013-09-08 Jeff Law <law@redhat.com>
9529 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9532 2013-09-08 Andi Kleen <ak@linux.intel.com>
9534 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9536 2013-09-08 Jan Hubicka <jh@suse.cz>
9538 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9540 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9542 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9543 for non-debug insns.
9544 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9545 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9546 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9549 2013-09-08 Jan Hubicka <jh@suse.cz>
9551 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9552 targets and devirtualize to BUILT_IN_UNREACHABLE.
9553 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9554 * ipa.c (walk_polymorphic_call_targets): New function.
9555 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9556 functions; use the new timevar.
9557 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9558 was removed from the program.
9559 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9560 not consider it in the walk when its vtable is dead.
9561 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9563 (devirt_variable_node_removal_hook): New function.
9564 (possible_polymorphic_call_targets): Also register
9565 devirt_variable_node_removal_hook.
9566 (ipa_devirt): Do not do non-speculative devirtualization.
9567 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9569 2013-09-08 Jan Hubicka <jh@suse.cz>
9571 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9572 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9573 varpool_remove_variable_insertion_hook): Declare.
9574 * varpool.c (varpool_node_hook_list): New structure.
9575 (first_varpool_node_removal_hook,
9576 first_varpool_variable_insertion_hook): New variables.
9577 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9578 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9579 varpool_remove_variable_insertion_hook,
9580 varpool_call_variable_insertion_hooks): New functions.
9581 (varpool_remove_node): Use it.
9583 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9586 * diagnostic.c (diagnostic_build_prefix): When s.file is
9587 "<built-in>" don't output line and column numbers.
9589 2013-09-06 Jan Hubicka <jh@suse.cz>
9591 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9592 * lto-streamer-out.c: Stream thunks, too.
9593 * lto-streamer-in.c (input_function): Pop cfun here
9594 (lto_read_body): Instead of here.
9596 2013-09-06 Caroline Tice <cmtice@google.com>
9598 * doc/install.texi: Add documentation for the --enable-vtable-verify
9599 and the --disable-libvtv configure options.
9601 2013-09-06 Jeff Law <law@redhat.com>
9603 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9604 edge implied equivalences into successor phis.
9606 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9608 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9610 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9612 * resource.c (mark_target_live_regs): Compute resources taking
9613 into account if a call is predicated or not.
9615 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9617 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9618 the compiler in the function names.
9620 2013-09-06 Jan Hubicka <jh@suse.cz>
9623 * ipa-inline.c (has_caller_p): New function.
9624 (want_inline_function_to_all_callers_p): Use it.
9625 (sum_callers, inline_to_all_callers): Break out from ...
9626 (ipa_inline): ... here.
9628 2013-09-06 Jan Hubicka <jh@suse.cz>
9630 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9631 only when AVX is enabled.
9633 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9635 * config/aarch64/aarch64.md
9636 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9637 is fpsimd_<load/store>2.
9638 (load_pair<mode>): Likewise.
9639 (store_pair<mode>): Likewise.
9641 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9643 * config/arm/types.md (type): Add "mrs" type.
9644 * config/aarch64/aarch64.md
9645 (aarch64_load_tp_hard): Make type "mrs".
9647 (load_tp_hard): Make type "mrs".
9648 * config/arm/cortex-a15.md: Update with new attributes.
9649 * config/arm/cortex-a5.md: Update with new attributes.
9650 * config/arm/cortex-a53.md: Update with new attributes.
9651 * config/arm/cortex-a7.md: Update with new attributes.
9652 * config/arm/cortex-a8.md: Update with new attributes.
9653 * config/arm/cortex-a9.md: Update with new attributes.
9654 * config/arm/cortex-m4.md: Update with new attributes.
9655 * config/arm/cortex-r4.md: Update with new attributes.
9656 * config/arm/fa526.md: Update with new attributes.
9657 * config/arm/fa606te.md: Update with new attributes.
9658 * config/arm/fa626te.md: Update with new attributes.
9659 * config/arm/fa726te.md: Update with new attributes.
9661 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9663 * config/aarch64/aarch64.md
9664 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9665 (*movtf_aarch64): Likewise.
9667 (thumb1_movdi_insn): Use "multiple" for type where more than one
9668 instruction is used for a move.
9669 (*arm32_movhf): Likewise.
9670 (*thumb_movdf_insn): Likewise.
9672 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9674 * config/arm/types.md (type): Rename fcpys to fmov.
9676 (*arm_movsi_vfp): Rename type fcpys as fmov.
9677 (*thumb2_movsi_vfp): Likewise
9678 (*movhf_vfp_neon): Likewise
9679 (*movhf_vfp): Likewise
9680 (*movsf_vfp): Likewise
9681 (*thumb2_movsf_vfp): Likewise
9682 (*movsfcc_vfp): Likewise
9683 (*thumb2_movsfcc_vfp): Likewise
9684 * config/aarch64/aarch64-simd.md
9685 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9686 * config/aarch64/aarch64.md
9687 (*movsi_aarch64): Replace type mov_reg with fmovs.
9688 (*movdi_aarch64): Likewise
9689 (*movsf_aarch64): Likewise
9690 (*movdf_aarch64): Likewise
9692 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9693 * config/arm/iwmmxt.md
9694 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9695 * config/arm/arm1020e.md: Update with new attributes.
9696 * config/arm/cortex-a15-neon.md: Update with new attributes.
9697 * config/arm/cortex-a5.md: Update with new attributes.
9698 * config/arm/cortex-a53.md: Update with new attributes.
9699 * config/arm/cortex-a7.md: Update with new attributes.
9700 * config/arm/cortex-a8-neon.md: Update with new attributes.
9701 * config/arm/cortex-a9.md: Update with new attributes.
9702 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9703 * config/arm/cortex-r4f.md: Update with new attributes.
9704 * config/arm/marvell-pj4.md: Update with new attributes.
9705 * config/arm/vfp11.md: Update with new attributes.
9707 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9709 * config/aarch64/aarch64.md
9710 (*madd<mode>): Fix type attribute.
9711 (*maddsi_uxtw): Likewise.
9712 (*msub<mode>): Likewise.
9713 (*msubsi_uxtw): Likewise.
9714 (<su_optab>maddsidi4): Likewise.
9715 (<su_optab>msubsidi4): Likewise.
9717 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9719 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9720 * config/arm/arm.md (core_cycles): Remove fdiv.
9721 * config/arm/vfp.md:
9722 (*sqrtsf2_vfp): Update for attribute changes.
9723 (*sqrtdf2_vfp): Likewise.
9724 * config/aarch64/aarch64.md:
9725 (sqrt<mode>2): Update for attribute changes.
9726 * config/arm/arm1020e.md: Update with new attributes.
9727 * config/arm/cortex-a15-neon.md: Update with new attributes.
9728 * config/arm/cortex-a5.md: Update with new attributes.
9729 * config/arm/cortex-a53.md: Update with new attributes.
9730 * config/arm/cortex-a7.md: Update with new attributes.
9731 * config/arm/cortex-a8-neon.md: Update with new attributes.
9732 * config/arm/cortex-a9.md: Update with new attributes.
9733 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9734 * config/arm/cortex-r4f.md: Update with new attributes.
9735 * config/arm/marvell-pj4.md: Update with new attributes.
9736 * config/arm/vfp11.md: Update with new attributes.
9738 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9740 * config/arm/types.md
9741 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9742 * config/aarch64/aarch64.md
9743 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9745 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9746 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9747 (float<GPI:mode><GPF:mode>2): Likewise.
9749 (*truncsisf2_vfp): Update with new attributes.
9750 (*truncsidf2_vfp): Likewise.
9751 (fixuns_truncsfsi2): Likewise.
9752 (fixuns_truncdfsi2): Likewise.
9753 (*floatsisf2_vfp): Likewise.
9754 (*floatsidf2_vfp): Likewise.
9755 (floatunssisf2): Likewise.
9756 (floatunssidf2): Likewise.
9757 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9758 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9759 * config/arm/arm1020e.md: Update with new attributes.
9760 * config/arm/cortex-a15-neon.md: Update with new attributes.
9761 * config/arm/cortex-a5.md: Update with new attributes.
9762 * config/arm/cortex-a53.md: Update with new attributes.
9763 * config/arm/cortex-a7.md: Update with new attributes.
9764 * config/arm/cortex-a8-neon.md: Update with new attributes.
9765 * config/arm/cortex-a9.md: Update with new attributes.
9766 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9767 * config/arm/cortex-r4f.md: Update with new attributes.
9768 * config/arm/marvell-pj4.md: Update with new attributes.
9769 * config/arm/vfp11.md: Update with new attributes.
9771 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9773 * config/aarch64/arm_neon.h
9774 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9775 (vqtbx<1,2,3,4><q>_s8): Likewise.
9777 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9779 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9780 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9781 (add<mode>3): Add type attribute.
9782 (add<mode>3): Likewise.
9783 (usadd<mode>3): Likewise.
9784 (ssadd<mode>3): Likewise.
9785 (sub<mode>3): Likewise.
9786 (sub<mode>3): Likewise.
9787 (ussub<mode>3): Likewise.
9788 (sssub<mode>3): Likewise.
9789 (ssmulsa3): Likewise.
9790 (usmulusa3): Likewise.
9791 (arm_usatsihi): Likewise.
9793 (*movdi_vfp): Add types for all instructions.
9794 (*movdi_vfp_cortexa8): Likewise.
9795 (*movhf_vfp_neon): Likewise.
9796 (*movhf_vfp): Likewise.
9797 (*movdf_vfp): Likewise.
9798 (*thumb2_movdf_vfp): Likewise.
9799 (*thumb2_movdfcc_vfp): Likewise.
9800 * config/arm/arm.md: Add type attribute to all insn patterns.
9801 (*thumb1_adddi3): Add type attribute.
9802 (*arm_adddi3): Likewise.
9803 (*adddi_sesidi_di): Likewise.
9804 (*adddi_zesidi_di): Likewise.
9805 (*thumb1_addsi3): Likewise.
9806 (addsi3_compare0): Likewise.
9807 (*addsi3_compare0_scratch): Likewise.
9808 (*compare_negsi_si): Likewise.
9809 (cmpsi2_addneg): Likewise.
9810 (*addsi3_carryin_<optab>): Likewise.
9811 (*addsi3_carryin_alt2_<optab>): Likewise.
9812 (*addsi3_carryin_clobercc_<optab>): Likewise.
9813 (*subsi3_carryin): Likewise.
9814 (*subsi3_carryin_const): Likewise.
9815 (*subsi3_carryin_compare): Likewise.
9816 (*subsi3_carryin_compare_const): Likewise.
9817 (*arm_subdi3): Likewise.
9818 (*thumb_subdi3): Likewise.
9819 (*subdi_di_zesidi): Likewise.
9820 (*subdi_di_sesidi): Likewise.
9821 (*subdi_zesidi_di): Likewise.
9822 (*subdi_sesidi_di): Likewise.
9823 (*subdi_zesidi_ze): Likewise.
9824 (thumb1_subsi3_insn): Likewise.
9825 (*arm_subsi3_insn): Likewise.
9826 (*anddi3_insn): Likewise.
9827 (*anddi_zesidi_di): Likewise.
9828 (*anddi_sesdi_di): Likewise.
9829 (*ne_zeroextracts): Likewise.
9830 (*ne_zeroextracts): Likewise.
9831 (*ite_ne_zeroextr): Likewise.
9832 (*ite_ne_zeroextr): Likewise.
9833 (*anddi_notdi_di): Likewise.
9834 (*anddi_notzesidi): Likewise.
9835 (*anddi_notsesidi): Likewise.
9836 (andsi_notsi_si): Likewise.
9837 (thumb1_bicsi3): Likewise.
9838 (*iordi3_insn): Likewise.
9839 (*iordi_zesidi_di): Likewise.
9840 (*iordi_sesidi_di): Likewise.
9841 (*thumb1_iorsi3_insn): Likewise.
9842 (*xordi3_insn): Likewise.
9843 (*xordi_zesidi_di): Likewise.
9844 (*xordi_sesidi_di): Likewise.
9845 (*arm_xorsi3): Likewise.
9846 (*andsi_iorsi3_no): Likewise.
9847 (*smax_0): Likewise.
9848 (*smax_m1): Likewise.
9849 (*arm_smax_insn): Likewise.
9850 (*smin_0): Likewise.
9851 (*arm_smin_insn): Likewise.
9852 (*arm_umaxsi3): Likewise.
9853 (*arm_uminsi3): Likewise.
9854 (*minmax_arithsi): Likewise.
9855 (*minmax_arithsi_): Likewise.
9856 (*satsi_<SAT:code>): Likewise.
9857 (arm_ashldi3_1bit): Likewise.
9858 (arm_ashrdi3_1bit): Likewise.
9859 (arm_lshrdi3_1bit): Likewise.
9860 (*arm_negdi2): Likewise.
9861 (*thumb1_negdi2): Likewise.
9862 (*arm_negsi2): Likewise.
9863 (*thumb1_negsi2): Likewise.
9864 (*negdi_extendsid): Likewise.
9865 (*negdi_zero_extend): Likewise.
9866 (*arm_abssi2): Likewise.
9867 (*thumb1_abssi2): Likewise.
9868 (*arm_neg_abssi2): Likewise.
9869 (*thumb1_neg_abss): Likewise.
9870 (one_cmpldi2): Likewise.
9871 (extend<mode>di2): Likewise.
9872 (*compareqi_eq0): Likewise.
9873 (*arm_extendhisi2addsi): Likewise.
9874 (*arm_movdi): Likewise.
9875 (*thumb1_movdi_insn): Likewise.
9876 (*arm_movt): Likewise.
9877 (*thumb1_movsi_insn): Likewise.
9878 (pic_add_dot_plus_four): Likewise.
9879 (pic_add_dot_plus_eight): Likewise.
9880 (tls_load_dot_plus_eight): Likewise.
9881 (*thumb1_movhi_insn): Likewise.
9882 (*thumb1_movsf_insn): Likewise.
9883 (*movdf_soft_insn): Likewise.
9884 (*thumb_movdf_insn): Likewise.
9885 (cbranchsi4_insn): Likewise.
9886 (cbranchsi4_scratch): Likewise.
9887 (*negated_cbranchsi4): Likewise.
9888 (*tbit_cbranch): Likewise.
9889 (*tlobits_cbranch): Likewise.
9890 (*tstsi3_cbranch): Likewise.
9891 (*cbranchne_decr1): Likewise.
9892 (*addsi3_cbranch): Likewise.
9893 (*addsi3_cbranch_scratch): Likewise.
9894 (*arm_cmpdi_insn): Likewise.
9895 (*arm_cmpdi_unsig): Likewise.
9896 (*arm_cmpdi_zero): Likewise.
9897 (*thumb_cmpdi_zero): Likewise.
9898 (*deleted_compare): Likewise.
9899 (*mov_scc): Likewise.
9900 (*mov_negscc): Likewise.
9901 (*mov_notscc): Likewise.
9902 (*cstoresi_eq0_thumb1_insn): Likewise.
9903 (cstoresi_nltu_thumb1): Likewise.
9904 (cstoresi_ltu_thu): Likewise.
9905 (thumb1_addsi3_addgeu): Likewise.
9906 (*arm_jump): Likewise.
9907 (*thumb_jump): Likewise.
9908 (*check_arch2): Likewise.
9909 (arm_casesi_internal): Likewise.
9910 (thumb1_casesi_dispatch): Likewise.
9911 (*arm_indirect_jump): Likewise.
9912 (*thumb1_indirect_jump): Likewise.
9914 (*and_scc): Likewise.
9915 (*ior_scc): Likewise.
9916 (*compare_scc): Likewise.
9917 (*cond_move): Likewise.
9918 (*cond_arith): Likewise.
9919 (*cond_sub): Likewise.
9920 (*cmp_ite0): Likewise.
9921 (*cmp_ite1): Likewise.
9922 (*cmp_and): Likewise.
9923 (*cmp_ior): Likewise.
9924 (*ior_scc_scc): Likewise.
9925 (*ior_scc_scc_cmp): Likewise.
9926 (*and_scc_scc): Likewise.
9927 (*and_scc_scc_cmp): Likewise.
9928 (*and_scc_scc_nod): Likewise.
9929 (*negscc): Likewise.
9930 (movcond_addsi): Likewise.
9931 (movcond): Likewise.
9932 (*ifcompare_plus_move): Likewise.
9933 (*if_plus_move): Likewise.
9934 (*ifcompare_move_plus): Likewise.
9935 (*if_move_plus): Likewise.
9936 (*ifcompare_arith_arith): Likewise.
9937 (*if_arith_arith): Likewise.
9938 (*ifcompare_arith_move): Likewise.
9939 (*if_arith_move): Likewise.
9940 (*ifcompare_move_arith): Likewise.
9941 (*if_move_arith): Likewise.
9942 (*ifcompare_move_not): Likewise.
9943 (*if_move_not): Likewise.
9944 (*ifcompare_not_move): Likewise.
9945 (*if_not_move): Likewise.
9946 (*ifcompare_shift_move): Likewise.
9947 (*if_shift_move): Likewise.
9948 (*ifcompare_move_shift): Likewise.
9949 (*if_move_shift): Likewise.
9950 (*ifcompare_shift_shift): Likewise.
9951 (*ifcompare_not_arith): Likewise.
9952 (*ifcompare_arith_not): Likewise.
9953 (*if_arith_not): Likewise.
9954 (*ifcompare_neg_move): Likewise.
9955 (*if_neg_move): Likewise.
9956 (*ifcompare_move_neg): Likewise.
9957 (*if_move_neg): Likewise.
9958 (prologue_thumb1_interwork): Likewise.
9959 (*cond_move_not): Likewise.
9960 (*sign_extract_onebit): Likewise.
9961 (*not_signextract_onebit): Likewise.
9962 (stack_tie): Likewise.
9963 (align_4): Likewise.
9964 (align_8): Likewise.
9965 (consttable_end): Likewise.
9966 (consttable_1): Likewise.
9967 (consttable_2): Likewise.
9968 (consttable_4): Likewise.
9969 (consttable_8): Likewise.
9970 (consttable_16): Likewise.
9971 (*thumb1_tablejump): Likewise.
9972 (prefetch): Likewise.
9973 (force_register_use): Likewise.
9974 (thumb_eh_return): Likewise.
9975 (load_tp_hard): Likewise.
9976 (load_tp_soft): Likewise.
9977 (tlscall): Likewise.
9978 (*arm_movtas_ze): Likewise.
9979 (*arm_rev): Likewise.
9980 (*arm_revsh): Likewise.
9981 (*arm_rev16): Likewise.
9982 * config/arm/thumb2.md
9983 (*thumb2_smaxsi3): Likewise.
9984 (*thumb2_sminsi3): Likewise.
9985 (*thumb32_umaxsi3): Likewise.
9986 (*thumb2_uminsi3): Likewise.
9987 (*thumb2_negdi2): Likewise.
9988 (*thumb2_abssi2): Likewise.
9989 (*thumb2_neg_abss): Likewise.
9990 (*thumb2_movsi_insn): Likewise.
9991 (tls_load_dot_plus_four): Likewise.
9992 (*thumb2_movhi_insn): Likewise.
9993 (*thumb2_mov_scc): Likewise.
9994 (*thumb2_mov_negs): Likewise.
9995 (*thumb2_mov_negs): Likewise.
9996 (*thumb2_mov_nots): Likewise.
9997 (*thumb2_mov_nots): Likewise.
9998 (*thumb2_movsicc_): Likewise.
9999 (*thumb2_movsfcc_soft_insn): Likewise.
10000 (*thumb2_indirect_jump): Likewise.
10001 (*thumb2_and_scc): Likewise.
10002 (*thumb2_ior_scc): Likewise.
10003 (*thumb2_ior_scc_strict_it): Likewise.
10004 (*thumb2_cond_move): Likewise.
10005 (*thumb2_cond_arith): Likewise.
10006 (*thumb2_cond_ari): Likewise.
10007 (*thumb2_cond_sub): Likewise.
10008 (*thumb2_negscc): Likewise.
10009 (*thumb2_movcond): Likewise.
10010 (thumb2_casesi_internal): Likewise.
10011 (thumb2_casesi_internal_pic): Likewise.
10012 (*thumb2_alusi3_short): Likewise.
10013 (*thumb2_mov<mode>_shortim): Likewise.
10014 (*thumb2_addsi_short): Likewise.
10015 (*thumb2_subsi_short): Likewise.
10016 (thumb2_addsi3_compare0): Likewise.
10017 (*thumb2_cbz): Likewise.
10018 (*thumb2_cbnz): Likewise.
10019 (*thumb2_one_cmplsi2_short): Likewise.
10020 (*thumb2_negsi2_short): Likewise.
10021 (*orsi_notsi_si): Likewise.
10022 * config/arm/arm1020e.md: Update with new attributes.
10023 * config/arm/arm1026ejs.md: Update with new attributes.
10024 * config/arm/arm1136jfs.md: Update with new attributes.
10025 * config/arm/arm926ejs.md: Update with new attributes.
10026 * config/arm/cortex-a15.md: Update with new attributes.
10027 * config/arm/cortex-a5.md: Update with new attributes.
10028 * config/arm/cortex-a53.md: Update with new attributes.
10029 * config/arm/cortex-a7.md: Update with new attributes.
10030 * config/arm/cortex-a8.md: Update with new attributes.
10031 * config/arm/cortex-a9.md: Update with new attributes.
10032 * config/arm/cortex-m4.md: Update with new attributes.
10033 * config/arm/cortex-r4.md: Update with new attributes.
10034 * config/arm/fa526.md: Update with new attributes.
10035 * config/arm/fa606te.md: Update with new attributes.
10036 * config/arm/fa626te.md: Update with new attributes.
10037 * config/arm/fa726te.md: Update with new attributes.
10039 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10041 * config/aarch64/aarch64-simd.md
10042 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
10043 <vwx> iterator to ensure correct register choice.
10044 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
10045 (aarch64_sqdmull_n<mode>): Likewise.
10046 (aarch64_sqdmull2_n<mode>_internal): Likewise.
10047 * config/aarch64/arm_neon.h
10048 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
10049 (vml<as><q>_n_<su>16): Likewise.
10050 (vml<as>l_high_lane<q>_<su>16): Likewise.
10051 (vml<as>l_high_n_<su>16): Likewise.
10052 (vml<as>l_lane<q>_<su>16): Likewise.
10053 (vml<as>l_n_<su>16): Likewise.
10054 (vmul<q>_lane<q>_<su>16): Likewise.
10055 (vmul<q>_n_<su>16): Likewise.
10056 (vmull_lane<q>_<su>16): Likewise.
10057 (vmull_n_<su>16): Likewise.
10058 (vmull_high_lane<q>_<su>16): Likewise.
10059 (vmull_high_n_<su>16): Likewise.
10060 (vqrdmulh<q>_n_s16): Likewise.
10062 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
10064 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
10065 have the correct lane parameter.
10067 2013-09-06 Richard Biener <rguenther@suse.de>
10069 * cfganal.c (control_dependences::~control_dependences):
10070 Properly free all of the vector.
10072 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
10075 * config/i386/i386.c (ix86_conditional_register_usage):
10076 Proper initialize extended SSE registers.
10078 2013-09-06 Jan Hubicka <jh@suse.cz>
10080 PR tree-optimization/58311
10081 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
10083 2013-09-06 Jan Hubicka <jh@suse.cz>
10085 * Makefile.in (tree-sra.o): Update dependencies.
10086 * tree-sra.c: Include ipa-utils.h
10087 (scan_function): Use recursive_call_p.
10088 (has_caller_p): New function.
10089 (cgraph_for_node_and_aliases): Count also callers of aliases.
10091 2013-09-06 Jan Hubicka <jh@suse.cz>
10093 PR middle-end/58094
10094 * cgraph.h (symtab_semantically_equivalent_p): Declare.
10095 * tree-tailcall.c: Include ipa-utils.h.
10096 (find_tail_calls): Use it.
10097 * ipa-pure-const.c (check_call): Likewise.
10098 * ipa-utils.c (recursive_call_p): New function.
10099 * ipa-utils.h (recursive_call_p): Dclare.
10100 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
10101 (symtab_semantically_equivalent_p): New function.
10102 * Makefile.in (tree-tailcall.o): Update dependencies.
10104 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10106 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
10107 non-inlinable part.
10109 2013-09-06 Richard Biener <rguenther@suse.de>
10111 * lto-streamer.h (lto_global_var_decls): Remove.
10112 * Makefile.in (OBJS): Remove lto-symtab.o.
10113 (lto-symtab.o): Remove.
10114 (GTFILES): Remove lto-symtab.c
10115 * lto-symtab.c: Move to lto/
10117 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10119 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
10120 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
10121 (UNSPEC_FPINT_RINT): New constant definitions.
10122 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
10123 definition with 2 attributes.
10124 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
10125 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
10128 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10130 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
10131 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
10132 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
10133 attribute to "z196_alone".
10134 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
10137 2013-09-06 Richard Biener <rguenther@suse.de>
10139 * basic-block.h (class control_dependences): New.
10140 * tree-ssa-dce.c (control_dependence_map): Remove.
10142 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
10143 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
10144 find_pdom, find_control_dependence, find_all_control_dependences):
10146 (mark_control_dependent_edges_necessary,
10147 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
10148 tree_dce_done, perform_tree_ssa_dce): Adjust.
10149 * cfganal.c (set_control_dependence_map_bit,
10150 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
10151 find_all_control_dependences): Move from tree-ssa-dce.c and
10152 implement as methods of control_dependences class.
10153 (control_dependences::control_dependences): New.
10154 (control_dependences::~control_dependences): Likewise.
10155 (control_dependences::get_edges_dependent_on): Likewise.
10156 (control_dependences::get_edge): Likewise.
10158 2013-09-04 Jan Hubicka <jh@suse.cz>
10160 * tree.c (types_same_for_odr): Drop overactive check.
10161 * ipa-devirt.c (hash_type_name): Likewise.
10163 2013-09-04 Jan Hubicka <jh@suse.cz>
10165 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
10166 (analyze_functions): ... here.
10168 2013-09-04 Jan Hubicka <jh@suse.cz>
10170 PR middle-end/58201
10171 * cgraphunit.c (analyze_functions): Clear AUX fields
10172 after processing; initialize assembler name has.
10174 2013-09-05 Jeff Law <law@redhat.com>
10176 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
10177 from thread_around_empty_block. Record threading path into PATH.
10178 Recurse if threading through the initial block is successful.
10179 (thread_across_edge): Corresponding changes to slightly simplify.
10181 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10183 * config/aarch64/aarch64.md
10184 (type): Remove frecpe, frecps, frecpx.
10185 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
10186 fix to be a TARGET_SIMD instruction.
10187 (aarch64_frecps): Remove.
10188 * config/aarch64/aarch64-simd.md
10189 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
10190 (aarch64_frecps<mode>): Handle all float/vector of float modes.
10192 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10193 Sofiane Naci <sofiane.naci@arm.com>
10195 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
10196 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
10197 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
10198 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
10199 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
10200 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
10201 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
10202 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
10203 into "clz, "rbit". Rename "shift" to "shift_imm".
10204 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
10205 changes. Update for attribute changes all occurrences of arlo_* and
10207 * config/arm/arm-fixed.md: Update for attribute changes
10208 all occurrences of arlo_* types.
10209 * config/arm/thumb2.md: Update for attribute changes all occurrences
10211 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
10212 (cortexa7_older_only): Likewise.
10213 (cortexa7_younger): Likewise.
10214 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10215 (1020alu_shift_op): Likewise.
10216 (1020alu_shift_reg_op): Likewise.
10217 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10218 (alu_shift_op): Likewise.
10219 (alu_shift_reg_op): Likewise.
10220 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10221 (11_alu_shift_op): Likewise.
10222 (11_alu_shift_reg_op): Likewise.
10223 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10224 (9_alu_shift_reg_op): Likewise.
10225 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
10227 (cortex_a15_alu_shift): Likewise.
10228 (cortex_a15_alu_shift_reg): Likewise.
10229 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
10231 (cortex_a5_alu_shift): Likewise.
10232 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
10234 (cortex_a53_alu_shift): Likewise.
10235 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
10237 (cortex_a7_alu_reg): Likewise.
10238 (cortex_a7_alu_shift): Likewise.
10239 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
10241 (cortex_a8_alu_shift): Likewise.
10242 (cortex_a8_alu_shift_reg): Likewise.
10243 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10244 (cortex_a9_dp_shift): Likewise.
10245 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
10247 * config/arm/cortex-r4.md
10248 (cortex_r4_alu): Update for attribute changes.
10249 (cortex_r4_mov): Likewise.
10250 (cortex_r4_alu_shift_reg): Likewise.
10251 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10252 (526_alu_shift_op): Likewise.
10253 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10254 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10255 (626te_alu_shift_op): Likewise.
10256 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
10257 (726te_alu_shift_op): Likewise.
10258 (726te_alu_shift_reg_op): Likewise.
10259 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10260 (mp626_alu_shift_op): Likewise.
10261 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
10262 (pj4_alu_conds): Likewise.
10263 (pj4_shift): Likewise.
10264 (pj4_shift_conds): Likewise.
10265 (pj4_alu_shift): Likewise.
10266 (pj4_alu_shift_conds): Likewise.
10267 * config/aarch64/aarch64.md: Update for attribute change
10268 all occurrences of arlo_* and shift* types.
10270 2013-09-05 Mike Stump <mikestump@comcast.net>
10272 * tree.h: Move documentation for tree_function_decl to tree-core.h
10273 with the declaration.
10275 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
10278 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
10279 looking for widest mode.
10281 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
10283 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
10285 2013-09-05 Richard Biener <rguenther@suse.de>
10287 PR tree-optimization/58137
10288 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
10289 Do not create vectors of pointers.
10290 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
10291 types for the components of the vector initializer.
10292 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
10293 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
10295 2013-09-05 Martin Jambor <mjambor@suse.cz>
10297 * ipa-prop.c (remove_described_reference): Accept missing references,
10298 return false if that hppens, otherwise return true.
10299 (cgraph_node_for_jfunc): New function.
10300 (try_decrement_rdesc_refcount): Likewise.
10301 (try_make_edge_direct_simple_call): Use them.
10302 (ipa_edge_removal_hook): Remove references from rdescs.
10303 (ipa_edge_duplication_hook): Clone rdescs and their references
10304 when the new edge has the same caller as the old one.
10305 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10306 reference before removing any edges.
10308 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10310 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10312 * thumb2.md (thumb2_storewb_parisi): New pattern.
10314 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10316 * config/aarch64/aarch64-option-extensions.def: Add
10317 AARCH64_OPT_EXTENSION of 'crc'.
10318 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10319 (AARCH64_ISA_CRC): Ditto.
10320 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10321 description of the CRC extension.
10323 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10325 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10326 * config/rs6000/linux.h: Ditto.
10327 * alpha/linux.h: Ditto.
10328 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10329 no_c99_libc_has_function.
10330 * config/c6x/uclinux-elf.h: Ditto.
10331 * config/lm32/uclinux-elf.h: Ditto.
10332 * config/m68k/uclinux.h: Ditto.
10333 * config/moxie/uclinux.h: Ditto.
10334 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10335 (crisv32-*-linux*, cris-*-linux*): Ditto.
10336 * config/bfin/bfin.c: Include "tm_p.h".
10338 2013-09-05 Richard Biener <rguenther@suse.de>
10340 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10341 check for a definition without a basic-block.
10343 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10344 Sofiane Naci <sofiane.naci@arm.com>
10346 * config/aarch64/aarch64.md
10347 (*movti_aarch64): Rename r_2_f and f_2_r.
10348 (*movsf_aarch64): Likewise.
10349 (*movdf_aarch64): Likewise.
10350 (*movtf_aarch64): Likewise.
10351 (aarch64_movdi_<mode>low): Likewise.
10352 (aarch64_movdi_<mode>high): Likewise.
10353 (aarch64_mov<mode>high_di): Likewise.
10354 (aarch64_mov<mode>low_di): Likewise.
10355 (aarch64_movtilow_tilow): Likewise.
10356 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10357 values to config/arm/types.md
10358 (attribute "conds"): Update for attribute change.
10359 (anddi3_insn): Likewise.
10360 (iordi3_insn): Likewise.
10361 (xordi3_insn): Likewise.
10362 (one_cmpldi2): Likewise.
10363 * config/arm/types.md (type): Add Neon types.
10364 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10365 use "type" attribute.
10366 (movmisalign<mode>_neon_store): Likewise.
10367 (movmisalign<mode>_neon_load): Likewise.
10368 (vec_set<mode>_internal): Likewise.
10369 (vec_setv2di_internal): Likewise.
10370 (vec_extract<mode>): Likewise.
10371 (vec_extractv2di): Likewise.
10372 (add<mode>3_neon): Likewise.
10373 (adddi3_neon): Likewise.
10374 (sub<mode>3_neon): Likewise.
10375 (subdi3_neon): Likewise.
10376 (mul<mode>3_neon): Likewise.
10377 (mul<mode>3add<mode>_neon): Likewise.
10378 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10379 (fma<VCVTF:mode>4)): Likewise.
10380 (fma<VCVTF:mode>4_intrinsic): Likewise.
10381 (fmsub<VCVTF:mode>4)): Likewise.
10382 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10383 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10384 (ior<mode>3): Likewise.
10385 (and<mode>3): Likewise.
10386 (anddi3_neon): Likewise.
10387 (orn<mode>3_neon): Likewise.
10388 (orndi3_neon): Likewise.
10389 (bic<mode>3_neon): Likewise.
10390 (bicdi3_neon): Likewise.
10391 (xor<mode>3): Likewise.
10392 (one_cmpl<mode>2): Likewise.
10393 (abs<mode>2): Likewise.
10394 (neg<mode>2): Likewise.
10395 (umin<mode>3_neon): Likewise.
10396 (umax<mode>3_neon): Likewise.
10397 (smin<mode>3_neon): Likewise.
10398 (smax<mode>3_neon): Likewise.
10399 (vashl<mode>3): Likewise.
10400 (vashr<mode>3_imm): Likewise.
10401 (vlshr<mode>3_imm): Likewise.
10402 (ashl<mode>3_signed): Likewise.
10403 (ashl<mode>3_unsigned): Likewise.
10404 (neon_load_count): Likewise.
10405 (ashldi3_neon_noclobber): Likewise.
10406 (signed_shift_di3_neon): Likewise.
10407 (unsigned_shift_di3_neon): Likewise.
10408 (ashrdi3_neon_imm_noclobber): Likewise.
10409 (lshrdi3_neon_imm_noclobber): Likewise.
10410 (widen_ssum<mode>3): Likewise.
10411 (widen_usum<mode>3): Likewise.
10412 (quad_halves_<code>v4si): Likewise.
10413 (quad_halves_<code>v4sf): Likewise.
10414 (quad_halves_<code>v8hi): Likewise.
10415 (quad_halves_<code>v16qi): Likewise.
10416 (reduc_splus_v2di): Likewise.
10417 (neon_vpadd_internal<mode>): Likewise.
10418 (neon_vpsmin<mode>): Likewise.
10419 (neon_vpsmax<mode>): Likewise.
10420 (neon_vpumin<mode>): Likewise.
10421 (neon_vpumax<mode>): Likewise.
10422 (ss_add<mode>_neon): Likewise.
10423 (us_add<mode>_neon): Likewise.
10424 (ss_sub<mode>_neon): Likewise.
10425 (us_sub<mode>_neon): Likewise.
10426 (neon_vadd<mode>_unspec): Likewise.
10427 (neon_vaddl<mode>): Likewise.
10428 (neon_vaddw<mode>): Likewise.
10429 (neon_vhadd<mode>): Likewise.
10430 (neon_vqadd<mode>): Likewise.
10431 (neon_vaddhn<mode>): Likewise.
10432 (neon_vmul<mode>): Likewise.
10433 (neon_vmla<mode>): Likewise.
10434 (neon_vmlal<mode>): Likewise.
10435 (neon_vmls<mode>): Likewise.
10436 (neon_vmlsl<mode>): Likewise.
10437 (neon_vqdmulh<mode>): Likewise.
10438 (neon_vqdmlal<mode>): Likewise.
10439 (neon_vqdmlsl<mode>): Likewise.
10440 (neon_vmull<mode>): Likewise.
10441 (neon_vqdmull<mode>): Likewise.
10442 (neon_vsub<mode>_unspec): Likewise.
10443 (neon_vsubl<mode>): Likewise.
10444 (neon_vsubw<mode>): Likewise.
10445 (neon_vqsub<mode>): Likewise.
10446 (neon_vhsub<mode>): Likewise.
10447 (neon_vsubhn<mode>): Likewise.
10448 (neon_vceq<mode>): Likewise.
10449 (neon_vcge<mode>): Likewise.
10450 (neon_vcgeu<mode>): Likewise.
10451 (neon_vcgt<mode>): Likewise.
10452 (neon_vcgtu<mode>): Likewise.
10453 (neon_vcle<mode>): Likewise.
10454 (neon_vclt<mode>): Likewise.
10455 (neon_vcage<mode>): Likewise.
10456 (neon_vcagt<mode>): Likewise.
10457 (neon_vtst<mode>): Likewise.
10458 (neon_vabd<mode>): Likewise.
10459 (neon_vabdl<mode>): Likewise.
10460 (neon_vaba<mode>): Likewise.
10461 (neon_vabal<mode>): Likewise.
10462 (neon_vmax<mode>): Likewise.
10463 (neon_vmin<mode>): Likewise.
10464 (neon_vpaddl<mode>): Likewise.
10465 (neon_vpadal<mode>): Likewise.
10466 (neon_vpmax<mode>): Likewise.
10467 (neon_vpmin<mode>): Likewise.
10468 (neon_vrecps<mode>): Likewise.
10469 (neon_vrsqrts<mode>): Likewise.
10470 (neon_vqabs<mode>): Likewise.
10471 (neon_vqneg<mode>): Likewise.
10472 (neon_vcls<mode>): Likewise.
10473 (clz<mode>2): Likewise.
10474 (popcount<mode>2): Likewise.
10475 (neon_vrecpe): Likewise.
10476 (neon_vrsqrte): Likewise.
10477 (neon_vget_lane<mode>_sext_internal): Likewise.
10478 (neon_vget_lane<mode>_zext_internal): Likewise.
10479 (neon_vdup_n<mode>): Likewise.
10480 (neon_vdup_nv2di): Likewise.
10481 (neon_vdpu_lane<mode>_internal): Likewise.
10482 (neon_vswp<mode>): Likewise.
10483 (float<mode><V_cvtto>2): Likewise.
10484 (floatuns<mode><V_cvtto>2): Likewise.
10485 (fix_trunc<mode><V_cvtto>)2): Likewise
10486 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10487 (neon_vcvt<mode>): Likewise.
10488 (neon_vcvtv4sfv4hf): Likewise.
10489 (neon_vcvtv4hfv4sf): Likewise.
10490 (neon_vcvt_n<mode>): Likewise.
10491 (neon_vmovn<mode>): Likewise.
10492 (neon_vqmovn<mode>): Likewise.
10493 (neon_vqmovun<mode>): Likewise.
10494 (neon_vmovl<mode>): Likewise.
10495 (neon_vmul_lane<mode>): Likewise.
10496 (neon_vmull_lane<mode>): Likewise.
10497 (neon_vqdmull_lane<mode>): Likewise.
10498 (neon_vqdmulh_lane<mode>): Likewise.
10499 (neon_vmla_lane<mode>): Likewise.
10500 (neon_vmlal_lane<mode>): Likewise.
10501 (neon_vqdmlal_lane<mode>): Likewise.
10502 (neon_vmls_lane<mode>): Likewise.
10503 (neon_vmlsl_lane<mode>): Likewise.
10504 (neon_vqdmlsl_lane<mode>): Likewise.
10505 (neon_vext<mode>): Likewise.
10506 (neon_vrev64<mode>): Likewise.
10507 (neon_vrev32<mode>): Likewise.
10508 (neon_vrev16<mode>): Likewise.
10509 (neon_vbsl<mode>_internal): Likewise.
10510 (neon_vshl<mode>): Likewise.
10511 (neon_vqshl<mode>): Likewise.
10512 (neon_vshr_n<mode>): Likewise.
10513 (neon_vshrn_n<mode>): Likewise.
10514 (neon_vqshrn_n<mode>): Likewise.
10515 (neon_vqshrun_n<mode>): Likewise.
10516 (neon_vshl_n<mode>): Likewise.
10517 (neon_vqshl_n<mode>): Likewise.
10518 (neon_vqshlu_n<mode>): Likewise.
10519 (neon_vshll_n<mode>): Likewise.
10520 (neon_vsra_n<mode>): Likewise.
10521 (neon_vsri_n<mode>): Likewise.
10522 (neon_vsli_n<mode>): Likewise.
10523 (neon_vtbl1v8qi): Likewise.
10524 (neon_vtbl2v8qi): Likewise.
10525 (neon_vtbl3v8qi): Likewise.
10526 (neon_vtbl4v8qi): Likewise.
10527 (neon_vtbx1v8qi): Likewise.
10528 (neon_vtbx2v8qi): Likewise.
10529 (neon_vtbx3v8qi): Likewise.
10530 (neon_vtbx4v8qi): Likewise.
10531 (neon_vtrn<mode>_internal): Likewise.
10532 (neon_vzip<mode>_internal): Likewise.
10533 (neon_vuzp<mode>_internal): Likewise.
10534 (neon_vld1<mode>): Likewise.
10535 (neon_vld1_lane<mode>): Likewise.
10536 (neon_vld1_dup<mode>): Likewise.
10537 (neon_vld1_dupv2di): Likewise.
10538 (neon_vst1<mode>): Likewise.
10539 (neon_vst1_lane<mode>): Likewise.
10540 (neon_vld2<mode>): Likewise.
10541 (neon_vld2_lane<mode>): Likewise.
10542 (neon_vld2_dup<mode>): Likewise.
10543 (neon_vst2<mode>): Likewise.
10544 (neon_vst2_lane<mode>): Likewise.
10545 (neon_vld3<mode>): Likewise.
10546 (neon_vld3qa<mode>): Likewise.
10547 (neon_vld3qb<mode>): Likewise.
10548 (neon_vld3_lane<mode>): Likewise.
10549 (neon_vld3_dup<mode>): Likewise.
10550 (neon_vst3<mode>): Likewise.
10551 (neon_vst3qa<mode>): Likewise.
10552 (neon_vst3qb<mode>): Likewise.
10553 (neon_vst3_lane<mode>): Likewise.
10554 (neon_vld4<mode>): Likewise.
10555 (neon_vld4qa<mode>): Likewise.
10556 (neon_vld4qb<mode>): Likewise.
10557 (neon_vld4_lane<mode>): Likewise.
10558 (neon_vld4_dup<mode>): Likewise.
10559 (neon_vst4<mode>): Likewise.
10560 (neon_vst4qa<mode>): Likewise.
10561 (neon_vst4qb<mode>): Likewise.
10562 (neon_vst4_lane<mode>): Likewise.
10563 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10564 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10565 (neon_vec_<US>mult_lo_<mode>): Likewise.
10566 (neon_vec_<US>mult_hi_<mode>): Likewise.
10567 (neon_vec_<US>shiftl_<mode>): Likewise.
10568 (neon_unpack<US>_<mode>): Likewise.
10569 (neon_vec_<US>mult_<mode>): Likewise.
10570 (vec_pack_trunc_<mode>): Likewise.
10571 (neon_vec_pack_trunk_<mode>): Likewise.
10572 (neon_vabd<mode>_2): Likewise.
10573 (neon_vabd<mode>_3): Likewise.
10574 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10575 (thumb2_movsi_vfp): Likewise.
10576 (movdi_vfp): Likewise.
10577 (movdi_vfp_cortexa8): Likewise.
10578 (movhf_vfp_neon): Likewise.
10579 (movhf_vfp): Likewiwse.
10580 (movsf_vfp): Likewiwse.
10581 (thumb2_movsf_vfp): Likewiwse.
10582 (movdf_vfp): Likewise.
10583 (thumb2_movdf_vfp): Likewise.
10584 (movsfcc_vfp): Likewise.
10585 (thumb2_movsfcc_vfp): Likewise.
10586 (movdfcc_vfp): Likewise.
10587 (thumb2_movdfcc_vfp): Likewise.
10588 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10589 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10590 (v10_v2c): Likewise.
10591 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10593 (cortex_a15_neon_int_2): Likewise.
10594 (cortex_a15_neon_int_3): Likewise.
10595 (cortex_a15_neon_int_4): Likewise.
10596 (cortex_a15_neon_int_5): Likewise.
10597 (cortex_a15_neon_vqneg_vqabs): Likewise.
10598 (cortex_a15_neon_vmov): Likewise.
10599 (cortex_a15_neon_vaba): Likewise.
10600 (cortex_a15_neon_vaba_qqq): Likewise.
10601 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10602 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10603 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10604 scalar_64_32_long_scalar): Likewise.
10605 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10606 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10607 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10608 lotype_qdd_64_32_long): Likewise.
10609 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10610 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10611 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10612 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10613 (cortex_a15_neon_shift_1): Likewise.
10614 (cortex_a15_neon_shift_2): Likewise.
10615 (cortex_a15_neon_shift_3): Likewise.
10616 (cortex_a15_neon_vshl_ddd): Likewise.
10617 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10618 (cortex_a15_neon_vsra_vrsra): Likewise.
10619 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10620 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10621 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10622 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10623 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10624 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10625 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10626 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10627 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10628 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10629 (cortex_a15_neon_bp_simple): Likewise.
10630 (cortex_a15_neon_bp_2cycle): Likewise.
10631 (cortex_a15_neon_bp_3cycle): Likewise.
10632 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10633 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10634 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10635 (cortex_a15_neon_vld2_4_regs): Likewise.
10636 (cortex_a15_neon_vld3_vld4): Likewise.
10637 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10638 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10639 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10640 (cortex_a15_neon_vst3_vst4): Likewise.
10641 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10642 (cortex_a15_neon_vld3_vld4_lane" 10
10643 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10644 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10645 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10646 (cortex_a15_neon_ldm_2): Likewise.0
10647 (cortex_a15_neon_stm_2): Likewise.
10648 (cortex_a15_neon_mcr): Likewise.
10649 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10650 (cortex_a15_neon_mrc): Likewise.
10651 (cortex_a15_neon_mrrc): Likewise.
10652 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10654 (cortex_a15_alu_shift): Likewise.
10655 (cortex_a15_alu_shift_reg): Likewise.
10656 (cortex_a15_mult32): Likewise.
10657 (cortex_a15_mult64): Likewise.
10658 (cortex_a15_block): Likewise.
10659 (cortex_a15_branch): Likewise.
10660 (cortex_a15_load1): Likewise.
10661 (cortex_a15_load3): Likewise.
10662 (cortex_a15_store1): Likewise.
10663 (cortex_a15_store3): Likewise.
10664 (cortex_a15_call): Likewise.
10665 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10666 (cortex_a5_f2r): Likewise.
10667 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10669 (cortex_a53_f2r): Likewise.
10670 * config/arm/cortex-a7.md
10671 (cortex_a7_branch): Update for attribute change.
10672 (cortex_a7_call): Likewise.
10673 (cortex_a7_alu_imm): Likewise.
10674 (cortex_a7_alu_reg): Likewise.
10675 (cortex_a7_alu_shift): Likewise.
10676 (cortex_a7_mul): Likewise.
10677 (cortex_a7_load1): Likewise.
10678 (cortex_a7_store1): Likewise.
10679 (cortex_a7_load2): Likewise.
10680 (cortex_a7_store2): Likewise.
10681 (cortex_a7_load3): Likewise.
10682 (cortex_a7_store3): Likewise.
10683 (cortex_a7_load4): Likewise.
10684 (cortex_a7_store4): Likewise.
10685 (cortex_a7_fpalu): Likewise.
10686 (cortex_a7_fconst): Likewise.
10687 (cortex_a7_fpmuls): Likewise.
10688 (cortex_a7_neon_mul): Likewise.
10689 (cortex_a7_fpmacs): Likewise.
10690 (cortex_a7_neon_mla: Likewise.
10691 (cortex_a7_fpmuld: Likewise.
10692 (cortex_a7_fpmacd: Likewise.
10693 (cortex_a7_fpfmad: Likewise.
10694 (cortex_a7_fdivs: Likewise.
10695 (cortex_a7_fdivd: Likewise.
10696 (cortex_a7_r2f: Likewise.
10697 (cortex_a7_f2r: Likewise.
10698 (cortex_a7_f_flags: Likewise.
10699 (cortex_a7_f_loads: Likewise.
10700 (cortex_a7_f_loadd: Likewise.
10701 (cortex_a7_f_stores: Likewise.
10702 (cortex_a7_f_stored: Likewise.
10703 (cortex_a7_neon): Likewise.
10704 * config/arm/cortex-a8-neon.md
10705 (cortex_a8_neon_mrc): Update for attribute change.
10706 (cortex_a8_neon_mrrc): Likewise.
10707 (cortex_a8_neon_int_1): Likewise.
10708 (cortex_a8_neon_int_2): Likewise.
10709 (cortex_a8_neon_int_3): Likewise.
10710 (cortex_a8_neon_int_4): Likewise.
10711 (cortex_a8_neon_int_5): Likewise.
10712 (cortex_a8_neon_vqneg_vqabs): Likewise.
10713 (cortex_a8_neon_vmov): Likewise.
10714 (cortex_a8_neon_vaba): Likewise.
10715 (cortex_a8_neon_vaba_qqq): Likewise.
10716 (cortex_a8_neon_vsma): Likewise.
10717 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10718 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10719 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10720 long_scalar): Likewise.
10721 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10722 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10723 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10724 long_scalar_qdd_64_32_long): Likewise.
10725 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10726 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10727 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10728 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10729 (cortex_a8_neon_shift_1): Likewise.
10730 (cortex_a8_neon_shift_2): Likewise.
10731 (cortex_a8_neon_shift_3): Likewise.
10732 (cortex_a8_neon_vshl_ddd): Likewise.
10733 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10734 (cortex_a8_neon_vsra_vrsra): Likewise.
10735 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10736 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10737 (cortex_a8_neon_fp_vsum): Likewise.
10738 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10739 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10740 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10741 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10742 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10743 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10744 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10745 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10746 (cortex_a8_neon_bp_simple): Likewise.
10747 (cortex_a8_neon_bp_2cycle): Likewise.
10748 (cortex_a8_neon_bp_3cycle): Likewise.
10749 (cortex_a8_neon_ldr): Likewise.
10750 (cortex_a8_neon_str): Likewise.
10751 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10752 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10753 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10754 (cortex_a8_neon_vld2_4_regs): Likewise.
10755 (cortex_a8_neon_vld3_vld4): Likewise.
10756 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10757 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10758 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10759 (cortex_a8_neon_vst3_vst4): Likewise.
10760 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10761 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10762 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10763 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10764 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10765 (cortex_a8_neon_mcr): Likewise.
10766 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10767 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10768 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10770 (ca9_neon_mrrc): Likewise.
10771 (cortex_a9_neon_int_1): Likewise.
10772 (cortex_a9_neon_int_2): Likewise.
10773 (cortex_a9_neon_int_3): Likewise.
10774 (cortex_a9_neon_int_4): Likewise.
10775 (cortex_a9_neon_int_5): Likewise.
10776 (cortex_a9_neon_vqneg_vqabs): Likewise.
10777 (cortex_a9_neon_vmov): Likewise.
10778 (cortex_a9_neon_vaba): Likewise.
10779 (cortex_a9_neon_vaba_qqq): Likewise.
10780 (cortex_a9_neon_vsma): Likewise.
10781 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10782 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10783 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10784 long_scalar): Likewise.
10785 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10786 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10787 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10788 long_scalar_qdd_64_32_long): Likewise.
10789 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10790 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10791 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10792 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10793 (cortex_a9_neon_shift_1): Likewise.
10794 (cortex_a9_neon_shift_2): Likewise.
10795 (cortex_a9_neon_shift_3): Likewise.
10796 (cortex_a9_neon_vshl_ddd): Likewise.
10797 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10798 (cortex_a9_neon_vsra_vrsra): Likewise.
10799 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10800 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10801 (cortex_a9_neon_fp_vsum): Likewise.
10802 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10803 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10804 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10805 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10806 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10807 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10808 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10809 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10810 (cortex_a9_neon_bp_simple): Likewise.
10811 (cortex_a9_neon_bp_2cycle): Likewise.
10812 (cortex_a9_neon_bp_3cycle): Likewise.
10813 (cortex_a9_neon_ldr): Likewise.
10814 (cortex_a9_neon_str): Likewise.
10815 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10816 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10817 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10818 (cortex_a9_neon_vld2_4_regs): Likewise.
10819 (cortex_a9_neon_vld3_vld4): Likewise.
10820 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10821 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10822 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10823 (cortex_a9_neon_vst3_vst4): Likewise.
10824 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10825 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10826 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10827 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10828 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10829 (cortex_a9_neon_mcr): Likewise.
10830 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10831 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10832 (cortex_a9_fps): Likewise.
10833 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10835 (cortex_m4_fmuls): Likewise.
10836 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10838 (cortex_r4_mrc): Likewise.
10839 * config/arm/iterators.md: Update comment referring to neon_type.
10840 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10841 (iwmmxt_movsi_insn): Likewise.
10842 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10844 (pj4_core_to_vfp): Likewise.
10845 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10847 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10848 (vfp_fstore): Likewise.
10849 * doc/md.texi: Change references to neon_type to refer to type.
10851 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10853 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10855 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10857 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10858 lglobl if not weak.
10860 2013-09-04 Easwaran Raman <eraman@google.com>
10862 PR middle-end/57370
10863 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10864 (build_and_add_sum): Use it.
10865 (appears_later_in_bb): Simplify code.
10867 2013-09-04 Teresa Johnson <tejohnson@google.com>
10869 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10871 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10873 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10875 2013-09-04 Jan Hubicka <jh@suse.cz>
10877 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10878 * ipa-devirt.c: Include diganostic.h
10879 (odr_type_d): Add types and types_set.
10880 (hash_type_name): Work for types with vtables during LTO.
10881 (odr_hasher::remove): Fix comment; destroy types_set.
10882 (add_type_duplicate): New function,
10883 (get_odr_type): Use it.
10884 (dump_type_inheritance_graph): Dump type duplicates.
10885 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10887 * tree.c (types_same_for_odr): Give exact answers on types with
10890 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10892 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10893 explaining their differences.
10895 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10897 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10899 2013-09-03 Jeff Law <law@redhat.com>
10901 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10902 when not threading through a joiner block. Pass joiner/no joiner
10903 state to register_jump_thread.
10904 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10905 state from argument rather than implying on path length.
10906 Dump the entire jump thread path into debugging dump.
10907 * tree-flow.h (register_jump_thread): Update prototype.
10909 2013-08-29 Xinliang David Li <davidxl@google.com>
10911 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10912 Remove a trivial gcc_assert.
10914 2013-08-29 Xinliang David Li <davidxl@google.com>
10916 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10917 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10918 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10919 Delay base decl alignment adjustment.
10920 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10921 * tree-vectorizer.h: New data structure.
10922 (set_dr_misalignment): New function.
10923 (dr_misalignment): Ditto.
10924 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10925 (vectorizable_load): Ditto.
10926 (ensure_base_align): New function.
10927 (vectorize_loops): Add dbg_cnt support.
10928 (execute_vect_slp): Ditto.
10929 * dbgcnt.def: New debug counter.
10930 * Makefile: New dependency.
10932 2013-09-03 Meador Inge <meadori@codesourcery.com>
10936 2013-08-30 Meador Inge <meadori@codesourcery.com>
10938 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10940 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10942 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10943 function descriptor.
10945 2013-09-03 Richard Biener <rguenther@suse.de>
10947 * tree-affine.c (add_elt_to_tree): Fix association issue,
10948 avoid useless converts and make sure to always return a
10949 properly typed result.
10951 2013-09-03 Richard Biener <rguenther@suse.de>
10953 PR middle-end/57656
10954 * fold-const.c (negate_expr_p): Fix division case.
10955 (negate_expr): Likewise.
10957 2013-09-03 Richard Biener <rguenther@suse.de>
10960 * tree-streamer-out.c: Include tm.h.
10961 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10963 2013-09-03 Jan Hubicka <jh@suse.cz>
10965 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10967 2013-09-03 Alan Modra <amodra@gmail.com>
10969 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10970 multilibs to big-endian target and vice versa.
10971 * config/rs6000/t-linux64: Use := assignment on all vars.
10972 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10973 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10974 * config/rs6000/t-linux64le: New file.
10975 * config/rs6000/t-linux64bele: New file.
10976 * config/rs6000/t-linux64lebe: New file.
10978 2013-09-02 Jan Hubicka <jh@suse.cz>
10980 * ipa-inline-transform.c (inline_transform): Do not
10981 optimize_inline_calls when not optimizing.
10983 2013-09-02 Jan Hubicka <jh@suse.cz>
10985 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10986 duplicated nodes for assembler names.
10987 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10988 hard registers from assembler name hash.
10990 2013-09-02 Jan Hubicka <jh@suse.cz>
10992 * ipa-split.c (execute_split_functions): Split externally visible
10993 functions called once.
10995 2013-09-02 Martin Jambor <mjambor@suse.cz>
10998 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10999 linked list. When finding the correct duplicate, also consider also
11000 the caller in additon to its inlined_to node.
11002 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
11004 * config/aarch64/aarch64-simd-builtins.def
11005 (dup_lane_scalar): Remove.
11006 * config/aarch64/aarch64-simd.md
11007 (aarch64_simd_dup): Add 'w->w' alternative.
11008 (aarch64_dup_lane<mode>): Allow for VALL.
11009 (aarch64_dup_lane_scalar<mode>): Remove.
11010 (aarch64_dup_lane_<vswap_width_name><mode>): New.
11011 (aarch64_get_lane_signed<mode>): Add w->w altenative.
11012 (aarch64_get_lane_unsigned<mode>): Likewise.
11013 (aarch64_get_lane<mode>): Likewise.
11014 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
11015 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
11016 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
11017 (VCON): Change container of V2SF.
11018 (vswap_width_name): Likewise.
11019 * config/aarch64/arm_neon.h
11020 (__aarch64_vdup_lane_any): New.
11021 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11022 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
11023 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11025 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
11027 PR middle-end/56382
11028 * expr.c (emit_move_complex): Do not move complex FP values as parts if
11029 the source or the destination is a single hard register.
11031 2013-09-02 Richard Biener <rguenther@suse.de>
11033 PR middle-end/57511
11034 * tree-scalar-evolution.c (instantiate_scev_name): Allow
11037 2013-09-02 Richard Biener <rguenther@suse.de>
11039 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
11040 arithmetic to sizetype.
11042 2013-09-02 Bin Cheng <bin.cheng@arm.com>
11044 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
11045 Find auto-increment use both before and after candidate.
11047 2013-09-02 Marek Polacek <polacek@redhat.com>
11049 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
11051 2013-09-01 Jan Hubicka <jh@suse.cz>
11053 * Makefile.in: Add ipa-profile.o
11054 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
11055 * cgraph.c (struct cgraph_propagate_frequency_data,
11056 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
11057 ipa-profile.c; replace cgraph_ by ipa_ prefix.
11058 * cgraph.h (cgraph_propagate_frequency): Remove.
11059 * ipa-inline-analysis.c: Include ipa-utils.h;
11060 drop duplicated cfgloop.h.
11061 (inline_update_callee_summaries): Update.
11062 * ipa-profile.c: New file.
11063 * ipa-utils.h (ipa_propagate_frequency): Declare.
11064 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
11065 data-streamer.h, value-prof.h.
11066 (symtab_remove_unreachable_nodes): Update profile.
11067 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
11068 account_time_size, cmp_counts, dump_histogram,
11069 ipa_profile_generate_summary, ipa_profile_write_summary,
11070 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
11071 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
11072 Move to ipa-profile.c.
11074 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
11076 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
11078 2013-09-01 Jan Hubicka <jh@suse.cz>
11080 * common.opt (fdevirtualize-speculatively): New function.
11081 * invoke.texi (fdevirtualize-speculatively): Document.
11082 * ipa-devirt.c: Include ipa-inline.h
11083 (likely_target_p): New function.
11084 (ipa_devirt): New function.
11085 (gate_ipa_devirt): New function.
11086 (pass_data_ipa_devirt): New static var.
11087 (pass_ipa_devirt): Likewise.
11088 (make_pass_ipa_devirt): New function.
11089 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
11090 (common_handle_option): Disable devirtualization when
11091 value range profiling is available.
11092 * passes.def (pass_ipa_devirt): Add.
11093 * timever.def (TV_IPA_DEVIRT): New timevar.
11094 * tree-pass.h (make_pass_ipa_devirt):
11096 2013-09-01 Iain Sandoe <iain@codesourcery.com>
11098 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
11099 include sanitize(undefined).
11101 2013-08-31 Diego Novillo <dnovillo@google.com>
11103 * Makefile.in (TREE_CORE_H): Define.
11105 (GTFILES): Add tree-core.h.
11106 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
11108 * tree-core.h: New file.
11109 Move all data structures, enum, typedefs, global
11110 declarations and constants from ...
11111 * tree.h: ... here.
11113 2013-08-31 Jan Hubicka <jh@suse.cz>
11115 * bulitins.c (expand_builtin): Do not early exit for gcov
11116 instrumented functions.
11118 2013-08-31 Marek Polacek <polacek@redhat.com>
11120 * ubsan.c: Include tm_p.h.
11122 2013-08-31 Jan Hubicka <jh@suse.cz>
11124 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
11127 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
11128 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
11129 * tree-cfg.c (verify_gimple_label): ... here.
11130 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
11131 (ipa_merge_profiles): New function.
11132 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
11133 (lto_input_function_body): Likewise.
11134 * ipa-utils.h (ipa_merge_profiles): Declare.
11135 * lto-streamer.h (lto_input_function_body): Update prototype.
11136 (emit_label_in_global_context_p): Remove.
11137 * lto-symtab.c: Include ipa-utils.h
11138 (lto_cgraph_replace_node): Use ipa_merge_profiles.
11140 2013-08-31 Jan Hubicka <jh@suse.cz>
11142 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
11144 2013-08-31 Jan Hubicka <jh@suse.cz>
11146 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
11148 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
11150 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
11151 "cmp" RTX before signed_comparison_operator check to account
11152 for "code" changes.
11154 2013-08-30 Jan Hubicka <jh@suse.cz>
11156 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
11157 (detect_type_change_1): Rename to ...
11158 (detect_type_change): ... this one; early return on non-polymorphic
11160 (detect_type_change_ssa): Add comp_type parameter; update
11161 use of detect_type_change.
11162 (compute_complex_assign_jump_func): Add param_type parameter;
11163 update use of detect_type_change_ssa.
11164 (compute_complex_ancestor_jump_func): Likewise.
11165 (ipa_get_callee_param_type): New function.
11166 (ipa_compute_jump_functions_for_edge): Compute parameter type;
11167 update calls to the jump function computation functions.
11169 2013-08-30 Teresa Johnson <tejohnson@google.com>
11170 Steven Bosscher <steven@gcc.gnu.org>
11172 * cfgrtl.c (fixup_new_cold_bb): New routine.
11173 (commit_edge_insertions): Invoke fixup_partitions.
11174 (find_partition_fixes): New routine.
11175 (fixup_partitions): Ditto.
11176 (verify_hot_cold_block_grouping): Update comments.
11177 (rtl_verify_edges): Invoke find_partition_fixes.
11178 (rtl_verify_bb_pointers): Update comments.
11179 (rtl_verify_bb_layout): Ditto.
11180 * basic-block.h (probably_never_executed_edge_p): Declare.
11181 (fixup_partitions): Ditto.
11182 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
11183 * bb-reorder.c (sanitize_hot_paths): New function.
11184 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
11185 sanitize_hot_paths.
11186 * predict.c (probably_never_executed_edge_p): New routine.
11187 * cfg.c (check_bb_profile): Add partition insanity warnings.
11189 2013-08-30 Meador Inge <meadori@codesourcery.com>
11191 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11193 2013-08-30 Marek Polacek <polacek@redhat.com>
11195 * Makefile.in (ubsan.o): Add.
11196 (c-family/c-ubsan.o): Add.
11197 (builtins.o): Add ubsan.h dependency.
11198 * ubsan.h: New file.
11199 * ubsan.c: New file.
11200 * common.opt: Add -fsanitize=undefined option.
11201 (flag_sanitize): Add variable.
11202 (fsanitize=): Add option. Add Driver.
11203 (fsanitize=thread): Remove option.
11204 (fsanitize=address): Likewise.
11205 (static-libubsan): New option.
11206 * doc/invoke.texi: Document the new flag and -static-libubsan.
11207 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
11208 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
11209 * builtin-attrs.def (ATTR_COLD): Define.
11210 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11211 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
11212 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
11213 * flag-types.h (sanitize_code): New enum.
11214 * opts.c (common_handle_option): Parse command line arguments
11215 of -fsanitize=. Add -fsanitize=unreachable option.
11216 * varasm.c (get_variable_section): Adjust.
11217 (assemble_noswitch_variable): Likewise.
11218 (assemble_variable): Likewise.
11219 (output_constant_def_contents): Likewise.
11220 (categorize_decl_for_section): Likewise.
11221 (place_block_symbol): Likewise.
11222 (output_object_block): Likewise.
11223 * builtins.def: Likewise.
11224 * toplev.c (compile_file): Likewise.
11225 (process_options): Likewise.
11226 * cppbuiltin.c: Likewise.
11227 * tsan.c (tsan_pass): Likewise.
11228 (tsan_gate): Likewise.
11229 (tsan_gate_O0): Likewise.
11230 * cfgexpand.c (partition_stack_vars): Likewise.
11231 (expand_stack_vars): Likewise.
11232 (defer_stack_allocation): Likewise.
11233 (expand_used_vars): Likewise.
11234 * cfgcleanup.c (old_insns_match_p): Likewise.
11235 * asan.c (asan_finish_file): Likewise.
11236 (asan_instrument): Likewise.
11237 (gate_asan): Likewise.
11238 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
11239 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11240 (asan_global_struct): Use pointer_sized_int_node instead
11241 calling build_nonstandard_integer_type.
11242 (initialize_sanitizer_builtins): Likewise.
11243 (asan_finish_file): Likewise.
11244 * gcc.c: Document %{%:function(args):X}.
11245 (static_spec_functions): Add sanitize.
11246 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
11247 store funcval != NULL there.
11248 (do_spec_1): Adjust handle_spec_function caller.
11249 (handle_braces): Allow %:function(args) as condition.
11250 (sanitize_spec_function): New function.
11251 (ADD_STATIC_LIBUBSAN_LIBS): Define.
11252 (LIBUBSAN_SPEC): Likewise.
11253 (LIBUBSAN_EARLY_SPEC): Likewise.
11254 (SANITIZER_SPEC): Handle libubsan.
11255 (SANITIZER_EARLY_SPEC): Likewise.
11256 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
11257 instead of fsanitize=address.
11258 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
11259 instead of fsanitize=address*.
11260 * builtins.c: Include ubsan.h.
11261 (fold_builtin_0): Instrument __builtin_unreachable.
11262 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
11263 instead of flag_asan.
11264 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
11265 (pointer_sized_int_node): Define.
11266 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
11268 2013-08-30 Mike Stump <mikestump@comcast.net>
11270 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
11273 2013-08-29 Jan Hubicka <jh@suse.cz>
11275 * cgraph.c (cgraph_function_body_availability): Handle weakref
11277 * passes.def: Remove pass_fixup_cfg.
11278 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
11279 track when we need to remove functions.
11280 (gate_ipa_inline): Execute inlining always; add comment why.
11281 (pass_data_ipa_inline): Remove TODO_remove_functions.
11282 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
11283 do not produce summaries.
11284 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
11285 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
11286 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
11287 constant pool and vtable.
11289 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
11291 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
11292 New arm_neon.h's internal macros to specify 64-bit constants.
11293 Avoid using stdint.h's macros.
11295 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
11297 * recog.c (verify_changes): Verify that changes[i].old is non-zero
11298 before applying REG_P.
11300 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11302 PR tree-optimization/58277
11303 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11304 after seeing too many stmts with vdef in between dombb and current
11305 bb, invalidate everything.
11307 2013-08-30 Richard Biener <rguenther@suse.de>
11309 * fold-const.c (fold_single_bit_test): Fix overflow test.
11311 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11313 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11314 and which can live in a register, always retrieve the value on entry.
11315 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11316 passed by invisible reference specially.
11317 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11318 (vt_add_function_parameter): Correctly deal with a parameter passed by
11319 invisible reference.
11321 2013-08-30 Jan Hubicka <jh@suse.cz>
11323 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11325 2013-08-30 Richard Biener <rguenther@suse.de>
11327 PR tree-optimization/58228
11328 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11329 allow invariant loads in nested loop vectorization.
11331 2013-08-30 Richard Biener <rguenther@suse.de>
11333 PR tree-optimization/58223
11334 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11335 (has_anti_or_output_dependence): ... this and adjust to also
11336 look for output dependences.
11337 (mark_nodes_having_upstream_mem_writes): Adjust.
11338 (rdg_flag_uses): Likewise.
11340 2013-08-30 Richard Biener <rguenther@suse.de>
11342 PR tree-optimization/58010
11343 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11344 assert that we have a loop-closed PHI.
11346 2013-08-29 Jan Hubicka <jh@suse.cz>
11348 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11349 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11350 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11351 (lto_free_function_in_decl_state_for_node): New function.
11353 2013-08-29 Xinliang David Li <davidxl@google.com>
11355 * loop-unroll.c (report_unroll_peel): Minor message change.
11356 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11357 Emit alignment peeling message with default -fopt-info.
11358 (vect_loop_versioning): Emit loop version info message.
11359 * tree-vectorizer.c (vectorize_loops): Minor message change.
11360 (execute_vect_slp): Ditto.
11362 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11364 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11365 of the clone from the DECL_NAME of the original function.
11367 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11369 * passes.c (register_pass): Add overload.
11370 * tree-pass.h (register_pass): Forward declare it. Add comment.
11372 2013-08-29 Jan Hubicka <jh@suse.cz>
11374 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11375 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11376 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11377 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11378 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11379 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11380 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11381 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11383 2013-08-29 Teresa Johnson <tejohnson@google.com>
11385 * dumpfile.c (dump_loc): Output column number.
11386 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11387 * doc/invoke.texi: Document optall -fopt-info flag.
11388 * profile.c (read_profile_edge_counts): Use new dump framework.
11389 (compute_branch_probabilities): Ditto.
11390 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11391 when pass not in any opt group.
11392 * pass_manager.h (pass_manager::get_pass_profile): New method.
11393 * value-prof.c (check_counter): Use new dump framework.
11394 (check_ic_target): Ditto.
11395 * coverage.c (get_coverage_counts): Ditto.
11396 (coverage_init): Setup new dump framework.
11398 2013-08-29 Richard Biener <rguenther@suse.de>
11400 PR tree-optimization/58246
11401 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11402 handle the dominance check inside a basic-block.
11404 2013-08-29 Richard Biener <rguenther@suse.de>
11406 PR middle-end/57287
11407 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11408 of default defs that appear in abnormal PHI nodes.
11410 2013-08-29 Richard Biener <rguenther@suse.de>
11412 PR tree-optimization/57685
11413 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11414 single-use operands to avoid exponential complexity.
11416 2013-08-28 Dehao Chen <dehao@google.com>
11418 * ipa-inline.c (edge_badness): Fix integer underflow.
11420 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11422 * gtm-builtins.def (_ITM_free): Declare leaf.
11424 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11427 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11428 (*tls_local_dynamic_base_64_largepic): Likewise.
11429 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11430 Remove predicate from call operand.
11431 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11432 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11434 2013-08-28 Jeff Law <law@redhat.com>
11436 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11437 checks for the number of predecessors and successors allowed.
11438 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11439 which require copying a joiner block if there is a request which
11440 is a subpath that requires no joiner block copying.
11442 2013-08-28 Jan Hubicka <jh@suse.cz>
11444 * lto-streamer-out.c (DFS_write_tree_body): Drop
11445 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11446 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11447 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11448 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11449 Do not read DECL_ERROR_ISSUED.
11450 (unpack_ts_decl_with_vis_value_fields): Do not read
11452 (lto_input_ts_binfo_tree_pointers): Do not read
11453 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11454 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11455 write DECL_ERROR_ISSUED..
11456 (pack_ts_decl_with_vis_value_fields): Do not write
11458 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11459 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11460 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11461 * tree.h (tree_decl_common): Update comment.
11462 (DECL_ERROR_ISSUED): Remove.
11464 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11466 PR middle-end/58257
11467 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11469 2013-08-28 Jan Hubicka <jh@suse.cz>
11471 * builtins.def (free): Declare leaf.
11473 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11475 * gdbhooks.py: New.
11476 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11477 * configure: Regenerate.
11479 2013-08-27 Martin Jambor <mjambor@suse.cz>
11481 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11482 (ipa_ancestor_jf_data): Likewise.
11483 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11484 (ipa_get_jf_pass_through_type_preserved): New function.
11485 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11486 (ipa_get_jf_ancestor_type_preserved): New function.
11487 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11488 (ipa_get_jf_ancestor_result): Likewise.
11489 (propagate_vals_accross_pass_through): Use
11490 ipa_get_jf_pass_through_result to do all the value mappings.
11491 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11492 type_preserved flag.
11493 (ipa_set_jf_cst_copy): New function.
11494 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11495 (ipa_set_jf_arith_pass_through): Likewise.
11496 (ipa_set_ancestor_jf): Likewise.
11497 (compute_complex_assign_jump_func): Set type_preserved instead of
11499 (ipa_compute_jump_functions_for_edge): Likewise.
11500 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11501 (update_jump_functions_after_inlining): Update type_preserved.
11502 Explicitely create jump functions when combining one with pass_through.
11503 (ipa_write_jump_function): Stream the type_preserved flags.
11504 (ipa_read_jump_function): Likewise.
11506 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11507 Aldy Hernandez <aldyh@redhat.com>
11509 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11510 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11511 * function.h (struct function): Add has_force_vect_loops and
11513 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11514 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11516 * gimple.def: Update CLAUSES comments.
11517 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11518 (gimple_build_omp_for): Add argument to prototype.
11519 (gimple_omp_for_kind): New.
11520 (gimple_omp_for_set_kind): New.
11521 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11522 GOVD_DATA_SHARE_CLASS.
11523 (enum omp_region_type): Add ORT_SIMD.
11524 (gimple_add_tmp_var): Handle ORT_SIMD.
11525 (gimplify_var_or_parm_decl): Same.
11526 (is_gimple_stmt): Same.
11527 (omp_firstprivatize_variable): Same.
11528 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11529 (omp_notice_variable): Handle ORT_SIMD.
11530 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11531 (omp_check_private): Handle ORT_SIMD.
11532 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11533 OMP_CLAUSE_SAFELEN.
11534 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11535 Handle OMP_CLAUSE_LASTPRIVATE.
11536 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11537 OMP_CLAUSE_SAFELEN.
11538 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11539 (gimplify_expr): Handle OMP_SIMD.
11540 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11541 (expand_GOMP_SIMD_VF): New.
11542 (expand_GOMP_SIMD_LAST_LANE): New.
11543 * internal-fn.def (GOMP_SIMD_LANE): New.
11544 (GOMP_SIMD_VF): New.
11545 (GOMP_SIMD_LAST_LANE): New.
11546 * omp-low.c: Include target.h.
11547 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11548 OMP_CLAUSE_SAFELEN.
11549 (check_omp_nesting_restrictions): Same.
11551 (lower_rec_simd_input_clauses): New.
11552 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11554 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11555 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11556 (expand_omp_build_assign): New.
11557 (expand_omp_for_init_counts): New.
11558 (expand_omp_for_init_vars): New.
11559 (extract_omp_for_update_vars): New.
11560 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11561 and rewrite accordingly.
11562 (expand_omp_simd): New.
11563 (expand_omp_for): Use expand_omp_simd.
11564 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11565 (lower_omp_for): Do not lower the body.
11566 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11567 in their own loops.
11568 * tree-flow.h (find_omp_clause): Remove prototype.
11569 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11570 forcing vectorization of the loop, or if flag_tree_vectorize.
11571 (gate_tree_if_conversion): Similarly.
11572 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11573 gimple_build_omp_for.
11574 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11575 * tree-parloops (create_parallel_loop): Pass kind argument to
11576 gimple_build_omp_for.
11577 * tree-pretty-print.c (dump_omp_clause): Add cases for
11578 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11579 OMP_CLAUSE__SIMDUID_.
11580 (dump_generic_node): Handle OMP_SIMD.
11581 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11582 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11583 unroll OMP_SIMD loops here.
11584 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11585 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11587 (vect_analyze_data_refs): Handle simd loops.
11588 * tree-vect-loop.c (vectorizable_live_operation): Handle
11590 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11591 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11592 (vectorizable_load): Same.
11593 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11594 (struct simduid_to_vf): New.
11595 (simduid_to_vf::hash): New.
11596 (simduid_to-vf::equal): New.
11597 (struct simd_array_to_simduid): New.
11598 (simd_array_to_simduid::hash): New.
11599 (simd_array_to_simduid::equal): New.
11600 (adjust_simduid_builtins): New.
11601 (struct note_simd_array_uses_struct): New.
11602 (note_simd_array_uses_cb): New.
11603 (note_simd_array_uses): New.
11604 (vectorize_loops): Handle simd hints and adjust simd builtins
11606 * tree-vectorizer.h (struct _stmt_vec_info): Add
11607 simd_lane_access_p field.
11608 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11609 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11610 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11611 (omp_clause_code_name): Same.
11612 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11613 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11614 * tree.def (OMP_SIMD): New entry.
11615 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11616 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11617 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11618 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11619 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11620 (OMP_CLAUSE_LINEAR_STEP): New.
11621 (OMP_CLAUSE_SAFELEN_EXPR): New.
11622 (OMP_CLAUSE__SIMDUID__DECL): New.
11623 (find_omp_clause): New prototype.
11625 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11627 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11628 Haswell processor detection.
11630 2013-08-27 Christian Widmer <shadow@umbrox.de>
11633 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11634 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11635 AVX2 capable processors.
11637 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11639 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11640 of vget_low_* with implementations in terms of other intrinsics.
11642 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11644 PR middle-end/57219
11645 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11646 values to -1, 0 and 1.
11648 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11650 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11651 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11652 (<optab><mode>3_insn): Remove.
11653 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11654 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11655 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11656 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11657 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11658 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11659 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11660 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11661 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11662 (ror<mode>3_insn): Likewise.
11663 * config/aarch64/predicates.md (aarch64_simd_register): New.
11665 2013-08-27 Richard Biener <rguenther@suse.de>
11667 PR tree-optimization/57521
11668 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11669 one edge is non-critical.
11670 (find_phi_replacement_condition): Make sure to use a non-critical
11671 edge. Cleanup and remove old bug workarounds.
11672 (bb_postdominates_preds): Remove.
11673 (if_convertible_loop_p_1): Do not compute post-dominators.
11674 (combine_blocks): Do not free post-dominators.
11675 (main_tree_if_conversion): Likewise.
11676 (pass_data_if_conversion): Add TODO_verify_ssa.
11678 2013-08-27 DJ Delorie <dj@redhat.com>
11680 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11682 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11684 * function.c (assign_parm_find_data_types): Set passed_mode and
11685 nominal_mode to the TYPE_MODE of nominal_type for the built
11686 pointer type in case of the struct-pass-by-reference.
11688 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11690 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11691 (UINT16_TYPE): Change default to "unsigned int".
11693 * config/avr/avr.opt (mfract-convert-truncate): New option.
11694 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11695 is set, round negative fractional integers according to n1169
11696 when converting to integer types.
11698 2013-08-26 Jan Hubicka <jh@suse.cz>
11700 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11701 methods can not be called indirectly when their address is not taken.
11703 2013-08-26 Jan Hubicka <jh@suse.cz>
11705 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11708 2013-08-26 Jan Hubicka <jh@suse.cz>
11710 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11713 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11715 * reload.c (find_valid_class): Allow classes that do not include
11716 FIRST_PSEUDO_REGISTER - 1.
11718 2013-08-26 Jan Hubicka <jh@suse.cz>
11720 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11721 fix edge count/frequency when speculation failed; fix type check
11722 for the direct call.
11724 2013-08-26 Jan Hubicka <jh@suse.cz>
11726 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11728 2013-08-26 Jan Hubicka <jh@suse.cz>
11730 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11731 to be changed by edge redirection.
11733 2013-08-26 Jan Hubicka <jh@suse.cz>
11735 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11736 formating; add sanity check.
11737 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11738 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11739 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11740 (resolve_noninline_speculation): Update callee keys, too.
11742 2013-08-26 Jan Hubicka <jh@suse.cz>
11744 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11745 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11747 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11749 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11752 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11754 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11755 (x86_64_elf_select_section): Ditto.
11756 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11758 (ix86_pass_by_reference): Ditto.
11759 (output_set_got): Ditto.
11760 (ix86_unary_operator_ok): Ditto.
11761 (ix86_expand_builtin): Ditto.
11763 2013-08-23 Jan Hubicka <jh@suse.cz>
11765 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11767 2013-08-23 Jan Hubicka <jh@suse.cz>
11769 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11770 (tree_decl_with_vis): Add FINAL field.
11772 2013-08-23 Jeff Law <law@redhat.com>
11774 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11775 the debugging dump when the expression is fully redundant.
11777 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11779 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11780 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11781 * pretty-print.c (pp_formatted_text_data): Likewise.
11782 (pp_write_text_to_stream): Likewise.
11783 (pp_write_text_as_dot_label_to_stream): Likewise.
11784 (pp_append_r): Likewise.
11785 (pp_format): Likewise.
11786 (pp_flush): Likewise.
11787 (pp_clear_output_area): Likewise.
11788 (pp_append_text): Likewise.
11789 (pp_formatted_text): Likewise.
11790 (pp_remaining_character_count_for_line): Likewise.
11791 (pp_newline): Likewise.
11792 (pp_character): Likewise.
11793 (output_buffer::~output_buffer): Define.
11794 (pretty_printer::~pretty_printer): Destruct output buffer.
11795 * pretty-print.h (output_buffer::~output_buffer): Declare.
11796 (pretty_printer::~pretty_printer): Declare virtual.
11798 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11801 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11802 HOST_WIDE_INT_M1U): New macros.
11803 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11804 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11805 unsigned -1 for lshift.
11806 * cse.c (cse_insn): Likewise.
11807 * double-int.c (rshift_double, lshift_double): Likewise.
11808 * builtins.c (fold_builtin_bitop): Likewise.
11809 * combine.c (force_to_mode): Likewise.
11810 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11811 * simplify-rtx.c (simplify_const_unary_operation,
11812 simplify_const_binary_operation): Likewise.
11813 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11814 check_va_list_escapes): Likewise.
11815 * rtlanal.c (nonzero_bits1): Likewise.
11816 * expmed.c (expand_smod_pow2): Likewise.
11817 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11819 2013-08-23 Jan Hubicka <jh@suse.cz>
11821 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11822 as having address taken.
11824 2013-08-23 Jan Hubicka <jh@suse.cz>
11826 * ipa-utils.h (method_class_type): Declare.
11827 * ipa-devirt.c (method_class_type): Export.
11829 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11830 do not walk base classes of anonymous types.
11832 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11834 PR rtl-optimization/58220
11835 PR regression/58221
11836 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11837 handle SEQUENCE insns properly.
11839 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11841 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11843 (pp_newline_and_indent): Likewise.
11844 (pp_separate_with): Likewise.
11845 * pretty-print.c (pp_newline_and_flush): Define.
11846 (pp_newline_and_indent): Likewise.
11847 (pp_separate_with): Likewise.
11849 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11852 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11853 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11855 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11857 * config/i386/predicates.md (ext_sse_reg_operand): New.
11858 * config/i386/i386.md (*movti_internal): Use
11859 predicate to determine if EVEX is needed.
11860 (*movsi_internal): Ditto.
11861 (*movdf_internal): Ditto.
11862 (*movsf_internal): Ditto.
11863 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11865 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11867 PR tree-optimization/58209
11868 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11869 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11870 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11871 emit POINTER_PLUS_EXPR.
11872 (create_tailcall_accumulator): For pointer result type accumulate in
11875 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11877 * configure.ac: Add backslashes missing from the last change.
11878 * configure: Regenerate.
11880 2013-08-22 Jan Hubicka <jh@susue.cz>
11882 * ipa.c (function_and_variable_visibility): First remember function
11883 was global and then make it local.
11885 2013-08-22 Julian Brown <julian@codesourcery.com>
11887 * configure.ac: Add aarch64 to list of arches which use "nop" in
11889 * configure: Regenerate.
11891 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11893 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11894 gnu_libc_has_function.
11895 * config/s390/tpf.h: Likewise.
11897 2013-08-22 Jan Hubicka <jh@susue.cz>
11899 * timevar.c (validate_phases): Add cast.
11901 2013-08-22 Jan Hubicka <jh@susue.cz>
11903 * timevar.c (validate_phases): Use size_t for memory.
11904 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11906 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11908 * pretty-print.h (output_buffer::output_buffer): Declare.
11909 (pretty_printer::pretty_printer): Likewise.
11910 (pp_construct): Remove.
11911 * pretty-print.c (output_buffer::output_buffer): Define.
11912 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11913 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11914 (print_gimple_expr): Likewise.
11915 (print_gimple_seq): Likewise.
11916 (gimple_dump_bb): Likewise.
11917 * sched-vis.c (dump_value_slim): Likewise.
11918 (dump_insn_slim): Likewise.
11919 (dump_rtl_slim): Likewise.
11920 (str_pattern_slim): Likewise.
11921 * tree-mudflap.c (mf_varname_tree): Likewise.
11922 * graph.c (print_graph_cfg): Likewise.
11923 (start_graph_dump): Likewise.
11924 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11926 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11927 pretty printer initialization.
11928 * coretypes.h (diagnostic_context): Remove superflous type alias
11930 (pretty_printer): Likewise. Declare directly as a class.
11931 (pretty_print_info): Remove declaration as class.
11932 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11933 and pp_clear_output_area.
11934 (asan_add_global): Likewise.
11936 2013-08-22 Jan Hubicka <jh@suse.cz>
11938 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11939 * ipa-utils.h (update_type_inheritance_graph): Declare.
11940 (possible_polymorphic_call_target_p): Declare.
11941 (possible_polymorphic_call_target_p): New.
11942 * ipa-devirt.c: Update toplevel comments.
11943 (cached_polymorphic_call_targets): Move up.
11944 (odr_type_d): Move ID down.
11945 (polymorphic_type_binfo_p): Update comment.
11946 (odr_hasher::remove): Likewise;
11947 (get_odr_type): Set anonymous_namespace.
11948 (dump_odr_type): Dump it.
11949 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11950 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11951 (record_binfo): Add comment.
11952 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11954 (devirt_node_removal_hook): Do not iCE when cache is freed.
11955 (possible_polymorphic_call_target_p): New predicate.
11956 (update_type_inheritance_graph): New function.
11958 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11959 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11960 Sergey Lega <sergey.s.lega@intel.com>
11961 Anna Tikhonova <anna.tikhonova@intel.com>
11962 Ilya Tocar <ilya.tocar@intel.com>
11963 Andrey Turetskiy <andrey.turetskiy@intel.com>
11964 Ilya Verbin <ilya.verbin@intel.com>
11965 Kirill Yukhin <kirill.yukhin@intel.com>
11966 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11968 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11969 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11970 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11971 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11972 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11973 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11974 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11975 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11976 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11977 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11978 OPT_mavx512pf, OPT_mavx512er cases.
11979 * config/i386/constraints.md (v): New constraint.
11980 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11981 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11982 (bit_AVX512CD): New.
11983 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11984 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11985 * config/i386/i386-c.c (ix86_target_macros_internal):
11986 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11988 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11989 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11990 * config/i386/i386.c (regclass_map, dbx_register_map)
11991 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11992 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11993 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11994 -mavx512pf options.
11995 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11996 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11997 -mavx512pf options. Fix formatting.
11998 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11999 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
12000 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
12001 -mavx512cd, -mavx512pf options.
12002 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
12003 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
12004 (ix86_preferred_output_reload_class): Replace SSE_REGS with
12006 (ix86_hard_regno_mode_ok): Support 512-bit registers.
12007 (ix86_set_reg_reg_cost): Ditto.
12008 (x86_order_regs_for_local_alloc): Ditto.
12009 (MAX_VECT_LEN): Extend to 64-byte.
12010 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
12011 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
12012 (TARGET_AVX512ER, TARGET_AVX512CD): New.
12013 (BIGGEST_ALIGNMENT): Extend to 512-bits.
12014 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
12015 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
12016 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
12017 (SSE_REG_MODE_P): Support new modes.
12018 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
12019 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
12020 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
12021 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
12022 (REG_CLASS_CONTENTS): Add new registers.
12023 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
12024 (EXT_REX_SSE_REGNO_P): New.
12025 (HI_REGISTER_NAMES): Add new registers.
12026 * config/i386/i386.md: Define constants for new registers.
12027 (mode): Add new 512-bit modes.
12028 (prefix): Support evex prefix.
12029 (isa): Support avx512f, noavx512f, fma_avx512f.
12030 (ssemodesuffix): Add new 512-bit modes.
12032 (*movxi_internal_avx512f): Ditto.
12033 (*movdi_internal): Replace constraint "x" with the new constraint "v".
12035 (*movsi_internal): Likewise.
12036 (*movdf_internal): Likewise.
12037 (*movsf_internal): Likewise.
12038 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
12039 (<code><mode>3): Likewise.
12040 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
12042 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
12043 with the new constraint "v".
12044 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
12046 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
12047 with the new constraint "v".
12048 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
12049 (<sse2>_storedqu<avxsizesuffix>): Likewise.
12050 (*<plusminus_insn><mode>3): Likewise.
12051 (<sse>_vm<plusminus_insn><mode>3): Likewise.
12052 (*mul<mode>3): Likewise.
12053 (<sse>_vmmul<mode>3): Likewise.
12054 (<sse>_div<mode>3): Likewise.
12055 (<sse>_vmdiv<mode>3): Likewise.
12056 (<sse>_sqrt<mode>2): Likewise.
12057 (<sse>_vmsqrt<mode>2): Likewise.
12058 (*<code><mode>3_finite): Likewise.
12059 (*<code><mode>3) <smaxmin>: Likewise.
12060 (<sse>_vm<code><mode>3): Likewise.
12061 (*<code><mode>3) <any_logic>: Likewise.
12062 (*fma_fmadd_<mode>): Likewise.
12063 (*fma_fmsub_<mode>): Likewise.
12064 (*fma_fnmadd_<mode>): Likewise.
12065 (*fma_fnmsub_<mode>): Likewise.
12066 (*fma_fmaddsub_<mode>): Likewise.
12067 (*fma_fmsubadd_<mode>): Likewise.
12068 (*fmai_fmadd_<mode>): Likewise.
12069 (*fmai_fmsub_<mode>): Likewise.
12070 (*fmai_fnmadd_<mode>): Likewise.
12071 (*fmai_fnmsub_<mode>): Likewise.
12072 (sse_cvtsi2ss): Likewise.
12073 (sse_cvtsi2ssq): Likewise.
12074 (sse_cvtss2si): Likewise.
12075 (sse_cvtss2si_2): Likewise.
12076 (sse_cvtss2siq): Likewise.
12077 (sse_cvtss2siq_2): Likewise.
12078 (sse_cvttss2si): Likewise.
12079 (sse_cvtss2siq_2): Likewise.
12080 (float<sseintvecmodelower><mode>2): Likewise.
12081 (sse2_cvtsd2si_2): Likewise.
12082 (sse2_cvtsd2siq_2): Likewise.
12083 (*<plusminus_insn><mode>3): Likewise.
12084 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
12085 (*<sse4_1_avx2>_mul<mode>3): Likewise.
12086 (ashr<mode>3): Likewise.
12087 (<shift_insn><mode>3): Likewise.
12088 (avx2_<code><mode>3): Likewise.
12089 (*avx2_<code><mode>3): Likewise.
12090 (*andnot<mode>3): Likewise.
12091 (*<code><mode>3) <any_logic>: Likewise.
12092 (abs<mode>2): Likewise.
12093 (avx2_permvar<mode>): Likewise.
12094 (avx2_perm<mode>_1): Likewise.
12095 (*avx_vpermilp<mode>): Likewise.
12096 (avx_vpermilvar<mode>3): Likewise.
12097 (avx2_ashrv<mode>): Likewise.
12098 (avx2_<shift_insn>v<mode>): Likewise.
12099 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
12101 * doc/rtl.texi: Document XImode.
12103 2013-08-21 Jeff Law <law@redhat.com>
12105 * tree-flow.h (register_jump_thread): Pass vector of edges
12106 instead of each important edge.
12107 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
12108 thread path into a vector and pass that to register_jump_thread.
12109 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
12110 passed in edge vector to the current 3-edge form.
12113 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12115 * dce.c (fini_dce): Call df_analyze again just in case
12116 delete_unmarked_insns removed anything.
12118 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
12120 * reload.h (struct reg_equivs): Rename to ..
12121 (struct reg_equivs_s): .. this.
12123 2013-08-20 Martin Liska <marxin.liska@gmail.com>
12125 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
12127 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12129 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
12131 2013-08-21 Jeff Law <law@redhat.com>
12133 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
12134 simplify assignments too. If the RHS collapses to a singleton
12135 range, then return the value for the range.
12137 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
12139 * config/i386/sse.md (V16): Rename to...
12141 (mov<mode>): Update iterator name.
12142 (*mov<mode>_internal): Ditto.
12143 (push<mode>1): Ditto.
12144 (movmisalign<mode>): Ditto.
12146 2013-08-20 Jan Hubicka <jh@suse.cz>
12149 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
12150 entry for direct edges.
12151 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
12153 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12155 Revert my last two changes, r201865 and r201864:
12158 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12160 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12161 instances can own GC refs.
12163 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12164 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12165 (gcc::context::gt_pch_nx): Likewise.
12166 (gcc::context::gt_pch_nx): Likewise.
12167 * ggc.h (gt_ggc_mx <T>): New.
12168 (gt_pch_nx_with_op <T>): New.
12169 (gt_pch_nx <T>): New.
12170 * passes.c (opt_pass::gt_ggc_mx): New.
12171 (opt_pass::gt_pch_nx): New.
12172 (opt_pass::gt_pch_nx_with_op): New.
12173 (pass_manager::gt_ggc_mx): New.
12174 (pass_manager::gt_pch_nx): New.
12175 (pass_manager::gt_pch_nx_with_op): New.
12176 (pass_manager::operator new): Use
12177 ggc_internal_cleared_alloc_stat rather than xcalloc.
12178 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12179 (pass_manager::gt_ggc_mx): New.
12180 (pass_manager::gt_pch_nx): New.
12181 (pass_manager::gt_pch_nx_with_op): New.
12182 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12183 (opt_pass::operator new): New.
12184 (opt_pass::gt_ggc_mx): New.
12185 (opt_pass::gt_pch_nx): New.
12186 (opt_pass::gt_pch_nx_with_op): New.
12189 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12191 * Makefile.in (GTFILES): Add context.h.
12192 * context.c (gcc::context::operator new): New.
12193 (gcc::context::gt_ggc_mx): New.
12194 (gcc::context::gt_pch_nx): New.
12195 (gcc::context::gt_pch_nx): New.
12196 * context.h (gcc::context): Add GTY((user)) marking.
12197 (gcc::context::operator new): New.
12198 (gcc::context::gt_ggc_mx): New.
12199 (gcc::context::gt_pch_nx): New.
12200 (gcc::context::gt_pch_nx): New.
12201 (g): Add GTY marking.
12202 (gt_ggc_mx (gcc::context *)): New.
12203 (gt_pch_nx (gcc::context *)): New.
12204 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12205 void *cookie)): New.
12206 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12208 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12210 * dce.c (fini_dce): Call df_analyze again just in case
12211 delete_unmarked_insns removed anything.
12213 2013-08-20 Teresa Johnson <tejohnson@google.com>
12215 PR rtl-optimizations/57451
12216 * final.c (reemit_insn_block_notes): Prevent lexical blocks
12217 from crossing split section boundaries.
12219 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12221 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
12222 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
12223 with MULTLIB_DEFAULTS.
12225 2013-08-20 Nick Clifton <nickc@redhat.com>
12227 * target.def (narrow_volatile_bitfield): Note that the default
12228 value is false, not !TARGET_STRICT_ALIGN.
12229 * doc/tm.texi: Regenerate.
12231 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
12233 Fix LIB_SPEC for systems without libpthread.
12235 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12236 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
12238 * config/i386/linux-common.h: Likewise.
12239 * config/mips/linux-common.h: Likewise.
12241 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12243 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
12246 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12248 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12249 instances can own GC refs.
12251 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12252 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12253 (gcc::context::gt_pch_nx): Likewise.
12254 (gcc::context::gt_pch_nx): Likewise.
12255 * ggc.h (gt_ggc_mx <T>): New.
12256 (gt_pch_nx_with_op <T>): New.
12257 (gt_pch_nx <T>): New.
12258 * passes.c (opt_pass::gt_ggc_mx): New.
12259 (opt_pass::gt_pch_nx): New.
12260 (opt_pass::gt_pch_nx_with_op): New.
12261 (pass_manager::gt_ggc_mx): New.
12262 (pass_manager::gt_pch_nx): New.
12263 (pass_manager::gt_pch_nx_with_op): New.
12264 (pass_manager::operator new): Use
12265 ggc_internal_cleared_alloc_stat rather than xcalloc.
12266 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12267 (pass_manager::gt_ggc_mx): New.
12268 (pass_manager::gt_pch_nx): New.
12269 (pass_manager::gt_pch_nx_with_op): New.
12270 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12271 (opt_pass::operator new): New.
12272 (opt_pass::gt_ggc_mx): New.
12273 (opt_pass::gt_pch_nx): New.
12274 (opt_pass::gt_pch_nx_with_op): New.
12276 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12278 * Makefile.in (GTFILES): Add context.h.
12279 * context.c (gcc::context::operator new): New.
12280 (gcc::context::gt_ggc_mx): New.
12281 (gcc::context::gt_pch_nx): New.
12282 (gcc::context::gt_pch_nx): New.
12283 * context.h (gcc::context): Add GTY((user)) marking.
12284 (gcc::context::operator new): New.
12285 (gcc::context::gt_ggc_mx): New.
12286 (gcc::context::gt_pch_nx): New.
12287 (gcc::context::gt_pch_nx): New.
12288 (g): Add GTY marking.
12289 (gt_ggc_mx (gcc::context *)): New.
12290 (gt_pch_nx (gcc::context *)): New.
12291 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12292 void *cookie)): New.
12293 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12295 2013-08-20 Alan Modra <amodra@gmail.com>
12298 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
12299 (rs6000_emit_epilogue): Likewise.
12301 2013-08-19 Dehao Chen <dehao@google.com>
12303 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12305 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12306 Jakub Jelinek <jakub@redhat.com>
12308 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12309 (BUILT_IN_FABSD64): Likewise.
12310 (BUILT_IN_FABSD128): Likewise.
12311 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12312 (fold_builtin_1): Likewise.
12313 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12314 destination and source operands.
12315 (*abstd2_fpr): Likewise.
12316 (*nabstd2_fpr): Likewise.
12318 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12320 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12323 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12325 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12328 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12330 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12331 * builtins.c (default_libc_has_function): New.
12332 (gnu_libc_has_function): Ditto.
12333 (no_c99_libc_has_function): Ditto.
12334 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12335 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12336 (fold_builtin_sincos): Likewise.
12337 (fold_builtin_cexp): Likewise.
12338 * builtins.def (DEF_C94_BUILTIN): Likewise.
12339 (DEF_C99_BUILTIN): Likewise.
12340 (DEF_C99_C90RES_BUILTIN): Likewise.
12341 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12342 definitions to using this define.
12343 * config/darwin-protos.h (darwin_libc_has_function): New.
12344 * config/darwin.c (darwin_libc_has_function): Ditto.
12345 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12346 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12347 * config/darwin.h: Ditto.
12348 * config/elfos.h: Ditto.
12349 * config/freebsd.h: Ditto.
12350 * config/i386/cygming.h: Ditto.
12351 * config/i386/djgpp.h: Ditto.
12352 * config/i386/i386-interix.h: Ditto.
12353 * config/microblaze/microblaze.h: Ditto.
12354 * config/mmix/mmix.h: Ditto.
12355 * config/gnu-user.h: Ditto.
12356 * config/ia64/hpux.h: Ditto.
12357 * config/pa/pa-hpux.h: Ditto.
12358 * config/pdp11/pdp11.h: Ditto.
12359 * config/picochip/picochip.h: Ditto.
12360 * config/linux.h: Ditto.
12361 * config/netbsd.h: Ditto.
12362 * config/openbsd.h: Ditto.
12363 * config/rs6000/aix43.h: Ditto.
12364 * config/rs6000/aix51.h: Ditto.
12365 * config/rs6000/aix52.h: Ditto.
12366 * config/rs6000/aix53.h: Ditto.
12367 * config/rs6000/aix61.h: Ditto.
12368 * config/rs6000/darwin.h: Ditto.
12369 * config/rs6000/linux.h: Ditto.
12370 * config/rs6000/linux64.h: Ditto.
12371 * config/s390/tpf.h: Ditto.
12372 * config/sol2-10.h: Ditto.
12373 * config/sol2.h: Ditto.
12374 * config/vms/vms.h: Ditto.
12375 * config/vxworks.h: Ditto.
12376 * config/linux-android.c (linux_android_libc_has_function):
12377 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12378 * config/linux-protos.h (linux_android_libc_has_function):
12380 * config/i386/i386.c (ix86_libc_has_function): New.
12381 * config/i386/i386-protos.h
12382 (ix86_libc_has_function): New declaration.
12383 * config/i386/i386.md
12384 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12385 ("isinf<mode>2): Likewise.
12386 * convert.c (convert_to_integer): Using new target hook
12387 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12388 TARGET_C99_FUNCTIONS.
12389 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12390 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12391 * coretypes.h (function_class): New enum for different
12392 classes of functions.
12393 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12394 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12395 (TARGET_HAS_SINCOS): Likewise.
12396 (TARGET_LIBC_HAS_FUNCTION): New.
12397 * doc/tm.texi: Regenerated.
12398 * targhooks.h (default_libc_has_function): New declaration.
12399 (no_c99_libc_has_function): Ditto.
12400 (gnu_libc_has_function): Ditto.
12401 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12402 and TARGET_HAS_SINCOS.
12404 2013-08-18 Jan Hubicka <jh@suse.cz>
12406 * Makeifle-in (ipa-devirt.o): New.
12407 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12408 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12409 (analyze_functions): Look into possible targets of polymorphic call.
12410 * dumpfile.c (dump_files): Add type-inheritance dump.
12411 * dumpfile.h (TDI_inheritance): New.
12412 * ipa-devirt.c: New file.
12413 * ipa-utils.h (odr_type_d): Forward declare.
12414 (odr_type): New type.
12415 (build_type_inheritance_graph): Declare.
12416 (possible_polymorphic_call_targets): Declare and introduce inline
12417 variant when only edge is pased.
12418 (dump_possible_polymorphic_call_targets): Likewise.
12419 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12420 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12421 (types_same_for_odr): ... here.
12422 * tree.h (type_in_anonymous_namespace_p): Declare.
12424 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12426 PR tree-optimization/58006
12427 * tree-parloops.c (take_address_of): Don't ICE if get_name
12429 (eliminate_local_variables_stmt): Remove clobber stmts.
12431 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12433 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12434 error message is issued for an alias to undefined symbol.
12436 2013-08-18 Jan Hubicka <jh@suse.cz>
12438 * cgraph.c (cgraph_create_indirect_edge): Discover
12439 polymorphic calls and record basic info into indirect_info.
12440 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12441 devirtualization, ignore objc function calls.
12442 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12443 call with no parm index info.
12444 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12445 * tree.c (virtual_method_call_p): New function.
12446 * tree.h (virtual_method_call_p): Declare.
12448 2013-08-16 Jan Hubicka <jh@suse.cz>
12450 PR middle-end/58179
12451 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12453 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12455 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12458 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12460 * gengtype.c (type_for_name): Add special-case support for
12461 locating types within the "gcc::" namespace.
12462 (open_base_files): Emit a "using namespace gcc" directive.
12464 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12467 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12468 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12470 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12471 array instead of each individual operand as a separate argument.
12472 (emit_fusion_gpr_load): Likewise.
12473 (expand_fusion_gpr_load): Add new function declaration.
12475 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12476 signature to have the operands passed as an array, instead of as
12477 separate arguments. Allow ZERO_EXTEND to be in the memory
12478 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12479 depend on the register live/dead flags when peepholes are run.
12480 (expand_fusion_gpr_load): New function to be called from the
12481 peephole2 pass, to change the register that addis sets to be the
12483 (emit_fusion_gpr_load): Change the calling signature to have the
12484 operands passed as an array, instead of as separate arguments.
12485 Allow ZERO_EXTEND to be in the memory address, and also
12486 SIGN_EXTEND if -mpower8-fusion-sign.
12488 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12489 unspec enumeration.
12490 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12491 adjust the register addis loads up in the peephole2 pass. Do not
12492 depend on the register live/dead state when the peephole pass is done.
12494 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12496 * gengtype.c (create_user_defined_type): Ensure that the kind
12497 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12498 declaration is seen before the GTY((user)) marking.
12500 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12503 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12505 2013-08-16 Jan Hubicka <jh@suse.cz>
12507 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12508 arugment expected_type.
12509 (gimple_fold_call): Use it.
12510 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12511 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12512 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12513 (try_make_edge_direct_virtual_call): Likewise.
12514 * tree.c (obj_type_ref_class): New.
12515 * tree.h (obj_type_ref_class): Use it.
12517 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12519 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12520 (rtl_slim_pp): Likewise.
12521 (init_rtl_slim_pretty_print): Likewise.
12522 (dump_value_slim): Don't call it. Use local pretty printer.
12523 (dump_insn_slim): Likewise.
12524 (dump_rtl_slim): Likewise.
12525 (str_pattern_slim): Likewise.
12526 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12529 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12531 PR tree-optimization/58164
12532 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12533 walk gimple_goto_dest of GIMPLE_GOTO.
12535 PR tree-optimization/58165
12536 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12537 bi_call must be the last stmt in a bb, don't split_block, instead
12538 use fallthru edge from it and give up if there is none.
12539 Release conds vector when returning early.
12541 2013-08-14 Xinliang David Li <davidxl@google.com>
12543 * config/i386/i386.c (ix86_option_override_internal):
12544 Remove unused variable and field.
12546 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12549 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12550 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12551 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12552 and Linux, correct BLKmode alignment when 128-bit alignment is
12553 required and compatibility flag is not set.
12554 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12555 for zero-size arguments when compatibility flag is not set.
12557 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12559 PR tree-optimization/58145
12560 * tree-sra.c (build_ref_for_offset): If prev_base has
12561 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12563 2013-08-14 Xinliang David Li <davidxl@google.com>
12565 * config/i386/i386.c (ix86_option_override_internal):
12566 Fix uninitialized variable error.
12568 2013-08-14 Xinliang David Li <davidxl@google.com>
12570 * config/i386/i386.opt: Define two new options.
12571 * config/i386/x86-tune.def: Add arch selector field in macros.
12572 * config/i386/i386.h: Adjust macro definition.
12573 * config/i386/i386.c (ix86_option_override_internal):
12575 (parse_mtune_ctrl_str): New function.
12576 (set_ix86_tune_features): New function.
12577 (ix86_function_specific_restore): Call the new helper function.
12579 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12581 PR rtl-optimization/57662
12582 * sel-sched.c (code_motion_process_successors): When the current insn
12583 is removed after the recursive traversal, break from the loop.
12584 Add comments and debug printouts.
12586 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12587 Alexandre Oliva <aoliva@redhat.com>
12590 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12591 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12592 there also UNSPEC_PLTOFF.
12594 2013-08-14 Marek Polacek <polacek@redhat.com>
12596 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12597 NUM_CONDITIONS bit positions.
12599 2013-08-13 Cary Coutant <ccoutant@google.com>
12601 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12602 (attr_checksum): Hash vector contents instead of pointer.
12603 (attr_checksum_ordered): Likewise.
12605 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12607 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12608 when Pmode != word_mode. Add length_address attribute.
12609 (sse3_monitor_<mode>): Merge from sse3_monitor and
12610 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12611 Pmode != word_mode. Update insn length attribute.
12612 * config/i386/i386.c (ix86_option_override_internal): Update
12613 ix86_gen_monitor selection for merged sse3_monitor insn.
12615 2013-08-13 Julian Brown <julian@codesourcery.com>
12617 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12618 perform invalid legitimization on greater-than-word-size modes for
12619 TARGET_E500_DOUBLE.
12621 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12623 * ira.c (setup_class_translate_array): Use aclass instead of cl
12624 for classes not fully covered by allocno classes.
12626 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12628 PR tree-optimization/57661
12629 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12630 * tree-inline.c (tree_function_versioning): Initialize it.
12631 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12632 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12633 that is not being copied.
12636 * asan.c (instrument_strlen_call): Fix typo in comment.
12637 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12639 2013-08-13 Steve Ellcey <sellcey@mips.com>
12641 * config/mips/mips.md (prefetch): Use lw instead of ld on
12642 loongson in 32bit mode.
12644 2013-08-13 Nick Clifton <nickc@redhat.com>
12646 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12648 2013-08-13 Jan Hubicka <jh@suse.cz>
12650 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12651 introduced edge; fix typo in sanity check.
12652 (cgraph_resolve_speculation): Export; improve diagnostic.
12653 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12654 speculation at type mismatch.
12655 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12656 (cgraph_resolve_speculation): Declare.
12657 (symtab_can_be_discarded): New function.
12658 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12659 * ipa-inline-transform.c (speculation_removed): New global var.
12660 (clone_inlined_nodes): See if speculation can be removed.
12661 (inline_call): If speculations was removed, we growths may not match.
12662 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12663 (speculation_useful_p): New function.
12664 (resolve_noninline_speculation): New function.
12665 (inline_small_functions): Resolve useless speculations.
12666 * ipa-inline.h (speculation_useful_p): Declare
12667 * ipa.c (can_replace_by_local_alias): Simplify.
12668 (ipa_profile): Produce speculative calls in non-lto, too;
12669 add simple cost model; produce local aliases.
12671 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12673 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12676 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12678 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12680 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12682 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12683 iterator instead of X87MODEF.
12685 2013-08-12 Perez Read <netfirewall@gmail.com>
12688 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12689 operand 0 for intel asm alternative.
12690 (*movabs<mode>_2): Ditto for operand 1.
12692 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12694 * config/aarch64/arm_none.h
12695 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12697 2013-08-12 Nick Clifton <nickc@redhat.com>
12699 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12701 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12703 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12704 expand for QI/HImode operand to produce more effictive code for
12705 unsigned char(short) --> float(double) conversion.
12707 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12709 * doc/invoke.texi: Mention that -ftls-model does not force the final
12712 2013-08-12 Marek Polacek <polacek@redhat.com>
12713 Marc Glisse <marc.glisse@inria.fr>
12715 PR tree-optimization/57980
12716 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12717 when creating -1 constant.
12719 2013-08-10 Jan Hubicka <jh@suse.cz>
12721 Workaround binutils PR14342.
12722 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12723 (gimple_init_edge_profiler): Likewise.
12724 (gimple_gen_ic_func_profiler): Likewise.
12726 2013-08-09 Jan Hubicka <jh@suse.cz>
12728 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12730 2013-08-09 Xinliang David Li <davidxl@google.com>
12732 * config/i386/stringop.def: New file.
12733 * config/i386/stringop.opt: New file.
12734 * config/i386/i386-opts.h: Include stringopt.def.
12735 * config/i386/i386.opt: Include stringopt.opt.
12736 * config/i386/i386.c (ix86_option_override_internal):
12737 Override default size based stringop inline strategies with options.
12738 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12741 2013-08-09 Jan Hubicka <jh@suse.cz>
12743 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12745 2013-08-09 Jan Hubicka <jh@suse.cz>
12747 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12749 (dump_cgraph_node): Dump profile-id.
12750 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12751 and common_target_probability.
12752 * lto-cgraph.c (lto_output_edge): Stream common targets.
12753 (lto_output_node): Stream profile ids.
12754 (input_node): Stream profile ids.
12755 (input_edge): Stream common targets.
12756 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12757 * ipa.c: Include value-prof.h
12758 (ipa_profile_generate_summary): Turn indirect call statement histograms
12759 into common targets.
12760 (ipa_profile): Turn common targets into speculative edges.
12762 2013-08-09 Jan Hubicka <jh@suse.cz>
12764 * cgraph.h (cgraph_node): Add profile_id.
12765 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12766 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12767 (del_node_map): Update.
12768 (find_func_by_funcdef_no): Replace by ...
12769 (find_func_by_profile_id): ... this one.
12770 (gimple_ic_transform): Do not remove useful histograms when
12771 speculation is not done; dump info when indirect call removal
12773 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12774 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12775 (__gcov_indirect_call_profiler_v2): .. this one.
12776 * profile.h (init_node_map): Update.
12777 * coverage.c (coverage_compute_profile_id): New function.
12778 * coverage.h (coverage_compute_profile_id): Declare.
12779 * tree-profile.c (init_ic_make_global_vars): Make
12780 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12781 (gimple_init_edge_profiler): Update prototype of
12782 __gcov_indirect_call_profiler.
12783 (gimple_gen_ic_func_profiler): Simplify.
12784 (tree_profiling): Use init_node_map
12786 2013-08-09 Jan Hubicka <jh@suse.cz>
12788 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12789 non-speculative refs.
12790 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12791 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12792 (cgraph_set_call_stmt): Likewise.
12793 (cgraph_create_edge_1): Fix release checking compilatoin;
12794 clear lto_stmt_uid.
12795 (cgraph_free_edge): Free indirect info.
12796 (cgraph_turn_edge_to_speculative): New function.
12797 (cgraph_speculative_call_info): New function.
12798 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12799 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12800 (dump_cgraph_node): Dump speculation.
12801 (verify_edge_count_and_frequency): Accept speculative edges.
12802 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12803 (verify_cgraph_node): Handle speculation.
12804 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12805 (cgraph_set_call_stmt): Update prototype.
12806 (cgraph_make_edge_direct): Update prototype.
12807 (cgraph_speculative_call_info): Declare.
12808 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12809 to change; update call of ipa_find_references.
12810 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12811 lto_stmt_uid and speculative flags.
12812 (ipa_dump_references): Dump speculation.
12813 (ipa_clone_references): Clone speculative flag.
12814 (ipa_clone_referring): Likewise.
12815 (ipa_clone_ref): New function.
12816 (ipa_find_reference): Look into lto_stmt_uids
12817 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12818 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12819 (ipa_find_reference): Update declaration.
12820 (ipa_clone_ref): Declare.
12821 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12822 stream speculative flag.
12823 (lto_output_ref): Stream statements uids and speculation.
12824 (input_ref): Likewise.
12825 (input_edge): Stream speuclation.
12826 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12827 (cgraph_set_call_stmt_including_clones): Handle speculation.
12828 * ipa-inline.c (heap_edge_removal_hook): New function.
12829 (inline_small_functions): Register it.
12830 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12831 also initialize refs.
12832 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12834 (try_make_edge_direct_simple_call): Likewise.
12835 (try_make_edge_direct_simple_call): Likewise.
12836 (update_indirect_edges_after_inlining): Likewise.
12837 (remove_described_reference): Look proper lto_stmt_uid.
12838 (propagate_controlled_uses): Likewise.
12839 (propagate_controlled_uses): Liekwise.
12840 * tree-inline.c (copy_bb): Copy speculative edges.
12841 (redirect_all_calls): New function.
12842 (copy_cfg_body): Do redirection after loop info is updated.
12843 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12845 2013-08-09 Jan Hubicka <jh@suse.cz>
12847 * lto-streamer-out.c (output_function): Renumber PHIs.
12848 * lto-streamer-in.c (input_function): Likewise.
12850 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12852 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12853 (get_lane_unsigned): Likewise.
12854 (dup_lane_scalar): Likewise.
12855 (get_lane): enable for VALL.
12856 * config/aarch64/aarch64-simd.md
12857 (aarch64_dup_lane_scalar<mode>): Remove.
12858 (aarch64_get_lane_signed<mode>): Likewise.
12859 (aarch64_get_lane_unsigned<mode>): Likewise.
12860 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12861 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12862 (aarch64_get_lane<mode>): Enable for all vector modes.
12863 (aarch64_get_lanedi): Remove misleading constraints.
12864 * config/aarch64/arm_neon.h
12865 (__aarch64_vget_lane_any): Define.
12866 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12867 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12868 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12869 * config/aarch64/iterators.md (VDQQH): New.
12870 (VDQQHS): Likewise.
12871 (vwcore): Likewise.
12873 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12875 * configure.ac: Add GAS check for LEON instructions on SPARC.
12876 * configure: Regenerate.
12877 * config.in: Likewise.
12878 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12880 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12881 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12883 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12884 (AS_LEON_FLAG): New macro.
12885 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12886 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12887 Deal with LEON and LEON3 for the memory model.
12888 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12889 (atomic_compare_and_swap<mode>_1): Likewise.
12890 (*atomic_compare_and_swap<mode>_1): Likewise.
12892 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12894 * config/arm/neon.md (vcond): Fix floating-point vector
12895 comparisons against 0.
12897 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12899 * lra-constraints.c (emit_spill_move): Remove assert.
12900 (process_alt_operands): Add more debugging
12901 output. Increase reject for spilling into memory. Decrease
12902 reject for reloading scratch.
12903 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12905 2013-08-08 Steve Ellcey <sellcey@mips.com>
12907 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12908 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12909 micromips incompatible. Add nan2008.
12910 (MULTILIB_DIRNAMES): Add nan2008.
12911 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12912 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12913 and micromips incompatible. Add nan2008.
12914 (MULTILIB_DIRNAMES): Add nan2008.
12915 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12917 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12919 PR rtl-optimization/58079
12920 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12921 simplify_comparison has widened a comparison with an integer.
12923 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12925 * config/arm/neon.md (movmisalign<mode>): Disable when we
12926 don't allow unaligned accesses.
12927 (*movmisalign<mode>_neon_store): Likewise.
12928 (*movmisalign<mode>_neon_load): Likewise.
12929 (*movmisalign<mode>_neon_store): Likewise.
12930 (*movmisalign<mode>_neon_load): Likewise.
12932 2013-08-08 Jan Hubicka <jh@suse.cz>
12934 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12935 (make_pass_rebuild_cgraph_edges): Also clear references.
12936 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12937 * ipa-inline-transform.c (inline_transform): Remove all references
12939 * cgraphunit.c (expand_function): Remove all references after
12941 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12942 (ipa_find_reference): Rewrite to iterator.
12943 (remove_stmt_references): Likewise.
12944 (ipa_clear_stmts_in_references): New function.
12945 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12946 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12948 * ipa-split.c (split_function): Remove references in split function.
12950 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12953 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12954 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12956 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12959 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12960 suggested mode for the assignment isn't compatible with the
12961 registers required.
12963 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12966 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12968 2013-08-07 Xinliang David Li <davidxl@google.com>
12970 * config/i386/i386.opt: New option -mtune-ctrl=.
12971 * config/i386/x86-tune.def: New file.
12972 * config/i386/i386.h: include x86-tune.def.
12973 * config/i386/i386.c (ix86_option_override_internal):
12974 Parsing -mtune-ctrl= option and set tune features.
12976 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12979 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12980 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12983 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12985 * rtl.h (update_alignments): Declare.
12986 * final.c (grow_label_align): New function extracted from...
12987 (shorten_branches): ...here. Call it.
12988 (update_alignments): New function.
12989 * reorg.c (sibling_labels): New variable.
12990 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12991 the new label along with it onto the sibling_labels vector.
12992 (fill_simple_delay_slots): Adjust call to get_label_before.
12993 (fill_slots_from_thread): Likewise.
12994 (relax_delay_slots): Likewise.
12995 (make_return_insns): Likewise.
12996 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12998 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13000 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
13001 document its semantics.
13002 (diagnostic_report_diagnostic): Adjust accordingly.
13004 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13006 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
13007 (sparc_option_override): ...and port to new C++ pass API.
13008 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
13010 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
13012 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
13014 2013-08-06 Caroline Tice <cmtice@google.com>
13016 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
13017 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
13018 * tree-pass.h: Add pass_vtable_verify.
13019 * varasm.c (assemble_variable): Add code to properly set the comdat
13020 section and name for the .vtable_map_vars section.
13021 (assemble_vtyv_preinit_initializer): New function.
13022 (default_sectin_type_flags): Make sure .vtable_map_vars section has
13024 * output.h: Add function decl for assemble_vtv_preinit_initializer.
13025 * vtable-verify.c: New file.
13026 * vtable-verify.h: New file.
13027 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
13028 initialiation levels.
13029 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
13030 * passes.def: Insert pass_vtable_verify.
13031 * aclocal.m4: Reorder includes.
13032 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
13033 -fvtv-counts options.
13034 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
13035 as appropriate, if -fvtable-verify=... is used.
13036 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
13037 -fvtable-verify=... is used.
13038 * Makefile.in (OBJS): Add vtable-verify.o to list.
13039 (vtable-verify.o): Add new build rule.
13040 (GTFILES): Add vtable-verify.c to list.
13041 * common.opt (fvtable-verify=): New flag.
13042 (vtv_priority): Values for fvtable-verify= flag.
13043 (fvtv-counts): New flag.
13044 (fvtv-debug): New flag.
13045 * tree.h (save_vtable_map_decl): New extern function decl.
13047 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13049 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
13050 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
13051 (pass_data_rl78_devirt): ...new pass_data instance and...
13052 (make_pass_rl78_devirt): ...new function.
13053 (rl78_asm_file_start): Port pass registration to new C++ API.
13055 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13057 * coretypes.h (rtl_opt_pass): Add.
13058 (gcc::context): Add.
13059 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
13060 (epiphany_init): Port to new C++ pass API.
13061 (epiphany_optimize_mode_switching): Likewise.
13062 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
13063 (pass_manager::get_pass_mode_switching): New.
13064 (pass_manager::get_pass_peephole2): New.
13065 * mode-switching.c (pass_mode_switching): Add clone method.
13066 * recog.c (pass_peephole2): Add clone method.
13067 (pass_split_all_insns): Add clone method.
13069 2013-08-06 David Malcolm <dmalcolm@redhat.com>
13071 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
13072 (mips_option_override): ...here, porting to new C++ API for passes.
13074 2013-08-06 Jan Hubicka <jh@suse.cz>
13076 * cgraph.c (cgraph_get_body): New function based on lto.c
13078 * cgraph.h (cgraph_get_body): Declare.
13079 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
13081 * cgraphunit.c (expand_function): Get body prior expanding.
13082 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
13083 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
13085 * passes.c (do_per_function_toporder): Get body.
13086 * tree-inline.c (expand_call_inline): Get body prior inlining it.
13087 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
13089 2013-08-06 Martin Jambor <mjambor@suse.cz>
13092 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
13093 is not re-finalized. Rename second parameter to no_collect.
13095 2013-08-06 Martin Jambor <mjambor@suse.cz>
13097 PR middle-end/58041
13098 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
13099 MEM_REF has proper alignment information.
13101 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
13104 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
13106 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
13107 rtx (*) (rtx, ...) with insn_gen_fn.
13108 * genoutput.c (output_insn_data): Cast gen_? function pointers to
13109 insn_gen_fn::stored_funcptr. Add initializer braces.
13111 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13113 Rewrite how instances of passes are cloned to remove assumptions
13114 about their sizes (thus allowing pass subclasses to have
13115 additional data fields, albeit non-GC-managed ones at this point).
13117 * passes.c (make_pass_instance): Now that passes have clone
13118 methods, rewrite this function to eliminate XNEW and memcpy
13119 calls that used hardcoded sizes. Since this function no longer
13120 creates pass instances, rename it to...
13121 (add_pass_instance): ...this. Document the old way that passes were
13122 numbered and flagged, and rework this function to continue using it.
13123 (next_pass_1): Add an initial_pass argument for use by
13125 (position_pass): When adding multiple instances of a pass, use
13126 the pass's clone method, rather than relying on the XNEW/memcpy
13127 within the former make_pass_instance (now add_pass_instance).
13128 (pass_manager::pass_manager): When invoking next_pass_1, also supply
13129 the initial instance of the current pass within the pass manager.
13131 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13133 This is the automated part of the conversion of passes from C
13134 structs to C++ classes.
13136 Patch autogenerated by refactor_passes.py from
13137 https://github.com/davidmalcolm/gcc-refactoring-scripts
13138 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
13140 * asan.c (pass_asan): Convert from a global struct to a subclass of
13141 gimple_opt_pass along with...
13142 (pass_data_asan): ...new pass_data instance and...
13143 (make_pass_asan): ...new function.
13144 (pass_asan_O0): Convert from a global struct to a subclass of
13145 gimple_opt_pass along with...
13146 (pass_data_asan_O0): ...new pass_data instance and...
13147 (make_pass_asan_O0): ...new function.
13148 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
13149 subclass of rtl_opt_pass along with...
13150 (pass_data_inc_dec): ...new pass_data instance and...
13151 (make_pass_inc_dec): ...new function.
13152 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
13153 a subclass of rtl_opt_pass along with...
13154 (pass_data_reorder_blocks): ...new pass_data instance and...
13155 (make_pass_reorder_blocks): ...new function.
13156 (pass_duplicate_computed_gotos): Convert from a global struct to a
13157 subclass of rtl_opt_pass along with...
13158 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
13159 (make_pass_duplicate_computed_gotos): ...new function.
13160 (pass_partition_blocks): Convert from a global struct to a subclass of
13161 rtl_opt_pass along with...
13162 (pass_data_partition_blocks): ...new pass_data instance and...
13163 (make_pass_partition_blocks): ...new function.
13164 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
13165 struct to a subclass of rtl_opt_pass along with...
13166 (pass_data_branch_target_load_optimize1): ...new pass_data instance
13168 (make_pass_branch_target_load_optimize1): ...new function.
13169 (pass_branch_target_load_optimize2): Convert from a global struct to a
13170 subclass of rtl_opt_pass along with...
13171 (pass_data_branch_target_load_optimize2): ...new pass_data instance
13173 (make_pass_branch_target_load_optimize2): ...new function.
13174 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
13175 of rtl_opt_pass along with...
13176 (pass_data_jump): ...new pass_data instance and...
13177 (make_pass_jump): ...new function.
13178 (pass_jump2): Convert from a global struct to a subclass of
13179 rtl_opt_pass along with...
13180 (pass_data_jump2): ...new pass_data instance and...
13181 (make_pass_jump2): ...new function.
13182 * cfgexpand.c (pass_expand): Convert from a global struct to a
13183 subclass of rtl_opt_pass along with...
13184 (pass_data_expand): ...new pass_data instance and...
13185 (make_pass_expand): ...new function.
13186 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
13187 of rtl_opt_pass along with...
13188 (pass_data_free_cfg): ...new pass_data instance and...
13189 (make_pass_free_cfg): ...new function.
13190 (pass_into_cfg_layout_mode): Convert from a global struct to a
13191 subclass of rtl_opt_pass along with...
13192 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
13193 (make_pass_into_cfg_layout_mode): ...new function.
13194 (pass_outof_cfg_layout_mode): Convert from a global struct to a
13195 subclass of rtl_opt_pass along with...
13196 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
13197 (make_pass_outof_cfg_layout_mode): ...new function.
13198 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
13199 struct to a subclass of gimple_opt_pass along with...
13200 (pass_data_build_cgraph_edges): ...new pass_data instance and...
13201 (make_pass_build_cgraph_edges): ...new function.
13202 (pass_rebuild_cgraph_edges): Convert from a global struct to a
13203 subclass of gimple_opt_pass along with...
13204 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
13205 (make_pass_rebuild_cgraph_edges): ...new function.
13206 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
13207 subclass of gimple_opt_pass along with...
13208 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
13210 (make_pass_remove_cgraph_callee_edges): ...new function.
13211 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
13212 struct to a subclass of rtl_opt_pass along with...
13213 (pass_data_stack_adjustments): ...new pass_data instance and...
13214 (make_pass_stack_adjustments): ...new function.
13215 * combine.c (pass_combine): Convert from a global struct to a subclass
13216 of rtl_opt_pass along with...
13217 (pass_data_combine): ...new pass_data instance and...
13218 (make_pass_combine): ...new function.
13219 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
13220 global struct to a subclass of rtl_opt_pass along with...
13221 (pass_data_compare_elim_after_reload): ...new pass_data instance
13223 (make_pass_compare_elim_after_reload): ...new function.
13224 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
13225 of rtl_opt_pass along with...
13226 (pass_data_rtl_cprop): ...new pass_data instance and...
13227 (make_pass_rtl_cprop): ...new function.
13228 * cse.c (pass_cse): Convert from a global struct to a subclass of
13229 rtl_opt_pass along with...
13230 (pass_data_cse): ...new pass_data instance and...
13231 (make_pass_cse): ...new function.
13232 (pass_cse2): Convert from a global struct to a subclass of
13233 rtl_opt_pass along with...
13234 (pass_data_cse2): ...new pass_data instance and...
13235 (make_pass_cse2): ...new function.
13236 (pass_cse_after_global_opts): Convert from a global struct to a
13237 subclass of rtl_opt_pass along with...
13238 (pass_data_cse_after_global_opts): ...new pass_data instance and...
13239 (make_pass_cse_after_global_opts): ...new function.
13240 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
13241 of rtl_opt_pass along with...
13242 (pass_data_ud_rtl_dce): ...new pass_data instance and...
13243 (make_pass_ud_rtl_dce): ...new function.
13244 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
13245 rtl_opt_pass along with...
13246 (pass_data_fast_rtl_dce): ...new pass_data instance and...
13247 (make_pass_fast_rtl_dce): ...new function.
13248 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
13249 a subclass of rtl_opt_pass along with...
13250 (pass_data_df_initialize_opt): ...new pass_data instance and...
13251 (make_pass_df_initialize_opt): ...new function.
13252 (pass_df_initialize_no_opt): Convert from a global struct to a
13253 subclass of rtl_opt_pass along with...
13254 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
13255 (make_pass_df_initialize_no_opt): ...new function.
13256 (pass_df_finish): Convert from a global struct to a subclass of
13257 rtl_opt_pass along with...
13258 (pass_data_df_finish): ...new pass_data instance and...
13259 (make_pass_df_finish): ...new function.
13260 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
13261 rtl_opt_pass along with...
13262 (pass_data_rtl_dse1): ...new pass_data instance and...
13263 (make_pass_rtl_dse1): ...new function.
13264 (pass_rtl_dse2): Convert from a global struct to a subclass of
13265 rtl_opt_pass along with...
13266 (pass_data_rtl_dse2): ...new pass_data instance and...
13267 (make_pass_rtl_dse2): ...new function.
13268 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
13269 subclass of rtl_opt_pass along with...
13270 (pass_data_dwarf2_frame): ...new pass_data instance and...
13271 (make_pass_dwarf2_frame): ...new function.
13272 * except.c (pass_set_nothrow_function_flags): Convert from a global
13273 struct to a subclass of rtl_opt_pass along with...
13274 (pass_data_set_nothrow_function_flags): ...new pass_data instance
13276 (make_pass_set_nothrow_function_flags): ...new function.
13277 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
13278 subclass of rtl_opt_pass along with...
13279 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
13281 (make_pass_convert_to_eh_region_ranges): ...new function.
13282 * final.c (pass_compute_alignments): Convert from a global struct to a
13283 subclass of rtl_opt_pass along with...
13284 (pass_data_compute_alignments): ...new pass_data instance and...
13285 (make_pass_compute_alignments): ...new function.
13286 (pass_final): Convert from a global struct to a subclass of
13287 rtl_opt_pass along with...
13288 (pass_data_final): ...new pass_data instance and...
13289 (make_pass_final): ...new function.
13290 (pass_shorten_branches): Convert from a global struct to a subclass of
13291 rtl_opt_pass along with...
13292 (pass_data_shorten_branches): ...new pass_data instance and...
13293 (make_pass_shorten_branches): ...new function.
13294 (pass_clean_state): Convert from a global struct to a subclass of
13295 rtl_opt_pass along with...
13296 (pass_data_clean_state): ...new pass_data instance and...
13297 (make_pass_clean_state): ...new function.
13298 * function.c (pass_instantiate_virtual_regs): Convert from a global
13299 struct to a subclass of rtl_opt_pass along with...
13300 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13301 (make_pass_instantiate_virtual_regs): ...new function.
13302 (pass_leaf_regs): Convert from a global struct to a subclass of
13303 rtl_opt_pass along with...
13304 (pass_data_leaf_regs): ...new pass_data instance and...
13305 (make_pass_leaf_regs): ...new function.
13306 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13307 subclass of rtl_opt_pass along with...
13308 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13310 (make_pass_thread_prologue_and_epilogue): ...new function.
13311 (pass_match_asm_constraints): Convert from a global struct to a
13312 subclass of rtl_opt_pass along with...
13313 (pass_data_match_asm_constraints): ...new pass_data instance and...
13314 (make_pass_match_asm_constraints): ...new function.
13315 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13316 subclass of rtl_opt_pass along with...
13317 (pass_data_rtl_fwprop): ...new pass_data instance and...
13318 (make_pass_rtl_fwprop): ...new function.
13319 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13320 rtl_opt_pass along with...
13321 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13322 (make_pass_rtl_fwprop_addr): ...new function.
13323 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13324 rtl_opt_pass along with...
13325 (pass_data_rtl_pre): ...new pass_data instance and...
13326 (make_pass_rtl_pre): ...new function.
13327 (pass_rtl_hoist): Convert from a global struct to a subclass of
13328 rtl_opt_pass along with...
13329 (pass_data_rtl_hoist): ...new pass_data instance and...
13330 (make_pass_rtl_hoist): ...new function.
13331 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13332 subclass of gimple_opt_pass along with...
13333 (pass_data_lower_cf): ...new pass_data instance and...
13334 (make_pass_lower_cf): ...new function.
13335 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13336 from a global struct to a subclass of gimple_opt_pass along with...
13337 (pass_data_strength_reduction): ...new pass_data instance and...
13338 (make_pass_strength_reduction): ...new function.
13339 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13340 of rtl_opt_pass along with...
13341 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13342 (make_pass_rtl_ifcvt): ...new function.
13343 (pass_if_after_combine): Convert from a global struct to a subclass of
13344 rtl_opt_pass along with...
13345 (pass_data_if_after_combine): ...new pass_data instance and...
13346 (make_pass_if_after_combine): ...new function.
13347 (pass_if_after_reload): Convert from a global struct to a subclass of
13348 rtl_opt_pass along with...
13349 (pass_data_if_after_reload): ...new pass_data instance and...
13350 (make_pass_if_after_reload): ...new function.
13351 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13352 a subclass of rtl_opt_pass along with...
13353 (pass_data_initialize_regs): ...new pass_data instance and...
13354 (make_pass_initialize_regs): ...new function.
13355 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13356 of ipa_opt_pass_d along with...
13357 (pass_data_ipa_cp): ...new pass_data instance and...
13358 (make_pass_ipa_cp): ...new function.
13359 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13360 global struct to a subclass of gimple_opt_pass along with...
13361 (pass_data_inline_parameters): ...new pass_data instance and...
13362 (make_pass_inline_parameters): ...new function.
13363 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13364 subclass of gimple_opt_pass along with...
13365 (pass_data_early_inline): ...new pass_data instance and...
13366 (make_pass_early_inline): ...new function.
13367 (pass_ipa_inline): Convert from a global struct to a subclass of
13368 ipa_opt_pass_d along with...
13369 (pass_data_ipa_inline): ...new pass_data instance and...
13370 (make_pass_ipa_inline): ...new function.
13371 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13372 struct to a subclass of gimple_opt_pass along with...
13373 (pass_data_local_pure_const): ...new pass_data instance and...
13374 (make_pass_local_pure_const): ...new function.
13375 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13376 ipa_opt_pass_d along with...
13377 (pass_data_ipa_pure_const): ...new pass_data instance and...
13378 (make_pass_ipa_pure_const): ...new function.
13379 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13380 to a subclass of ipa_opt_pass_d along with...
13381 (pass_data_ipa_reference): ...new pass_data instance and...
13382 (make_pass_ipa_reference): ...new function.
13383 * ipa-split.c (pass_split_functions): Convert from a global struct to
13384 a subclass of gimple_opt_pass along with...
13385 (pass_data_split_functions): ...new pass_data instance and...
13386 (make_pass_split_functions): ...new function.
13387 (pass_feedback_split_functions): Convert from a global struct to a
13388 subclass of gimple_opt_pass along with...
13389 (pass_data_feedback_split_functions): ...new pass_data instance and...
13390 (make_pass_feedback_split_functions): ...new function.
13391 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13392 global struct to a subclass of simple_ipa_opt_pass along with...
13393 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13395 (make_pass_ipa_function_and_variable_visibility): ...new function.
13396 (pass_ipa_free_inline_summary): Convert from a global struct to a
13397 subclass of simple_ipa_opt_pass along with...
13398 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13399 (make_pass_ipa_free_inline_summary): ...new function.
13400 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13401 subclass of ipa_opt_pass_d along with...
13402 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13404 (make_pass_ipa_whole_program_visibility): ...new function.
13405 (pass_ipa_profile): Convert from a global struct to a subclass of
13406 ipa_opt_pass_d along with...
13407 (pass_data_ipa_profile): ...new pass_data instance and...
13408 (make_pass_ipa_profile): ...new function.
13409 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13410 ipa_opt_pass_d along with...
13411 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13412 (make_pass_ipa_cdtor_merge): ...new function.
13413 * ira.c (pass_ira): Convert from a global struct to a subclass of
13414 rtl_opt_pass along with...
13415 (pass_data_ira): ...new pass_data instance and...
13416 (make_pass_ira): ...new function.
13417 (pass_reload): Convert from a global struct to a subclass of
13418 rtl_opt_pass along with...
13419 (pass_data_reload): ...new pass_data instance and...
13420 (make_pass_reload): ...new function.
13421 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13422 subclass of rtl_opt_pass along with...
13423 (pass_data_cleanup_barriers): ...new pass_data instance and...
13424 (make_pass_cleanup_barriers): ...new function.
13425 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13426 of rtl_opt_pass along with...
13427 (pass_data_loop2): ...new pass_data instance and...
13428 (make_pass_loop2): ...new function.
13429 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13430 rtl_opt_pass along with...
13431 (pass_data_rtl_loop_init): ...new pass_data instance and...
13432 (make_pass_rtl_loop_init): ...new function.
13433 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13434 rtl_opt_pass along with...
13435 (pass_data_rtl_loop_done): ...new pass_data instance and...
13436 (make_pass_rtl_loop_done): ...new function.
13437 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13438 subclass of rtl_opt_pass along with...
13439 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13440 (make_pass_rtl_move_loop_invariants): ...new function.
13441 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13442 rtl_opt_pass along with...
13443 (pass_data_rtl_unswitch): ...new pass_data instance and...
13444 (make_pass_rtl_unswitch): ...new function.
13445 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13446 subclass of rtl_opt_pass along with...
13447 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13449 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13450 (pass_rtl_doloop): Convert from a global struct to a subclass of
13451 rtl_opt_pass along with...
13452 (pass_data_rtl_doloop): ...new pass_data instance and...
13453 (make_pass_rtl_doloop): ...new function.
13454 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13455 a subclass of rtl_opt_pass along with...
13456 (pass_data_lower_subreg): ...new pass_data instance and...
13457 (make_pass_lower_subreg): ...new function.
13458 (pass_lower_subreg2): Convert from a global struct to a subclass of
13459 rtl_opt_pass along with...
13460 (pass_data_lower_subreg2): ...new pass_data instance and...
13461 (make_pass_lower_subreg2): ...new function.
13462 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13463 struct to a subclass of ipa_opt_pass_d along with...
13464 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13465 (make_pass_ipa_lto_gimple_out): ...new function.
13466 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13467 of ipa_opt_pass_d along with...
13468 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13469 (make_pass_ipa_lto_finish_out): ...new function.
13470 * mode-switching.c (pass_mode_switching): Convert from a global struct
13471 to a subclass of rtl_opt_pass along with...
13472 (pass_data_mode_switching): ...new pass_data instance and...
13473 (make_pass_mode_switching): ...new function.
13474 * modulo-sched.c (pass_sms): Convert from a global struct to a
13475 subclass of rtl_opt_pass along with...
13476 (pass_data_sms): ...new pass_data instance and...
13477 (make_pass_sms): ...new function.
13478 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13479 subclass of gimple_opt_pass along with...
13480 (pass_data_expand_omp): ...new pass_data instance and...
13481 (make_pass_expand_omp): ...new function.
13482 (pass_lower_omp): Convert from a global struct to a subclass of
13483 gimple_opt_pass along with...
13484 (pass_data_lower_omp): ...new pass_data instance and...
13485 (make_pass_lower_omp): ...new function.
13486 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13487 of gimple_opt_pass along with...
13488 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13489 (make_pass_diagnose_omp_blocks): ...new function.
13490 * passes.c (pass_early_local_passes): Convert from a global struct to
13491 a subclass of simple_ipa_opt_pass along with...
13492 (pass_data_early_local_passes): ...new pass_data instance and...
13493 (make_pass_early_local_passes): ...new function.
13494 (pass_all_early_optimizations): Convert from a global struct to a
13495 subclass of gimple_opt_pass along with...
13496 (pass_data_all_early_optimizations): ...new pass_data instance and...
13497 (make_pass_all_early_optimizations): ...new function.
13498 (pass_all_optimizations): Convert from a global struct to a subclass
13499 of gimple_opt_pass along with...
13500 (pass_data_all_optimizations): ...new pass_data instance and...
13501 (make_pass_all_optimizations): ...new function.
13502 (pass_all_optimizations_g): Convert from a global struct to a subclass
13503 of gimple_opt_pass along with...
13504 (pass_data_all_optimizations_g): ...new pass_data instance and...
13505 (make_pass_all_optimizations_g): ...new function.
13506 (pass_rest_of_compilation): Convert from a global struct to a subclass
13507 of rtl_opt_pass along with...
13508 (pass_data_rest_of_compilation): ...new pass_data instance and...
13509 (make_pass_rest_of_compilation): ...new function.
13510 (pass_postreload): Convert from a global struct to a subclass of
13511 rtl_opt_pass along with...
13512 (pass_data_postreload): ...new pass_data instance and...
13513 (make_pass_postreload): ...new function.
13514 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13515 subclass of rtl_opt_pass along with...
13516 (pass_data_gcse2): ...new pass_data instance and...
13517 (make_pass_gcse2): ...new function.
13518 * postreload.c (pass_postreload_cse): Convert from a global struct to
13519 a subclass of rtl_opt_pass along with...
13520 (pass_data_postreload_cse): ...new pass_data instance and...
13521 (make_pass_postreload_cse): ...new function.
13522 * predict.c (pass_profile): Convert from a global struct to a subclass
13523 of gimple_opt_pass along with...
13524 (pass_data_profile): ...new pass_data instance and...
13525 (make_pass_profile): ...new function.
13526 (pass_strip_predict_hints): Convert from a global struct to a subclass
13527 of gimple_opt_pass along with...
13528 (pass_data_strip_predict_hints): ...new pass_data instance and...
13529 (make_pass_strip_predict_hints): ...new function.
13530 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13531 of rtl_opt_pass along with...
13532 (pass_data_peephole2): ...new pass_data instance and...
13533 (make_pass_peephole2): ...new function.
13534 (pass_split_all_insns): Convert from a global struct to a subclass of
13535 rtl_opt_pass along with...
13536 (pass_data_split_all_insns): ...new pass_data instance and...
13537 (make_pass_split_all_insns): ...new function.
13538 (pass_split_after_reload): Convert from a global struct to a subclass
13539 of rtl_opt_pass along with...
13540 (pass_data_split_after_reload): ...new pass_data instance and...
13541 (make_pass_split_after_reload): ...new function.
13542 (pass_split_before_regstack): Convert from a global struct to a
13543 subclass of rtl_opt_pass along with...
13544 (pass_data_split_before_regstack): ...new pass_data instance and...
13545 (make_pass_split_before_regstack): ...new function.
13546 (pass_split_before_sched2): Convert from a global struct to a subclass
13547 of rtl_opt_pass along with...
13548 (pass_data_split_before_sched2): ...new pass_data instance and...
13549 (make_pass_split_before_sched2): ...new function.
13550 (pass_split_for_shorten_branches): Convert from a global struct to a
13551 subclass of rtl_opt_pass along with...
13552 (pass_data_split_for_shorten_branches): ...new pass_data instance
13554 (make_pass_split_for_shorten_branches): ...new function.
13555 * ree.c (pass_ree): Convert from a global struct to a subclass of
13556 rtl_opt_pass along with...
13557 (pass_data_ree): ...new pass_data instance and...
13558 (make_pass_ree): ...new function.
13559 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13560 subclass of rtl_opt_pass along with...
13561 (pass_data_stack_regs): ...new pass_data instance and...
13562 (make_pass_stack_regs): ...new function.
13563 (pass_stack_regs_run): Convert from a global struct to a subclass of
13564 rtl_opt_pass along with...
13565 (pass_data_stack_regs_run): ...new pass_data instance and...
13566 (make_pass_stack_regs_run): ...new function.
13567 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13568 subclass of rtl_opt_pass along with...
13569 (pass_data_cprop_hardreg): ...new pass_data instance and...
13570 (make_pass_cprop_hardreg): ...new function.
13571 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13572 subclass of rtl_opt_pass along with...
13573 (pass_data_reginfo_init): ...new pass_data instance and...
13574 (make_pass_reginfo_init): ...new function.
13575 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13576 of rtl_opt_pass along with...
13577 (pass_data_regmove): ...new pass_data instance and...
13578 (make_pass_regmove): ...new function.
13579 * regrename.c (pass_regrename): Convert from a global struct to a
13580 subclass of rtl_opt_pass along with...
13581 (pass_data_regrename): ...new pass_data instance and...
13582 (make_pass_regrename): ...new function.
13583 * reorg.c (pass_delay_slots): Convert from a global struct to a
13584 subclass of rtl_opt_pass along with...
13585 (pass_data_delay_slots): ...new pass_data instance and...
13586 (make_pass_delay_slots): ...new function.
13587 (pass_machine_reorg): Convert from a global struct to a subclass of
13588 rtl_opt_pass along with...
13589 (pass_data_machine_reorg): ...new pass_data instance and...
13590 (make_pass_machine_reorg): ...new function.
13591 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13592 of rtl_opt_pass along with...
13593 (pass_data_sched): ...new pass_data instance and...
13594 (make_pass_sched): ...new function.
13595 (pass_sched2): Convert from a global struct to a subclass of
13596 rtl_opt_pass along with...
13597 (pass_data_sched2): ...new pass_data instance and...
13598 (make_pass_sched2): ...new function.
13599 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13600 to a subclass of rtl_opt_pass along with...
13601 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13602 (make_pass_stack_ptr_mod): ...new function.
13603 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13604 to a subclass of rtl_opt_pass along with...
13605 (pass_data_rtl_store_motion): ...new pass_data instance and...
13606 (make_pass_rtl_store_motion): ...new function.
13607 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13608 of gimple_opt_pass along with...
13609 (pass_data_tracer): ...new pass_data instance and...
13610 (make_pass_tracer): ...new function.
13611 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13612 to a subclass of gimple_opt_pass along with...
13613 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13614 (make_pass_diagnose_tm_blocks): ...new function.
13615 (pass_lower_tm): Convert from a global struct to a subclass of
13616 gimple_opt_pass along with...
13617 (pass_data_lower_tm): ...new pass_data instance and...
13618 (make_pass_lower_tm): ...new function.
13619 (pass_tm_init): Convert from a global struct to a subclass of
13620 gimple_opt_pass along with...
13621 (pass_data_tm_init): ...new pass_data instance and...
13622 (make_pass_tm_init): ...new function.
13623 (pass_tm_mark): Convert from a global struct to a subclass of
13624 gimple_opt_pass along with...
13625 (pass_data_tm_mark): ...new pass_data instance and...
13626 (make_pass_tm_mark): ...new function.
13627 (pass_tm_edges): Convert from a global struct to a subclass of
13628 gimple_opt_pass along with...
13629 (pass_data_tm_edges): ...new pass_data instance and...
13630 (make_pass_tm_edges): ...new function.
13631 (pass_tm_memopt): Convert from a global struct to a subclass of
13632 gimple_opt_pass along with...
13633 (pass_data_tm_memopt): ...new pass_data instance and...
13634 (make_pass_tm_memopt): ...new function.
13635 (pass_ipa_tm): Convert from a global struct to a subclass of
13636 simple_ipa_opt_pass along with...
13637 (pass_data_ipa_tm): ...new pass_data instance and...
13638 (make_pass_ipa_tm): ...new function.
13639 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13640 subclass of gimple_opt_pass along with...
13641 (pass_data_call_cdce): ...new pass_data instance and...
13642 (make_pass_call_cdce): ...new function.
13643 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13644 subclass of gimple_opt_pass along with...
13645 (pass_data_build_cfg): ...new pass_data instance and...
13646 (make_pass_build_cfg): ...new function.
13647 (pass_split_crit_edges): Convert from a global struct to a subclass of
13648 gimple_opt_pass along with...
13649 (pass_data_split_crit_edges): ...new pass_data instance and...
13650 (make_pass_split_crit_edges): ...new function.
13651 (pass_warn_function_return): Convert from a global struct to a
13652 subclass of gimple_opt_pass along with...
13653 (pass_data_warn_function_return): ...new pass_data instance and...
13654 (make_pass_warn_function_return): ...new function.
13655 (pass_warn_function_noreturn): Convert from a global struct to a
13656 subclass of gimple_opt_pass along with...
13657 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13658 (make_pass_warn_function_noreturn): ...new function.
13659 (pass_warn_unused_result): Convert from a global struct to a subclass
13660 of gimple_opt_pass along with...
13661 (pass_data_warn_unused_result): ...new pass_data instance and...
13662 (make_pass_warn_unused_result): ...new function.
13663 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13664 a subclass of gimple_opt_pass along with...
13665 (pass_data_merge_phi): ...new pass_data instance and...
13666 (make_pass_merge_phi): ...new function.
13667 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13668 a subclass of gimple_opt_pass along with...
13669 (pass_data_lower_complex): ...new pass_data instance and...
13670 (make_pass_lower_complex): ...new function.
13671 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13672 gimple_opt_pass along with...
13673 (pass_data_lower_complex_O0): ...new pass_data instance and...
13674 (make_pass_lower_complex_O0): ...new function.
13675 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13676 subclass of gimple_opt_pass along with...
13677 (pass_data_lower_eh): ...new pass_data instance and...
13678 (make_pass_lower_eh): ...new function.
13679 (pass_refactor_eh): Convert from a global struct to a subclass of
13680 gimple_opt_pass along with...
13681 (pass_data_refactor_eh): ...new pass_data instance and...
13682 (make_pass_refactor_eh): ...new function.
13683 (pass_lower_resx): Convert from a global struct to a subclass of
13684 gimple_opt_pass along with...
13685 (pass_data_lower_resx): ...new pass_data instance and...
13686 (make_pass_lower_resx): ...new function.
13687 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13688 of gimple_opt_pass along with...
13689 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13690 (make_pass_lower_eh_dispatch): ...new function.
13691 (pass_cleanup_eh): Convert from a global struct to a subclass of
13692 gimple_opt_pass along with...
13693 (pass_data_cleanup_eh): ...new pass_data instance and...
13694 (make_pass_cleanup_eh): ...new function.
13695 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13696 to a subclass of simple_ipa_opt_pass along with...
13697 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13698 (make_pass_ipa_lower_emutls): ...new function.
13699 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13700 a subclass of gimple_opt_pass along with...
13701 (pass_data_if_conversion): ...new pass_data instance and...
13702 (make_pass_if_conversion): ...new function.
13703 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13704 subclass of gimple_opt_pass along with...
13705 (pass_data_build_ssa): ...new pass_data instance and...
13706 (make_pass_build_ssa): ...new function.
13707 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13708 global struct to a subclass of gimple_opt_pass along with...
13709 (pass_data_loop_distribution): ...new pass_data instance and...
13710 (make_pass_loop_distribution): ...new function.
13711 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13712 subclass of gimple_opt_pass along with...
13713 (pass_data_mudflap_1): ...new pass_data instance and...
13714 (make_pass_mudflap_1): ...new function.
13715 (pass_mudflap_2): Convert from a global struct to a subclass of
13716 gimple_opt_pass along with...
13717 (pass_data_mudflap_2): ...new pass_data instance and...
13718 (make_pass_mudflap_2): ...new function.
13719 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13720 subclass of gimple_opt_pass along with...
13721 (pass_data_mudflap_1): ...new pass_data instance and...
13722 (make_pass_mudflap_1): ...new function.
13723 (pass_mudflap_2): Convert from a global struct to a subclass of
13724 gimple_opt_pass along with...
13725 (pass_data_mudflap_2): ...new pass_data instance and...
13726 (make_pass_mudflap_2): ...new function.
13727 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13728 gimple_opt_pass along with...
13729 (pass_data_nrv): ...new pass_data instance and...
13730 (make_pass_nrv): ...new function.
13731 (pass_return_slot): Convert from a global struct to a subclass of
13732 gimple_opt_pass along with...
13733 (pass_data_return_slot): ...new pass_data instance and...
13734 (make_pass_return_slot): ...new function.
13735 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13736 to a subclass of gimple_opt_pass along with...
13737 (pass_data_object_sizes): ...new pass_data instance and...
13738 (make_pass_object_sizes): ...new function.
13739 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13740 global struct to a subclass of gimple_opt_pass along with...
13741 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13743 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13744 (pass_fixup_cfg): Convert from a global struct to a subclass of
13745 gimple_opt_pass along with...
13746 (pass_data_fixup_cfg): ...new pass_data instance and...
13747 (make_pass_fixup_cfg): ...new function.
13748 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13749 (make_pass_mudflap_1): ...new function.
13750 (pass_mudflap_2): Replace declaration with that of...
13751 (make_pass_mudflap_2): ...new function.
13752 (pass_asan): Replace declaration with that of...
13753 (make_pass_asan): ...new function.
13754 (pass_asan_O0): Replace declaration with that of...
13755 (make_pass_asan_O0): ...new function.
13756 (pass_tsan): Replace declaration with that of...
13757 (make_pass_tsan): ...new function.
13758 (pass_tsan_O0): Replace declaration with that of...
13759 (make_pass_tsan_O0): ...new function.
13760 (pass_lower_cf): Replace declaration with that of...
13761 (make_pass_lower_cf): ...new function.
13762 (pass_refactor_eh): Replace declaration with that of...
13763 (make_pass_refactor_eh): ...new function.
13764 (pass_lower_eh): Replace declaration with that of...
13765 (make_pass_lower_eh): ...new function.
13766 (pass_lower_eh_dispatch): Replace declaration with that of...
13767 (make_pass_lower_eh_dispatch): ...new function.
13768 (pass_lower_resx): Replace declaration with that of...
13769 (make_pass_lower_resx): ...new function.
13770 (pass_build_cfg): Replace declaration with that of...
13771 (make_pass_build_cfg): ...new function.
13772 (pass_early_tree_profile): Replace declaration with that of...
13773 (make_pass_early_tree_profile): ...new function.
13774 (pass_cleanup_eh): Replace declaration with that of...
13775 (make_pass_cleanup_eh): ...new function.
13776 (pass_sra): Replace declaration with that of...
13777 (make_pass_sra): ...new function.
13778 (pass_sra_early): Replace declaration with that of...
13779 (make_pass_sra_early): ...new function.
13780 (pass_early_ipa_sra): Replace declaration with that of...
13781 (make_pass_early_ipa_sra): ...new function.
13782 (pass_tail_recursion): Replace declaration with that of...
13783 (make_pass_tail_recursion): ...new function.
13784 (pass_tail_calls): Replace declaration with that of...
13785 (make_pass_tail_calls): ...new function.
13786 (pass_tree_loop): Replace declaration with that of...
13787 (make_pass_tree_loop): ...new function.
13788 (pass_tree_loop_init): Replace declaration with that of...
13789 (make_pass_tree_loop_init): ...new function.
13790 (pass_lim): Replace declaration with that of...
13791 (make_pass_lim): ...new function.
13792 (pass_tree_unswitch): Replace declaration with that of...
13793 (make_pass_tree_unswitch): ...new function.
13794 (pass_predcom): Replace declaration with that of...
13795 (make_pass_predcom): ...new function.
13796 (pass_iv_canon): Replace declaration with that of...
13797 (make_pass_iv_canon): ...new function.
13798 (pass_scev_cprop): Replace declaration with that of...
13799 (make_pass_scev_cprop): ...new function.
13800 (pass_empty_loop): Replace declaration with that of...
13801 (make_pass_empty_loop): ...new function.
13802 (pass_record_bounds): Replace declaration with that of...
13803 (make_pass_record_bounds): ...new function.
13804 (pass_graphite): Replace declaration with that of...
13805 (make_pass_graphite): ...new function.
13806 (pass_graphite_transforms): Replace declaration with that of...
13807 (make_pass_graphite_transforms): ...new function.
13808 (pass_if_conversion): Replace declaration with that of...
13809 (make_pass_if_conversion): ...new function.
13810 (pass_loop_distribution): Replace declaration with that of...
13811 (make_pass_loop_distribution): ...new function.
13812 (pass_vectorize): Replace declaration with that of...
13813 (make_pass_vectorize): ...new function.
13814 (pass_slp_vectorize): Replace declaration with that of...
13815 (make_pass_slp_vectorize): ...new function.
13816 (pass_complete_unroll): Replace declaration with that of...
13817 (make_pass_complete_unroll): ...new function.
13818 (pass_complete_unrolli): Replace declaration with that of...
13819 (make_pass_complete_unrolli): ...new function.
13820 (pass_parallelize_loops): Replace declaration with that of...
13821 (make_pass_parallelize_loops): ...new function.
13822 (pass_loop_prefetch): Replace declaration with that of...
13823 (make_pass_loop_prefetch): ...new function.
13824 (pass_iv_optimize): Replace declaration with that of...
13825 (make_pass_iv_optimize): ...new function.
13826 (pass_tree_loop_done): Replace declaration with that of...
13827 (make_pass_tree_loop_done): ...new function.
13828 (pass_ch): Replace declaration with that of...
13829 (make_pass_ch): ...new function.
13830 (pass_ccp): Replace declaration with that of...
13831 (make_pass_ccp): ...new function.
13832 (pass_phi_only_cprop): Replace declaration with that of...
13833 (make_pass_phi_only_cprop): ...new function.
13834 (pass_build_ssa): Replace declaration with that of...
13835 (make_pass_build_ssa): ...new function.
13836 (pass_build_alias): Replace declaration with that of...
13837 (make_pass_build_alias): ...new function.
13838 (pass_build_ealias): Replace declaration with that of...
13839 (make_pass_build_ealias): ...new function.
13840 (pass_dominator): Replace declaration with that of...
13841 (make_pass_dominator): ...new function.
13842 (pass_dce): Replace declaration with that of...
13843 (make_pass_dce): ...new function.
13844 (pass_dce_loop): Replace declaration with that of...
13845 (make_pass_dce_loop): ...new function.
13846 (pass_cd_dce): Replace declaration with that of...
13847 (make_pass_cd_dce): ...new function.
13848 (pass_call_cdce): Replace declaration with that of...
13849 (make_pass_call_cdce): ...new function.
13850 (pass_merge_phi): Replace declaration with that of...
13851 (make_pass_merge_phi): ...new function.
13852 (pass_split_crit_edges): Replace declaration with that of...
13853 (make_pass_split_crit_edges): ...new function.
13854 (pass_pre): Replace declaration with that of...
13855 (make_pass_pre): ...new function.
13856 (pass_profile): Replace declaration with that of...
13857 (make_pass_profile): ...new function.
13858 (pass_strip_predict_hints): Replace declaration with that of...
13859 (make_pass_strip_predict_hints): ...new function.
13860 (pass_lower_complex_O0): Replace declaration with that of...
13861 (make_pass_lower_complex_O0): ...new function.
13862 (pass_lower_complex): Replace declaration with that of...
13863 (make_pass_lower_complex): ...new function.
13864 (pass_lower_vector): Replace declaration with that of...
13865 (make_pass_lower_vector): ...new function.
13866 (pass_lower_vector_ssa): Replace declaration with that of...
13867 (make_pass_lower_vector_ssa): ...new function.
13868 (pass_lower_omp): Replace declaration with that of...
13869 (make_pass_lower_omp): ...new function.
13870 (pass_diagnose_omp_blocks): Replace declaration with that of...
13871 (make_pass_diagnose_omp_blocks): ...new function.
13872 (pass_expand_omp): Replace declaration with that of...
13873 (make_pass_expand_omp): ...new function.
13874 (pass_expand_omp_ssa): Replace declaration with that of...
13875 (make_pass_expand_omp_ssa): ...new function.
13876 (pass_object_sizes): Replace declaration with that of...
13877 (make_pass_object_sizes): ...new function.
13878 (pass_strlen): Replace declaration with that of...
13879 (make_pass_strlen): ...new function.
13880 (pass_fold_builtins): Replace declaration with that of...
13881 (make_pass_fold_builtins): ...new function.
13882 (pass_stdarg): Replace declaration with that of...
13883 (make_pass_stdarg): ...new function.
13884 (pass_early_warn_uninitialized): Replace declaration with that of...
13885 (make_pass_early_warn_uninitialized): ...new function.
13886 (pass_late_warn_uninitialized): Replace declaration with that of...
13887 (make_pass_late_warn_uninitialized): ...new function.
13888 (pass_cse_reciprocals): Replace declaration with that of...
13889 (make_pass_cse_reciprocals): ...new function.
13890 (pass_cse_sincos): Replace declaration with that of...
13891 (make_pass_cse_sincos): ...new function.
13892 (pass_optimize_bswap): Replace declaration with that of...
13893 (make_pass_optimize_bswap): ...new function.
13894 (pass_optimize_widening_mul): Replace declaration with that of...
13895 (make_pass_optimize_widening_mul): ...new function.
13896 (pass_warn_function_return): Replace declaration with that of...
13897 (make_pass_warn_function_return): ...new function.
13898 (pass_warn_function_noreturn): Replace declaration with that of...
13899 (make_pass_warn_function_noreturn): ...new function.
13900 (pass_cselim): Replace declaration with that of...
13901 (make_pass_cselim): ...new function.
13902 (pass_phiopt): Replace declaration with that of...
13903 (make_pass_phiopt): ...new function.
13904 (pass_forwprop): Replace declaration with that of...
13905 (make_pass_forwprop): ...new function.
13906 (pass_phiprop): Replace declaration with that of...
13907 (make_pass_phiprop): ...new function.
13908 (pass_tree_ifcombine): Replace declaration with that of...
13909 (make_pass_tree_ifcombine): ...new function.
13910 (pass_dse): Replace declaration with that of...
13911 (make_pass_dse): ...new function.
13912 (pass_nrv): Replace declaration with that of...
13913 (make_pass_nrv): ...new function.
13914 (pass_rename_ssa_copies): Replace declaration with that of...
13915 (make_pass_rename_ssa_copies): ...new function.
13916 (pass_sink_code): Replace declaration with that of...
13917 (make_pass_sink_code): ...new function.
13918 (pass_fre): Replace declaration with that of...
13919 (make_pass_fre): ...new function.
13920 (pass_check_data_deps): Replace declaration with that of...
13921 (make_pass_check_data_deps): ...new function.
13922 (pass_copy_prop): Replace declaration with that of...
13923 (make_pass_copy_prop): ...new function.
13924 (pass_vrp): Replace declaration with that of...
13925 (make_pass_vrp): ...new function.
13926 (pass_uncprop): Replace declaration with that of...
13927 (make_pass_uncprop): ...new function.
13928 (pass_return_slot): Replace declaration with that of...
13929 (make_pass_return_slot): ...new function.
13930 (pass_reassoc): Replace declaration with that of...
13931 (make_pass_reassoc): ...new function.
13932 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13933 (make_pass_rebuild_cgraph_edges): ...new function.
13934 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13935 (make_pass_remove_cgraph_callee_edges): ...new function.
13936 (pass_build_cgraph_edges): Replace declaration with that of...
13937 (make_pass_build_cgraph_edges): ...new function.
13938 (pass_local_pure_const): Replace declaration with that of...
13939 (make_pass_local_pure_const): ...new function.
13940 (pass_tracer): Replace declaration with that of...
13941 (make_pass_tracer): ...new function.
13942 (pass_warn_unused_result): Replace declaration with that of...
13943 (make_pass_warn_unused_result): ...new function.
13944 (pass_diagnose_tm_blocks): Replace declaration with that of...
13945 (make_pass_diagnose_tm_blocks): ...new function.
13946 (pass_lower_tm): Replace declaration with that of...
13947 (make_pass_lower_tm): ...new function.
13948 (pass_tm_init): Replace declaration with that of...
13949 (make_pass_tm_init): ...new function.
13950 (pass_tm_mark): Replace declaration with that of...
13951 (make_pass_tm_mark): ...new function.
13952 (pass_tm_memopt): Replace declaration with that of...
13953 (make_pass_tm_memopt): ...new function.
13954 (pass_tm_edges): Replace declaration with that of...
13955 (make_pass_tm_edges): ...new function.
13956 (pass_split_functions): Replace declaration with that of...
13957 (make_pass_split_functions): ...new function.
13958 (pass_feedback_split_functions): Replace declaration with that of...
13959 (make_pass_feedback_split_functions): ...new function.
13960 (pass_strength_reduction): Replace declaration with that of...
13961 (make_pass_strength_reduction): ...new function.
13962 (pass_ipa_lower_emutls): Replace declaration with that of...
13963 (make_pass_ipa_lower_emutls): ...new function.
13964 (pass_ipa_function_and_variable_visibility): Replace declaration with
13966 (make_pass_ipa_function_and_variable_visibility): ...new function.
13967 (pass_ipa_tree_profile): Replace declaration with that of...
13968 (make_pass_ipa_tree_profile): ...new function.
13969 (pass_early_local_passes): Replace declaration with that of...
13970 (make_pass_early_local_passes): ...new function.
13971 (pass_ipa_whole_program_visibility): Replace declaration with that
13973 (make_pass_ipa_whole_program_visibility): ...new function.
13974 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13975 (make_pass_ipa_lto_gimple_out): ...new function.
13976 (pass_ipa_increase_alignment): Replace declaration with that of...
13977 (make_pass_ipa_increase_alignment): ...new function.
13978 (pass_ipa_inline): Replace declaration with that of...
13979 (make_pass_ipa_inline): ...new function.
13980 (pass_ipa_free_lang_data): Replace declaration with that of...
13981 (make_pass_ipa_free_lang_data): ...new function.
13982 (pass_ipa_free_inline_summary): Replace declaration with that of...
13983 (make_pass_ipa_free_inline_summary): ...new function.
13984 (pass_ipa_cp): Replace declaration with that of...
13985 (make_pass_ipa_cp): ...new function.
13986 (pass_ipa_reference): Replace declaration with that of...
13987 (make_pass_ipa_reference): ...new function.
13988 (pass_ipa_pure_const): Replace declaration with that of...
13989 (make_pass_ipa_pure_const): ...new function.
13990 (pass_ipa_pta): Replace declaration with that of...
13991 (make_pass_ipa_pta): ...new function.
13992 (pass_ipa_lto_finish_out): Replace declaration with that of...
13993 (make_pass_ipa_lto_finish_out): ...new function.
13994 (pass_ipa_tm): Replace declaration with that of...
13995 (make_pass_ipa_tm): ...new function.
13996 (pass_ipa_profile): Replace declaration with that of...
13997 (make_pass_ipa_profile): ...new function.
13998 (pass_ipa_cdtor_merge): Replace declaration with that of...
13999 (make_pass_ipa_cdtor_merge): ...new function.
14000 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
14002 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14003 (pass_init_datastructures): Replace declaration with that of...
14004 (make_pass_init_datastructures): ...new function.
14005 (pass_fixup_cfg): Replace declaration with that of...
14006 (make_pass_fixup_cfg): ...new function.
14007 (pass_expand): Replace declaration with that of...
14008 (make_pass_expand): ...new function.
14009 (pass_instantiate_virtual_regs): Replace declaration with that of...
14010 (make_pass_instantiate_virtual_regs): ...new function.
14011 (pass_rtl_fwprop): Replace declaration with that of...
14012 (make_pass_rtl_fwprop): ...new function.
14013 (pass_rtl_fwprop_addr): Replace declaration with that of...
14014 (make_pass_rtl_fwprop_addr): ...new function.
14015 (pass_jump): Replace declaration with that of...
14016 (make_pass_jump): ...new function.
14017 (pass_jump2): Replace declaration with that of...
14018 (make_pass_jump2): ...new function.
14019 (pass_lower_subreg): Replace declaration with that of...
14020 (make_pass_lower_subreg): ...new function.
14021 (pass_cse): Replace declaration with that of...
14022 (make_pass_cse): ...new function.
14023 (pass_fast_rtl_dce): Replace declaration with that of...
14024 (make_pass_fast_rtl_dce): ...new function.
14025 (pass_ud_rtl_dce): Replace declaration with that of...
14026 (make_pass_ud_rtl_dce): ...new function.
14027 (pass_rtl_dce): Replace declaration with that of...
14028 (make_pass_rtl_dce): ...new function.
14029 (pass_rtl_dse1): Replace declaration with that of...
14030 (make_pass_rtl_dse1): ...new function.
14031 (pass_rtl_dse2): Replace declaration with that of...
14032 (make_pass_rtl_dse2): ...new function.
14033 (pass_rtl_dse3): Replace declaration with that of...
14034 (make_pass_rtl_dse3): ...new function.
14035 (pass_rtl_cprop): Replace declaration with that of...
14036 (make_pass_rtl_cprop): ...new function.
14037 (pass_rtl_pre): Replace declaration with that of...
14038 (make_pass_rtl_pre): ...new function.
14039 (pass_rtl_hoist): Replace declaration with that of...
14040 (make_pass_rtl_hoist): ...new function.
14041 (pass_rtl_store_motion): Replace declaration with that of...
14042 (make_pass_rtl_store_motion): ...new function.
14043 (pass_cse_after_global_opts): Replace declaration with that of...
14044 (make_pass_cse_after_global_opts): ...new function.
14045 (pass_rtl_ifcvt): Replace declaration with that of...
14046 (make_pass_rtl_ifcvt): ...new function.
14047 (pass_into_cfg_layout_mode): Replace declaration with that of...
14048 (make_pass_into_cfg_layout_mode): ...new function.
14049 (pass_outof_cfg_layout_mode): Replace declaration with that of...
14050 (make_pass_outof_cfg_layout_mode): ...new function.
14051 (pass_loop2): Replace declaration with that of...
14052 (make_pass_loop2): ...new function.
14053 (pass_rtl_loop_init): Replace declaration with that of...
14054 (make_pass_rtl_loop_init): ...new function.
14055 (pass_rtl_move_loop_invariants): Replace declaration with that of...
14056 (make_pass_rtl_move_loop_invariants): ...new function.
14057 (pass_rtl_unswitch): Replace declaration with that of...
14058 (make_pass_rtl_unswitch): ...new function.
14059 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
14060 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14061 (pass_rtl_doloop): Replace declaration with that of...
14062 (make_pass_rtl_doloop): ...new function.
14063 (pass_rtl_loop_done): Replace declaration with that of...
14064 (make_pass_rtl_loop_done): ...new function.
14065 (pass_web): Replace declaration with that of...
14066 (make_pass_web): ...new function.
14067 (pass_cse2): Replace declaration with that of...
14068 (make_pass_cse2): ...new function.
14069 (pass_df_initialize_opt): Replace declaration with that of...
14070 (make_pass_df_initialize_opt): ...new function.
14071 (pass_df_initialize_no_opt): Replace declaration with that of...
14072 (make_pass_df_initialize_no_opt): ...new function.
14073 (pass_reginfo_init): Replace declaration with that of...
14074 (make_pass_reginfo_init): ...new function.
14075 (pass_inc_dec): Replace declaration with that of...
14076 (make_pass_inc_dec): ...new function.
14077 (pass_stack_ptr_mod): Replace declaration with that of...
14078 (make_pass_stack_ptr_mod): ...new function.
14079 (pass_initialize_regs): Replace declaration with that of...
14080 (make_pass_initialize_regs): ...new function.
14081 (pass_combine): Replace declaration with that of...
14082 (make_pass_combine): ...new function.
14083 (pass_if_after_combine): Replace declaration with that of...
14084 (make_pass_if_after_combine): ...new function.
14085 (pass_ree): Replace declaration with that of...
14086 (make_pass_ree): ...new function.
14087 (pass_partition_blocks): Replace declaration with that of...
14088 (make_pass_partition_blocks): ...new function.
14089 (pass_match_asm_constraints): Replace declaration with that of...
14090 (make_pass_match_asm_constraints): ...new function.
14091 (pass_regmove): Replace declaration with that of...
14092 (make_pass_regmove): ...new function.
14093 (pass_split_all_insns): Replace declaration with that of...
14094 (make_pass_split_all_insns): ...new function.
14095 (pass_fast_rtl_byte_dce): Replace declaration with that of...
14096 (make_pass_fast_rtl_byte_dce): ...new function.
14097 (pass_lower_subreg2): Replace declaration with that of...
14098 (make_pass_lower_subreg2): ...new function.
14099 (pass_mode_switching): Replace declaration with that of...
14100 (make_pass_mode_switching): ...new function.
14101 (pass_sms): Replace declaration with that of...
14102 (make_pass_sms): ...new function.
14103 (pass_sched): Replace declaration with that of...
14104 (make_pass_sched): ...new function.
14105 (pass_ira): Replace declaration with that of...
14106 (make_pass_ira): ...new function.
14107 (pass_reload): Replace declaration with that of...
14108 (make_pass_reload): ...new function.
14109 (pass_clean_state): Replace declaration with that of...
14110 (make_pass_clean_state): ...new function.
14111 (pass_branch_prob): Replace declaration with that of...
14112 (make_pass_branch_prob): ...new function.
14113 (pass_value_profile_transformations): Replace declaration with that
14115 (make_pass_value_profile_transformations): ...new function.
14116 (pass_postreload_cse): Replace declaration with that of...
14117 (make_pass_postreload_cse): ...new function.
14118 (pass_gcse2): Replace declaration with that of...
14119 (make_pass_gcse2): ...new function.
14120 (pass_split_after_reload): Replace declaration with that of...
14121 (make_pass_split_after_reload): ...new function.
14122 (pass_branch_target_load_optimize1): Replace declaration with that
14124 (make_pass_branch_target_load_optimize1): ...new function.
14125 (pass_thread_prologue_and_epilogue): Replace declaration with that
14127 (make_pass_thread_prologue_and_epilogue): ...new function.
14128 (pass_stack_adjustments): Replace declaration with that of...
14129 (make_pass_stack_adjustments): ...new function.
14130 (pass_peephole2): Replace declaration with that of...
14131 (make_pass_peephole2): ...new function.
14132 (pass_if_after_reload): Replace declaration with that of...
14133 (make_pass_if_after_reload): ...new function.
14134 (pass_regrename): Replace declaration with that of...
14135 (make_pass_regrename): ...new function.
14136 (pass_cprop_hardreg): Replace declaration with that of...
14137 (make_pass_cprop_hardreg): ...new function.
14138 (pass_reorder_blocks): Replace declaration with that of...
14139 (make_pass_reorder_blocks): ...new function.
14140 (pass_branch_target_load_optimize2): Replace declaration with that
14142 (make_pass_branch_target_load_optimize2): ...new function.
14143 (pass_leaf_regs): Replace declaration with that of...
14144 (make_pass_leaf_regs): ...new function.
14145 (pass_split_before_sched2): Replace declaration with that of...
14146 (make_pass_split_before_sched2): ...new function.
14147 (pass_compare_elim_after_reload): Replace declaration with that of...
14148 (make_pass_compare_elim_after_reload): ...new function.
14149 (pass_sched2): Replace declaration with that of...
14150 (make_pass_sched2): ...new function.
14151 (pass_stack_regs): Replace declaration with that of...
14152 (make_pass_stack_regs): ...new function.
14153 (pass_stack_regs_run): Replace declaration with that of...
14154 (make_pass_stack_regs_run): ...new function.
14155 (pass_df_finish): Replace declaration with that of...
14156 (make_pass_df_finish): ...new function.
14157 (pass_compute_alignments): Replace declaration with that of...
14158 (make_pass_compute_alignments): ...new function.
14159 (pass_duplicate_computed_gotos): Replace declaration with that of...
14160 (make_pass_duplicate_computed_gotos): ...new function.
14161 (pass_variable_tracking): Replace declaration with that of...
14162 (make_pass_variable_tracking): ...new function.
14163 (pass_free_cfg): Replace declaration with that of...
14164 (make_pass_free_cfg): ...new function.
14165 (pass_machine_reorg): Replace declaration with that of...
14166 (make_pass_machine_reorg): ...new function.
14167 (pass_cleanup_barriers): Replace declaration with that of...
14168 (make_pass_cleanup_barriers): ...new function.
14169 (pass_delay_slots): Replace declaration with that of...
14170 (make_pass_delay_slots): ...new function.
14171 (pass_split_for_shorten_branches): Replace declaration with that of...
14172 (make_pass_split_for_shorten_branches): ...new function.
14173 (pass_split_before_regstack): Replace declaration with that of...
14174 (make_pass_split_before_regstack): ...new function.
14175 (pass_convert_to_eh_region_ranges): Replace declaration with that
14177 (make_pass_convert_to_eh_region_ranges): ...new function.
14178 (pass_shorten_branches): Replace declaration with that of...
14179 (make_pass_shorten_branches): ...new function.
14180 (pass_set_nothrow_function_flags): Replace declaration with that of...
14181 (make_pass_set_nothrow_function_flags): ...new function.
14182 (pass_dwarf2_frame): Replace declaration with that of...
14183 (make_pass_dwarf2_frame): ...new function.
14184 (pass_final): Replace declaration with that of...
14185 (make_pass_final): ...new function.
14186 (pass_rtl_seqabstr): Replace declaration with that of...
14187 (make_pass_rtl_seqabstr): ...new function.
14188 (pass_release_ssa_names): Replace declaration with that of...
14189 (make_pass_release_ssa_names): ...new function.
14190 (pass_early_inline): Replace declaration with that of...
14191 (make_pass_early_inline): ...new function.
14192 (pass_inline_parameters): Replace declaration with that of...
14193 (make_pass_inline_parameters): ...new function.
14194 (pass_update_address_taken): Replace declaration with that of...
14195 (make_pass_update_address_taken): ...new function.
14196 (pass_convert_switch): Replace declaration with that of...
14197 (make_pass_convert_switch): ...new function.
14198 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
14199 to a subclass of simple_ipa_opt_pass along with...
14200 (pass_data_ipa_tree_profile): ...new pass_data instance and...
14201 (make_pass_ipa_tree_profile): ...new function.
14202 * tree-sra.c (pass_sra_early): Convert from a global struct to a
14203 subclass of gimple_opt_pass along with...
14204 (pass_data_sra_early): ...new pass_data instance and...
14205 (make_pass_sra_early): ...new function.
14206 (pass_sra): Convert from a global struct to a subclass of
14207 gimple_opt_pass along with...
14208 (pass_data_sra): ...new pass_data instance and...
14209 (make_pass_sra): ...new function.
14210 (pass_early_ipa_sra): Convert from a global struct to a subclass of
14211 gimple_opt_pass along with...
14212 (pass_data_early_ipa_sra): ...new pass_data instance and...
14213 (make_pass_early_ipa_sra): ...new function.
14214 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
14215 subclass of gimple_opt_pass along with...
14216 (pass_data_ccp): ...new pass_data instance and...
14217 (make_pass_ccp): ...new function.
14218 (pass_fold_builtins): Convert from a global struct to a subclass of
14219 gimple_opt_pass along with...
14220 (pass_data_fold_builtins): ...new pass_data instance and...
14221 (make_pass_fold_builtins): ...new function.
14222 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
14223 subclass of gimple_opt_pass along with...
14224 (pass_data_copy_prop): ...new pass_data instance and...
14225 (make_pass_copy_prop): ...new function.
14226 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
14227 global struct to a subclass of gimple_opt_pass along with...
14228 (pass_data_rename_ssa_copies): ...new pass_data instance and...
14229 (make_pass_rename_ssa_copies): ...new function.
14230 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
14231 subclass of gimple_opt_pass along with...
14232 (pass_data_dce): ...new pass_data instance and...
14233 (make_pass_dce): ...new function.
14234 (pass_dce_loop): Convert from a global struct to a subclass of
14235 gimple_opt_pass along with...
14236 (pass_data_dce_loop): ...new pass_data instance and...
14237 (make_pass_dce_loop): ...new function.
14238 (pass_cd_dce): Convert from a global struct to a subclass of
14239 gimple_opt_pass along with...
14240 (pass_data_cd_dce): ...new pass_data instance and...
14241 (make_pass_cd_dce): ...new function.
14242 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
14243 subclass of gimple_opt_pass along with...
14244 (pass_data_dominator): ...new pass_data instance and...
14245 (make_pass_dominator): ...new function.
14246 (pass_phi_only_cprop): Convert from a global struct to a subclass of
14247 gimple_opt_pass along with...
14248 (pass_data_phi_only_cprop): ...new pass_data instance and...
14249 (make_pass_phi_only_cprop): ...new function.
14250 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
14251 subclass of gimple_opt_pass along with...
14252 (pass_data_dse): ...new pass_data instance and...
14253 (make_pass_dse): ...new function.
14254 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
14255 a subclass of gimple_opt_pass along with...
14256 (pass_data_forwprop): ...new pass_data instance and...
14257 (make_pass_forwprop): ...new function.
14258 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
14259 struct to a subclass of gimple_opt_pass along with...
14260 (pass_data_tree_ifcombine): ...new pass_data instance and...
14261 (make_pass_tree_ifcombine): ...new function.
14262 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
14263 subclass of gimple_opt_pass along with...
14264 (pass_data_ch): ...new pass_data instance and...
14265 (make_pass_ch): ...new function.
14266 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
14267 subclass of gimple_opt_pass along with...
14268 (pass_data_tree_loop): ...new pass_data instance and...
14269 (make_pass_tree_loop): ...new function.
14270 (pass_tree_loop_init): Convert from a global struct to a subclass of
14271 gimple_opt_pass along with...
14272 (pass_data_tree_loop_init): ...new pass_data instance and...
14273 (make_pass_tree_loop_init): ...new function.
14274 (pass_lim): Convert from a global struct to a subclass of
14275 gimple_opt_pass along with...
14276 (pass_data_lim): ...new pass_data instance and...
14277 (make_pass_lim): ...new function.
14278 (pass_tree_unswitch): Convert from a global struct to a subclass of
14279 gimple_opt_pass along with...
14280 (pass_data_tree_unswitch): ...new pass_data instance and...
14281 (make_pass_tree_unswitch): ...new function.
14282 (pass_predcom): Convert from a global struct to a subclass of
14283 gimple_opt_pass along with...
14284 (pass_data_predcom): ...new pass_data instance and...
14285 (make_pass_predcom): ...new function.
14286 (pass_vectorize): Convert from a global struct to a subclass of
14287 gimple_opt_pass along with...
14288 (pass_data_vectorize): ...new pass_data instance and...
14289 (make_pass_vectorize): ...new function.
14290 (pass_graphite): Convert from a global struct to a subclass of
14291 gimple_opt_pass along with...
14292 (pass_data_graphite): ...new pass_data instance and...
14293 (make_pass_graphite): ...new function.
14294 (pass_graphite_transforms): Convert from a global struct to a subclass
14295 of gimple_opt_pass along with...
14296 (pass_data_graphite_transforms): ...new pass_data instance and...
14297 (make_pass_graphite_transforms): ...new function.
14298 (pass_check_data_deps): Convert from a global struct to a subclass of
14299 gimple_opt_pass along with...
14300 (pass_data_check_data_deps): ...new pass_data instance and...
14301 (make_pass_check_data_deps): ...new function.
14302 (pass_iv_canon): Convert from a global struct to a subclass of
14303 gimple_opt_pass along with...
14304 (pass_data_iv_canon): ...new pass_data instance and...
14305 (make_pass_iv_canon): ...new function.
14306 (pass_scev_cprop): Convert from a global struct to a subclass of
14307 gimple_opt_pass along with...
14308 (pass_data_scev_cprop): ...new pass_data instance and...
14309 (make_pass_scev_cprop): ...new function.
14310 (pass_record_bounds): Convert from a global struct to a subclass of
14311 gimple_opt_pass along with...
14312 (pass_data_record_bounds): ...new pass_data instance and...
14313 (make_pass_record_bounds): ...new function.
14314 (pass_complete_unroll): Convert from a global struct to a subclass of
14315 gimple_opt_pass along with...
14316 (pass_data_complete_unroll): ...new pass_data instance and...
14317 (make_pass_complete_unroll): ...new function.
14318 (pass_complete_unrolli): Convert from a global struct to a subclass of
14319 gimple_opt_pass along with...
14320 (pass_data_complete_unrolli): ...new pass_data instance and...
14321 (make_pass_complete_unrolli): ...new function.
14322 (pass_parallelize_loops): Convert from a global struct to a subclass
14323 of gimple_opt_pass along with...
14324 (pass_data_parallelize_loops): ...new pass_data instance and...
14325 (make_pass_parallelize_loops): ...new function.
14326 (pass_loop_prefetch): Convert from a global struct to a subclass of
14327 gimple_opt_pass along with...
14328 (pass_data_loop_prefetch): ...new pass_data instance and...
14329 (make_pass_loop_prefetch): ...new function.
14330 (pass_iv_optimize): Convert from a global struct to a subclass of
14331 gimple_opt_pass along with...
14332 (pass_data_iv_optimize): ...new pass_data instance and...
14333 (make_pass_iv_optimize): ...new function.
14334 (pass_tree_loop_done): Convert from a global struct to a subclass of
14335 gimple_opt_pass along with...
14336 (pass_data_tree_loop_done): ...new pass_data instance and...
14337 (make_pass_tree_loop_done): ...new function.
14338 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14339 struct to a subclass of gimple_opt_pass along with...
14340 (pass_data_cse_reciprocals): ...new pass_data instance and...
14341 (make_pass_cse_reciprocals): ...new function.
14342 (pass_cse_sincos): Convert from a global struct to a subclass of
14343 gimple_opt_pass along with...
14344 (pass_data_cse_sincos): ...new pass_data instance and...
14345 (make_pass_cse_sincos): ...new function.
14346 (pass_optimize_bswap): Convert from a global struct to a subclass of
14347 gimple_opt_pass along with...
14348 (pass_data_optimize_bswap): ...new pass_data instance and...
14349 (make_pass_optimize_bswap): ...new function.
14350 (pass_optimize_widening_mul): Convert from a global struct to a
14351 subclass of gimple_opt_pass along with...
14352 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14353 (make_pass_optimize_widening_mul): ...new function.
14354 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14355 subclass of gimple_opt_pass along with...
14356 (pass_data_phiopt): ...new pass_data instance and...
14357 (make_pass_phiopt): ...new function.
14358 (pass_cselim): Convert from a global struct to a subclass of
14359 gimple_opt_pass along with...
14360 (pass_data_cselim): ...new pass_data instance and...
14361 (make_pass_cselim): ...new function.
14362 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14363 subclass of gimple_opt_pass along with...
14364 (pass_data_phiprop): ...new pass_data instance and...
14365 (make_pass_phiprop): ...new function.
14366 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14367 subclass of gimple_opt_pass along with...
14368 (pass_data_pre): ...new pass_data instance and...
14369 (make_pass_pre): ...new function.
14370 (pass_fre): Convert from a global struct to a subclass of
14371 gimple_opt_pass along with...
14372 (pass_data_fre): ...new pass_data instance and...
14373 (make_pass_fre): ...new function.
14374 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14375 subclass of gimple_opt_pass along with...
14376 (pass_data_reassoc): ...new pass_data instance and...
14377 (make_pass_reassoc): ...new function.
14378 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14379 subclass of gimple_opt_pass along with...
14380 (pass_data_sink_code): ...new pass_data instance and...
14381 (make_pass_sink_code): ...new function.
14382 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14383 subclass of gimple_opt_pass along with...
14384 (pass_data_strlen): ...new pass_data instance and...
14385 (make_pass_strlen): ...new function.
14386 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14387 struct to a subclass of gimple_opt_pass along with...
14388 (pass_data_build_alias): ...new pass_data instance and...
14389 (make_pass_build_alias): ...new function.
14390 (pass_build_ealias): Convert from a global struct to a subclass of
14391 gimple_opt_pass along with...
14392 (pass_data_build_ealias): ...new pass_data instance and...
14393 (make_pass_build_ealias): ...new function.
14394 (pass_ipa_pta): Convert from a global struct to a subclass of
14395 simple_ipa_opt_pass along with...
14396 (pass_data_ipa_pta): ...new pass_data instance and...
14397 (make_pass_ipa_pta): ...new function.
14398 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14399 subclass of gimple_opt_pass along with...
14400 (pass_data_uncprop): ...new pass_data instance and...
14401 (make_pass_uncprop): ...new function.
14402 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14403 global struct to a subclass of gimple_opt_pass along with...
14404 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14405 (make_pass_late_warn_uninitialized): ...new function.
14406 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14407 to a subclass of gimple_opt_pass along with...
14408 (pass_data_init_datastructures): ...new pass_data instance and...
14409 (make_pass_init_datastructures): ...new function.
14410 (pass_early_warn_uninitialized): Convert from a global struct to a
14411 subclass of gimple_opt_pass along with...
14412 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14413 (make_pass_early_warn_uninitialized): ...new function.
14414 (pass_update_address_taken): Convert from a global struct to a
14415 subclass of gimple_opt_pass along with...
14416 (pass_data_update_address_taken): ...new pass_data instance and...
14417 (make_pass_update_address_taken): ...new function.
14418 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14419 struct to a subclass of gimple_opt_pass along with...
14420 (pass_data_release_ssa_names): ...new pass_data instance and...
14421 (make_pass_release_ssa_names): ...new function.
14422 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14423 subclass of gimple_opt_pass along with...
14424 (pass_data_stdarg): ...new pass_data instance and...
14425 (make_pass_stdarg): ...new function.
14426 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14427 global struct to a subclass of gimple_opt_pass along with...
14428 (pass_data_convert_switch): ...new pass_data instance and...
14429 (make_pass_convert_switch): ...new function.
14430 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14431 to a subclass of gimple_opt_pass along with...
14432 (pass_data_tail_recursion): ...new pass_data instance and...
14433 (make_pass_tail_recursion): ...new function.
14434 (pass_tail_calls): Convert from a global struct to a subclass of
14435 gimple_opt_pass along with...
14436 (pass_data_tail_calls): ...new pass_data instance and...
14437 (make_pass_tail_calls): ...new function.
14438 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14439 struct to a subclass of gimple_opt_pass along with...
14440 (pass_data_lower_vector): ...new pass_data instance and...
14441 (make_pass_lower_vector): ...new function.
14442 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14443 gimple_opt_pass along with...
14444 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14445 (make_pass_lower_vector_ssa): ...new function.
14446 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14447 to a subclass of gimple_opt_pass along with...
14448 (pass_data_slp_vectorize): ...new pass_data instance and...
14449 (make_pass_slp_vectorize): ...new function.
14450 (pass_ipa_increase_alignment): Convert from a global struct to a
14451 subclass of simple_ipa_opt_pass along with...
14452 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14453 (make_pass_ipa_increase_alignment): ...new function.
14454 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14455 gimple_opt_pass along with...
14456 (pass_data_vrp): ...new pass_data instance and...
14457 (make_pass_vrp): ...new function.
14458 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14459 subclass of simple_ipa_opt_pass along with...
14460 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14461 (make_pass_ipa_free_lang_data): ...new function.
14462 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14463 gimple_opt_pass along with...
14464 (pass_data_tsan): ...new pass_data instance and...
14465 (make_pass_tsan): ...new function.
14466 (pass_tsan_O0): Convert from a global struct to a subclass of
14467 gimple_opt_pass along with...
14468 (pass_data_tsan_O0): ...new pass_data instance and...
14469 (make_pass_tsan_O0): ...new function.
14470 * var-tracking.c (pass_variable_tracking): Convert from a global
14471 struct to a subclass of rtl_opt_pass along with...
14472 (pass_data_variable_tracking): ...new pass_data instance and...
14473 (make_pass_variable_tracking): ...new function.
14474 * web.c (pass_web): Convert from a global struct to a subclass of
14475 rtl_opt_pass along with...
14476 (pass_data_web): ...new pass_data instance and...
14477 (make_pass_web): ...new function.
14478 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14479 declaration with that of...
14480 (make_pass_mode_switch_use): ...new function.
14481 (pass_resolve_sw_modes): Replace declaration with that of...
14482 (make_pass_resolve_sw_modes): ...new function.
14483 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14484 from a global struct to a subclass of rtl_opt_pass along with...
14485 (pass_data_mode_switch_use): ...new pass_data instance and...
14486 (make_pass_mode_switch_use): ...new function.
14487 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14488 from a global struct to a subclass of rtl_opt_pass along with...
14489 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14490 (make_pass_resolve_sw_modes): ...new function.
14491 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14492 struct to a subclass of rtl_opt_pass along with...
14493 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14494 (make_pass_insert_vzeroupper): ...new function.
14495 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14496 global struct to a subclass of rtl_opt_pass along with...
14497 (pass_data_work_around_errata): ...new pass_data instance and...
14498 (make_pass_work_around_errata): ...new function.
14499 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14500 struct to a subclass of rtl_opt_pass along with...
14501 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14502 (make_pass_mips_machine_reorg2): ...new function.
14504 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14506 * passes.c (pass_manager::operator new): New.
14508 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14510 Handwritten part of conversion of passes to C++ classes.
14512 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14513 (toplev.o): Add dep on PASS_MANAGER_H.
14514 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14515 of early local pases to reflect this moving from a global to a
14516 member of gcc::pass_manager.
14517 (cgraph_add_new_function): Likewise.
14518 * lto-cgraph.c (lto_output_node): Update for conversion of
14519 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14520 * passes.c (opt_pass::clone): New.
14521 (opt_pass::gate): New.
14522 (opt_pass::execute): New.
14523 (opt_pass::opt_pass): New.
14524 (pass_manager::execute_early_local_passes): New.
14525 (pass_manager::execute_pass_mode_switching): new.
14526 (finish_optimization_passes): Convert to...
14527 (pass_manager::finish_optimization_passes): ...this.
14528 (finish_optimization_passes): Update for conversion of passes to
14530 (register_dump_files_1): Use has_gate since we cannot portably
14531 check a vtable entry against NULL.
14532 (dump_one_pass): Likewise.
14533 (ipa_write_summaries_2): Likewise.
14534 (ipa_write_optimization_summaries_1): Likewise.
14535 (ipa_read_summaries_1): Likewise.
14536 (ipa_read_optimization_summaries_1): Likewise.
14537 (execute_ipa_stmt_fixups): Likewise.
14538 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14539 pass-creation functions rather than wiring up globals, and
14540 storing the results in fields of pass_manager generated using
14541 pass-instances.def.
14542 (pass_manager::dump_profile_report): Update for conversion of
14543 passes to C++ classes.
14544 (pass_manager::execute_ipa_summary_passes): Likewise.
14545 (execute_one_ipa_transform_pass): Likewise.
14546 (execute_one_pass): Use has_gate and has_execute since we cannot
14547 portably check a vtable entry against NULL.
14548 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14549 (pass_manager): Use pass-instances.def to add fields for the
14550 various pass instances.
14551 * toplev.c (finalize): Update for move of
14552 finish_optimization_passes to a method of gcc::pass_manager.
14553 * toplev.h (finish_optimization_passes): Move to method of class
14555 * tree-pass.h (struct pass_data): New.
14556 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14557 (opt_pass::gate): Convert to virtual function.
14558 (opt_pass::~opt_pass): New.
14559 (opt_pass::clone): New.
14560 (opt_pass::execute): Convert to virtual function.
14561 (opt_pass::opt_pass): New.
14562 (opt_pass::ctxt_): new.
14563 (gimple_opt_pass): Convert to subclass of opt_pass.
14564 (gimple_opt_pass::gimple_opt_pass): New.
14565 (rtl_opt_pass): Convert to subclass of opt_pass.
14566 (rtl_opt_pass::rtl_opt_pass): New.
14567 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14568 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14569 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14570 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14571 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14572 invocation of pass_mode_switching to reflect this moving from a
14573 global to a member of gcc::pass_manager.
14574 (ix86_option_override): Rework how pass_insert_vzeroupper is
14575 added to the pass_manager to reflect autogenerated changes.
14576 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14578 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14580 PR rtl-optimization/57708
14581 * recog.c (peep2_find_free_register): Validate all regs in a
14584 2013-08-05 Jan Hubicka <jh@suse.cz>
14587 * cgraph.c (verify_cgraph_node): Accept local flags from other
14589 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14590 (function_and_variable_visibility): Likewise.
14591 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14593 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14595 * graph.c (init_graph_slim_pretty_print): Remove.
14596 (print_graph_cfg): Do not call it. Use local pretty printer.
14597 (start_graph_dump): Likewise.
14599 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14601 * gimple-pretty-print.c (buffer): Remove.
14602 (initialized): Likewise.
14603 (maybe_init_pretty_print): Likewise.
14604 (print_gimple_stmt): Do not call it. Use non-static local
14605 pretty_printer variable.
14606 (print_gimple_expr): Likewise.
14607 (print_gimple_seq): Likewise.
14608 (gimple_dump_bb): Likewise.
14610 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14612 * asan.c (asan_pp): Remove.
14613 (asan_pp_initialized): Likewise.
14614 (asan_pp_initialize): Likewise.
14615 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14616 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14617 (asan_add_global): Likewise.
14619 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14621 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14622 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14623 * pretty-print.c (pp_formatted_text_data): Likewise.
14624 (pp_indent): Rename from pp_base_indent.
14625 (pp_format): Rename from pp_base_format.
14626 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14627 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14628 (pp_flush): Rename from pp_base_flush.
14629 (pp_set_line_maximum_length): Rename from
14630 pp_base_set_line_maximum_length.
14631 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14632 (pp_set_prefix): Rename from pp_base_set_prefix.
14633 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14634 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14635 (pp_append_text): Rename from pp_base_append_text.
14636 (pp_formatted_text): Rename from pp_base_formatted_text.
14637 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14638 (pp_remaining_character_count_for_line): Rename from
14639 pp_base_remaining_character_count_for_line.
14640 (pp_newline): Rename from pp_base_newline.
14641 (pp_character): Rename from pp_base_character.
14642 (pp_string): Rename from pp_base_string.
14643 (pp_maybe_space): Rename from pp_base_maybe_space.
14644 * asan.c (asan_pp_string): Adjust.
14645 (asan_emit_stack_protection): Likewise.
14646 (asan_add_global): Likewise.
14647 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14648 * tree-mudflap.c (mf_varname_tree): Likewise.
14649 * tree-pretty-print.c (pp_tree_identifier): Rename from
14650 pp_base_tree_identifier.
14651 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14652 Declare as function.
14654 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14656 * pretty-print.h (pp_bar_bar): New.
14657 (pp_ampersand_ampersand): Likewise.
14658 (pp_less_equal): Likewise.
14659 (pp_greater_equal): Likewise.
14660 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14661 printer functions instead of pp_string or operators and punctuators.
14662 (dump_gimple_call): Likewise.
14663 (dump_gimple_omp_for): Likewise.
14664 (dump_gimple_transaction): Likewise.
14665 (dump_gimple_phi): Likewise.
14666 (pp_gimple_stmt_1): Likewise.
14667 * sched-vis.c (print_insn): Likewise.
14668 * tree-mudflap.c (mf_varname_tree): Likewise.
14669 * tree-pretty-print.c (dump_block_node): Likewise.
14670 (dump_generic_node): Likewise.
14672 2013-08-02 Jan Hubicka <jh@suse.cz>
14674 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14676 * lto-streamer-out.c (tree_is_indexable): Results decls and
14677 parm decls are not indexable.
14678 (DFS_write_tree_body): Do not follow args and results.
14679 (hash_tree): Likewise.
14680 (output_functions): Rearrange so struct function is needed
14681 only when real body is output; be able to also ouptut abstract
14682 functions; output DECL_ARGUMENTS and DECL_RESULT.
14683 (lto_output): When not in WPA, ale store abstract functions.
14684 (write_symbol): Do not care about RESULT_DECL.
14685 (output_symbol_p): Handle correctly sbtract decls.
14686 * lto-streamer-in.c (input_function): Rearrange so struct
14687 function can be NULL at entry; allow streaming of
14688 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14689 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14690 sanity check during LTO.
14691 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14692 RESULT_DECl and DECL_ARGUMENTS.
14693 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14696 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14698 * pretty-print.h (pp_underscore): New.
14700 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14701 printer functions instead of pp_character.
14702 (dump_binary_rhs): Likewise.
14703 (dump_ternary_rhs): Likewise.
14704 (dump_gimple_call_args): Likewise.
14705 (pp_points_to_solution): Likewise.
14706 (dump_gimple_call): Likewise.
14707 (dump_gimple_switch): Likewise.
14708 (dump_gimple_cond): Likewise.
14709 (dump_gimple_bind): Likewise.
14710 (dump_gimple_try): Likewise.
14711 (dump_gimple_omp_for): Likewise.
14712 (dump_gimple_omp_continue): Likewise.
14713 (dump_gimple_omp_single): Likewise.
14714 (dump_gimple_omp_sections): Likewise.
14715 (dump_gimple_omp_block): Likewise.
14716 (dump_gimple_omp_critical): Likewise.
14717 (dump_gimple_transaction): Likewise.
14718 (dump_gimple_asm): Likewise.
14719 (dump_gimple_phi): Likewise.
14720 (dump_gimple_omp_parallel): Likewise.
14721 (dump_gimple_omp_task): Likewise.
14722 (dump_gimple_omp_atomic_load): Likewise.
14723 (dump_gimple_omp_atomic_store): Likewise.
14724 (dump_gimple_mem_ops): Likewise.
14725 (pp_gimple_stmt_1): Likewise.
14726 (pp_cfg_jump): Likewise.
14727 (dump_implicit_edges): Likewise.
14728 (gimple_dump_bb_for_graph): Likewise.
14729 * graph.c (draw_cfg_node): Likewise.
14730 * langhooks.c (lhd_print_error_function): Likewise.
14731 * sched-vis.c (print_exp): Likewise.
14732 (print_value): Likewise.
14733 (print_pattern): Likewise.
14734 (print_insn): Likewise.
14735 (rtl_dump_bb_for_graph): Likewise.
14736 * tree-pretty-print.c (dump_function_declaration): Likewise.
14737 (dump_array_domain): Likewise.
14738 (dump_omp_clause): Likewise.
14739 (dump_location): Likewise.
14740 (dump_generic_node): Likewise.
14741 (print_struct_decl): Likewise.
14742 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14744 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14746 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14747 candidate table when replacing a candidate statement.
14748 (replace_rhs_if_not_dup): Likewise.
14749 (replace_one_candidate): Likewise.
14751 2013-08-02 Jan Hubicka <jh@suse.cz>
14752 Martin Liska <marxin.liska@gmail.com>
14754 * cgraphunit.c (add_new_function): Fix logic when adding from
14756 (assemble_thunk): Rename to ...
14757 (expand_thunk); .. this one; export; get it working with
14758 general functions; make produced gimple valid.
14759 * cgraph.h (expand_thunk): Declare.
14761 2013-08-02 Jan Hubicka <jh@suse.cz>
14763 * ipa-cp.c (gather_context_independent_values): Use
14764 ipa_get_param_move_cost.
14765 (get_replacement_map): Remove PARAM; move parameter folding
14767 (create_specialized_node): Update.
14768 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14769 assert that we have gimple body; update move_cost.
14770 (count_formal_params): Assert that we have gimple body.
14771 (ipa_dump_param): New function.
14772 (ipa_alloc_node_params): Break out from ...
14773 (ipa_initialize_node_params): ... here.
14774 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14775 (ipa_write_node_info): Stream move costs.
14776 (ipa_read_node_info): Read move costs.
14777 (ipa_update_after_lto_read): Do not recompute node params.
14778 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14779 (ipa_get_param): Check we are not in WPA.
14780 (ipa_get_param_move_cost): New.
14781 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14782 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14783 parm numbers to be present.
14785 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14787 PR rtl-optimization/58048
14788 * lra-constraints.c (process_alt_operands): Don't check asm
14789 operand on register.
14791 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14793 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14794 the implied StoreLoad barrier for atomic operations if before.
14796 2013-08-02 Jan Hubicka <jh@suse.cz>
14797 Martin Liska <marxin.liska@gmail.com>
14799 * cgraph.c (cgraph_function_body_availability): Do not check
14801 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14802 symtab_node_availability): Declare.
14803 * ipa.c (can_replace_by_local_alias): New.
14804 (function_and_variable_visibility): Use it.
14805 * symtab.c (symtab_for_node_and_aliases,
14806 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14808 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14810 PR rtl-optimization/57963
14811 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14812 (lra_constraints): Use them.
14814 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14816 * config/arm/types.md (define_attr "type"): Add "load_acq"
14818 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14820 (cortex_a53_store1): Likewise.
14822 2013-08-01 Jan Hubicka <jh@suse.cz>
14824 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14825 partitions are not needed.
14827 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14829 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14830 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14831 MAYBE_NON_Q_CLASS_P where appropriate.
14833 2013-08-01 Jan Hubicka <jh@suse.cz>
14835 * cgraph.h (release_function_body): Declare.
14836 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14837 of unused delcarations.
14839 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14841 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14842 RTL form when subtracting a constant.
14844 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14846 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14847 Generate canonical plus rtx with negated immediate instead of minus
14849 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14851 2013-08-01 Jan Hubicka <jh@suse.cz>
14853 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14854 (cgraph_release_function_body): Likewise.
14855 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14856 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14857 to used_as_abstract_origin.
14858 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14859 symbols used as abstract origins.
14860 * cgraphunit.c (analyze_functions): Update.
14861 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14862 used_as_abstract_origin.
14863 * tree-inline.c (tree_function_versioning): Update
14864 used_as_abstract_origin; be ready for DECL_RESULT and
14865 DECL_ARGUMENTS to be NULL.
14867 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14868 for abstract functions.
14869 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14872 2013-08-01 Jan Hubicka <jh@suse.cz>
14874 * profile.c (compute_value_histograms): Fix thinko.
14876 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14878 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14879 aarch-common-protos.h to extra_headers.
14880 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14881 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14882 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14884 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14886 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14887 Include "../arm/types.md". Define "type" attribute for all patterns.
14888 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14891 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14893 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14894 to support power8 load fusion.
14895 (fusion_gpr_mem_load): Likewise.
14897 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14899 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14900 declarations for power8 load fusion.
14901 (emit_fusion_gpr_load): Likewise.
14903 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14904 tuning for power8, turn on fusion mode by default. Turn on sign
14905 extending fusion mode if normal fusion mode is on, and we are at
14907 (fusion_gpr_load_p): New function, return true if we can fuse an
14908 addis instruction with a dependent load to a GPR.
14909 (emit_fusion_gpr_load): Emit the instructions for power8 load
14912 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14913 (VSX load fusion peepholes): New peepholes to fuse together an
14914 addi instruction with a VSX load instruction.
14916 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14917 peepholes to fuse an addis instruction with a load to a GPR base
14918 register. If we are supporting sign extending fusions, convert
14919 sign extending loads to zero extending loads and add an explicit
14922 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14924 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14925 aarch-common-protos.h to extra_headers.
14926 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14927 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14928 (arm_early_store_addr_dep): Likewise.
14929 (arm_no_early_alu_shift_dep): Likewise.
14930 (arm_no_early_alu_shift_value_dep): Likewise.
14931 (arm_no_early_mul_dep): Likewise.
14932 (arm_no_early_store_addr_dep): Likewise.
14933 (arm_mac_accumulator_is_mul_result): Likewise.
14934 (arm_mac_accumulator_is_result): Likewise.
14935 * config/arm/aarch-common.c: ... here. New file.
14936 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14938 (arm_early_store_addr_dep): Likewise.
14939 (arm_no_early_alu_shift_dep): Likewise.
14940 (arm_no_early_alu_shift_value_dep): Likewise.
14941 (arm_no_early_mul_dep): Likewise.
14942 (arm_no_early_store_addr_dep): Likewise.
14943 (arm_mac_accumulator_is_mul_result): Likewise.
14944 (arm_mac_accumulator_is_result): Likewise.
14945 * config/arm/aarch-common-protos.h: ... here. New file.
14946 * config/arm/t-arm (aarch-common.o): Define.
14948 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14950 * config/arm/arm.md: Include new file "types.md".
14951 (define_attr "type"): Move from here to ...
14952 (define_attr "mul32"): Likewise.
14953 (define_attr "mul64"): Likewise.
14954 * config/arm/types.md: ... here. New file.
14956 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14958 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14959 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14961 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14963 * gen-pass-instances.awk: Fix offset of substr().
14965 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14967 * Makefile.in (pass-instances.def): New.
14968 (passes.o): Replace dependency on passes.def with one on
14971 * gen-pass-instances.awk: New.
14973 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14974 rather than passes.def, updating local definition of NEXT_PASS
14975 macro to add an extra NUM parameter (currently unused).
14977 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14979 * Makefile.in (PASS_MANAGER_H): New.
14980 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14981 (passes.o): Likewise.
14982 (statistics.o): Likewise.
14983 (cgraphunit.o): Likewise.
14984 (context.o): Depend on PASS_MANAGER_H.
14986 * pass_manager.h: New.
14988 * cgraphunit.c (cgraph_add_new_function): Update for moves
14989 of globals to fields of pass_manager.
14990 (analyze_function): Likewise.
14991 (expand_function): Likewise.
14992 (ipa_passes): Likewise.
14993 (compile): Likewise.
14995 * context.c (context::context): New.
14996 * context.h (context::context): New.
14997 (context::get_passes): New.
14998 (context::passes_): New.
15000 * lto-cgraph.c (input_node): Update for moves of globals to
15001 fields of pass_manager.
15003 * passes.c (all_passes): Remove, in favor of a field of the
15004 same name within the new class pass_manager.
15005 (all_small_ipa_passes): Likewise.
15006 (all_lowering_passes): Likewise.
15007 (all_regular_ipa_passes): Likewise.
15008 (all_late_ipa_passes): Likewise.
15009 (all_lto_gen_passes): Likewise.
15010 (passes_by_id): Likewise.
15011 (passes_by_id_size): Likewise.
15012 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15013 the new class pass_manager.
15014 (set_pass_for_id): Convert to...
15015 (pass_manager::set_pass_for_id): ...method.
15016 (get_pass_for_id): Convert to...
15017 (pass_manager::get_pass_for_id): ...method.
15018 (register_one_dump_file): Move body of implementation into...
15019 (pass_manager::register_one_dump_file): ...here.
15020 (register_dump_files_1): Convert to...
15021 (pass_manager::register_dump_files_1): ...method.
15022 (register_dump_files): Convert to...
15023 (pass_manager::register_dump_files): ...method.
15024 (create_pass_tab): Update for moves of globals to fields of
15026 (dump_passes): Move body of implementation into...
15027 (pass_manager::dump_passes): ...here.
15028 (register_pass): Move body of implementation into...
15029 (pass_manager::register_pass): ...here.
15030 (init_optimization_passes): Convert into...
15031 (pass_manager::pass_manager): ...constructor for new
15032 pass_manager class, and initialize the pass_lists array.
15033 (check_profile_consistency): Update for moves of globals to
15034 fields of pass_manager.
15035 (dump_profile_report): Move body of implementation into...
15036 (pass_manager::dump_profile_report): ...here.
15037 (ipa_write_summaries_1): Update for moves of pass lists from
15038 being globals to fields of pass_manager.
15039 (ipa_write_optimization_summaries): Likewise.
15040 (ipa_read_summaries): Likewise.
15041 (ipa_read_optimization_summaries): Likewise.
15042 (execute_all_ipa_stmt_fixups): Likewise.
15044 * statistics.c (statistics_fini): Update for moves of globals to
15045 fields of pass_manager.
15047 * toplev.c (general_init): Replace call to
15048 init_optimization_passes with construction of the pass_manager
15051 * tree-pass.h (all_passes): Remove, in favor of a field of the
15052 same name within the new class pass_manager.
15053 (all_small_ipa_passes): Likewise.
15054 (all_lowering_passes): Likewise.
15055 (all_regular_ipa_passes): Likewise.
15056 (all_lto_gen_passes): Likewise.
15057 (all_late_ipa_passes): Likewise.
15058 (passes_by_id): Likewise.
15059 (passes_by_id_size): Likewise.
15060 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15061 the new class pass_manager.
15062 (get_pass_for_id): Remove.
15064 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15066 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
15069 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15071 * arm.md (mulhi3): New expand pattern.
15073 2013-07-30 Jan Hubicka <jh@suse.cz>
15074 Martin Liska <marxin.liska@gmail.com>
15076 * profile.c (compute_value_histograms): Do not ICE when
15077 there is mismatch only on some counters.
15079 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15081 PR rtl-optimization/57637
15082 * function.c (move_insn_for_shrink_wrap): Also check the
15083 GEN set of the LIVE problem for the liveness analysis
15084 if it exists, otherwise give up.
15086 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15088 PR tree-optimization/57993
15089 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
15090 replaced statement in the candidate table.
15091 (phi_add_costs): Return infinite cost when the hidden basis does
15092 not dominate all phis on which the candidate is dependent.
15093 (replace_one_candidate): Record replaced statement in the
15096 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
15098 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
15099 (ashlv2si3): New expander.
15100 (*ashlv2si3_i): New define_insn_and_split.
15101 * predicates.md (float_operation): Allow patterns with three
15102 basic sub-patterns.
15104 PR rtl-optimization/58021
15105 * mode-switching.c (create_pre_exit): Always split off preceding
15106 insns if we are not at the basic block head.
15108 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
15110 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
15111 (UCLIBC_DYNAMIC_LINKER): New macro.
15112 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
15114 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
15115 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
15117 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
15118 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
15119 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
15120 for SF and DF modes. Use ieee_quad_format for TF mode.
15121 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
15122 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
15123 (mips_option_override): Handle `-mnan=legacy'.
15124 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
15125 `-mabs=2008' and `-mnan=2008'.
15126 (OPTION_DEFAULT_SPECS): Add "nan" default.
15127 (ASM_SPEC): Handle `-mnan='.
15128 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
15129 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
15130 comment accordingly.
15131 (neg<mode>2): Likewise.
15132 * config/mips/mips.opt (mabs, mnan): New options.
15133 * doc/install.texi (Configuration): Document `--with-nan=' option.
15134 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
15136 (MIPS Options): Document them.
15137 * config.gcc <mips*-*-*>: Handle `--with-nan='.
15138 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
15139 * configure: Regenerate.
15140 * config.in: Regenerate.
15142 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15144 * config/i386/i386.md (float post-reload splitters): Do not check
15145 for subregs of SSE registers.
15147 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15148 H.J. Lu <hongjiu.lu@intel.com>
15152 * config/i386/i386.md (post-reload splitter
15153 to avoid partial SSE reg dependency stalls): New pattern.
15155 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
15157 * config/s390/s390.md ("movcc"): Swap load and store instructions.
15159 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
15161 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15162 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
15164 2013-07-26 Cary Coutant <ccoutant@google.com>
15166 * dwarf2out.c (die_checksum_ordered): Don't include template
15167 instantiations in signature.
15168 (is_template_parameter): New function.
15169 (is_template_instantiation): New function.
15170 (generate_skeleton_bottom_up): Don't include template instantiations
15172 (generate_skeleton): Likewise.
15173 (break_out_comdat_types): Move recursive call to break out nested
15175 (prune_unused_types_mark_generic_parms_dies): Call
15176 is_template_parameter.
15178 2013-07-26 Ian Bolton <ian.bolton@arm.com>
15180 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
15181 uses vector registers.
15182 * config/aarch64/iterators.md: Add attributes rtn and vas.
15184 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15185 Richard Earnshaw <richard.earnshaw@arm.com>
15187 * combine.c (simplify_comparison): Re-canonicalize operands
15189 * config/arm/arm.md (movcond_addsi): New splitter.
15191 2013-07-25 Sterling Augustine <saugustine@google.com>
15193 * dwarf2out.c (size_of_pubnames): Move code to...
15194 (include_pubname_in_output): ...here. New.
15195 (want_pubnames): Rearrange.
15196 (output_pubnames): Call include_pubname_in_output. Move assertion.
15198 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15200 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
15202 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15205 * doc/extend.texi: Remove obsolete builtins. Fix
15206 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
15208 2013-07-25 Jan Hubicka <jh@suse.cz>
15210 * cgraph.c (release_function_body): Break out from ...
15211 (cgraph_release_function_body): ... this one; also release DECL_RESULT
15212 and DECL_ARGUMENTS.
15213 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
15214 old_tree in the map.
15215 (create_specialized_node): Update.
15216 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
15218 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
15219 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
15220 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
15221 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
15224 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15226 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
15227 addsi3_carryin_alt2_<optab>): Correct output template.
15229 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15231 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
15232 Adjust for arm_restrict_it.
15233 Remove trailing whitespace.
15235 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
15237 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
15238 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
15240 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
15242 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
15244 PR rtl-optimization/57960
15245 * lra-constraints.c (process_alt_operands): Use the right mode
15246 when checking strict_low.
15248 2013-07-25 Jan Hubicka <jh@suse.cz>
15250 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
15251 * cgraph.c (cgraph_remove_node): Do not release function body
15252 when in cgraph streaming.
15253 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
15254 in other partitions are not considered reachable; fix handling of
15257 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15259 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
15261 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15266 * config/arm/arm.md ("*sibcall_insn): Replace use of
15267 Ss with US. Adjust output for v5 and v4t.
15268 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
15270 * config/arm/constraints.md ("Ss"): Rename to US.
15272 2013-07-25 Terry Guo <terry.guo@arm.com>
15274 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
15275 shift_add/shift_sub0/shift_sub1 RTXs.
15277 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
15278 Anton Blanchard <anton@au1.ibm.com>
15280 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
15281 (altivec_vpks<VI_char>ss): Likewise.
15282 (altivec_vpks<VI_char>us): Likewise.
15283 (altivec_vpku<VI_char>us): Likewise.
15284 (altivec_vpku<VI_char>um): Likewise.
15286 2013-07-24 David Malcolm <dmalcolm@redhat.com>
15288 Introduce context class.
15290 * Makefile.in (CONTEXT_H): New.
15291 (OBJS): Add context.o.
15292 (toplev.o): Add CONTEXT_H to dependencies.
15295 * toplev.c (general_init): Create the singleton gcc::context instance.
15301 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15303 PR rtl-optimization/57968
15304 * mode-switching.c (create_pre_exit): Allow instructions that
15305 don't set a return register to need a non-exit mode.
15307 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15308 Anton Blanchard <anton@au1.ibm.com>
15310 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15311 operands to vperm for little endian.
15312 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15313 of lvsl to create the control mask for a vperm for little endian.
15315 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15316 Anton Blanchard <anton@au1.ibm.com>
15318 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15319 two operands for little-endian.
15321 2013-07-23 Steve Ellcey <sellcey@mips.com>
15323 * config/mips/mips.c (mips_case_values_threshold): New.
15324 (TARGET_CASE_VALUES_THRESHOLD): Define.
15326 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15327 Anton Blanchard <anton@au1.ibm.com>
15329 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15330 selection of field for vector splat in little endian mode.
15332 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15334 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15335 expanders to rs6000.md.
15336 (ior<mode>3): Likewise.
15337 (and<mode>3): Likewise.
15338 (one_cmpl<mode>2): Likewise.
15339 (nor<mode>3): Likewise.
15340 (andc<mode>3): Likewise.
15341 (eqv<mode>3): Likewise.
15342 (nand<mode>3): Likewise.
15343 (orc<mode>3): Likewise.
15345 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15348 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15349 to split multi-word logical operations.
15350 (rs6000_split_logical_di): Likewise.
15351 (rs6000_split_logical): Likewise.
15353 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15354 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15355 and allow TImode operations in 32-bit.
15356 (vsx_and<mode>3_64bit): Likewise.
15357 (vsx_ior<mode>3_32bit): Likewise.
15358 (vsx_ior<mode>3_64bit): Likewise.
15359 (vsx_xor<mode>3_32bit): Likewise.
15360 (vsx_xor<mode>3_64bit): Likewise.
15361 (vsx_one_cmpl<mode>2_32bit): Likewise.
15362 (vsx_one_cmpl<mode>2_64bit): Likewise.
15363 (vsx_nor<mode>3_32bit): Likewise.
15364 (vsx_nor<mode>3_64bit): Likewise.
15365 (vsx_andc<mode>3_32bit): Likewise.
15366 (vsx_andc<mode>3_64bit): Likewise.
15367 (vsx_eqv<mode>3_32bit): Likewise.
15368 (vsx_eqv<mode>3_64bit): Likewise.
15369 (vsx_nand<mode>3_32bit): Likewise.
15370 (vsx_nand<mode>3_64bit): Likewise.
15371 (vsx_orc<mode>3_32bit): Likewise.
15372 (vsx_orc<mode>3_64bit): Likewise.
15374 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15375 logical types in GPRs.
15377 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15378 logical insns to rs6000.md, and allow TImode operations in
15380 (altivec_ior<mode>3): Likewise.
15381 (altivec_xor<mode>3): Likewise.
15382 (altivec_one_cmpl<mode>2): Likewise.
15383 (altivec_nor<mode>3): Likewise.
15384 (altivec_andc<mode>3): Likewise.
15386 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15387 attributes for moving the 128-bit logical operations into
15389 (BOOL_REGS_OUTPUT): Likewise.
15390 (BOOL_REGS_OP1): Likewise.
15391 (BOOL_REGS_OP2): Likewise.
15392 (BOOL_REGS_UNARY): Likewise.
15393 (BOOL_REGS_AND_CR0): Likewise.
15394 (one_cmpl<mode>2): Add support for DI logical operations on
15395 32-bit, splitting the operations to 32-bit.
15396 (anddi3): Likewise.
15397 (iordi3): Likewise.
15398 (xordi3): Likewise.
15399 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15400 changes to combine the 32/64-bit code, allow logical operations on
15401 TI mode in 32-bit, and to use similar match_operator patterns like
15402 scalar mode uses. Combine the Altivec and VSX code for logical
15403 operations, and move it here.
15404 (ior<mode>3, 128-bit types): Likewise.
15405 (xor<mode>3, 128-bit types): Likewise.
15406 (one_cmpl<mode>3, 128-bit types): Likewise.
15407 (nor<mode>3, 128-bit types): Likewise.
15408 (andc<mode>3, 128-bit types): Likewise.
15409 (eqv<mode>3, 128-bit types): Likewise.
15410 (nand<mode>3, 128-bit types): Likewise.
15411 (orc<mode>3, 128-bit types): Likewise.
15412 (and<mode>3_internal): Likewise.
15413 (bool<mode>3_internal): Likewise.
15414 (boolc<mode>3_internal1): Likewise.
15415 (boolc<mode>3_internal2): Likewise.
15416 (boolcc<mode>3_internal1): Likewise.
15417 (boolcc<mode>3_internal2): Likewise.
15418 (eqv<mode>3_internal1): Likewise.
15419 (eqv<mode>3_internal2): Likewise.
15420 (one_cmpl1<mode>3_internal): Likewise.
15422 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15424 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15425 Rename flag_stack_usage to flag_stack_usage_info.
15427 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15429 * config/microblaze/sync.md: New file.
15430 * config/microblaze/microblaze.md: Include sync.md
15431 * config/microblaze/microblaze.c: Add print_operand 'y'.
15432 * config/microblaze/constraints.md: Add memory_contraint
15433 'Q' which is a single register.
15435 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15437 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15439 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15441 * reload.c (find_reloads): Exit loop once we find this operand
15442 cannot be reloaded somehow for this alternative.
15444 * reload.c (find_reloads): Exit loop once we find a hard register.
15446 * rtlanal.c (computed_jump_p): Exit loop once we find label
15449 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15451 * cfgloopmanip.c (remove_path): Exit loop after setting
15454 * gensupport.c (subst_dup): Avoid loop if code is not
15455 MATCH_DUP nor MATCH_OP_DUP.
15457 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15459 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15461 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15463 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15464 true for SP_REGNUM if mode == ptr_mode.
15465 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15466 with value R0_REGNUM + 31.
15468 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15470 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15471 pad pointer-typed argument downward.
15473 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15475 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15476 and __ILP32__ when the ILP32 model is in use.
15478 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15480 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15481 (aarch64_load_symref_appropriately): In the case of
15482 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15483 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15484 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15485 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15486 if 'mode' doesn't equal to 'ptr_mode'.
15487 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15488 'vcall_offset'; change to call aarch64_emit_move differently depending
15489 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15490 to calculate the upper bound of 'vcall_offset'.
15491 (aarch64_cannot_force_const_mem): Change to also return true if
15493 (aarch64_legitimize_reload_address): In the case of large
15494 displacements, add new local variable 'xmode' and an assertion
15495 based on it; change to use 'xmode' to generate the new rtx and
15497 (aarch64_asm_trampoline_template): Change to generate the template
15498 differently depending on TARGET_ILP32 or not; change to use
15499 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15500 (aarch64_trampoline_size): Removed.
15501 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15502 and replace immediate literals with it. Change to use 'ptr_mode'
15503 instead of 'DImode' and call convert_memory_address if the mode
15504 of 'fnaddr' doesn't equal to 'ptr_mode'.
15505 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15507 (aarch64_elf_asm_destructor): Likewise.
15508 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15509 on TARGET_ILP32 instead of aarch64_trampoline_size.
15510 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15511 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15512 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15513 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15514 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15515 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15516 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15517 depending on the value of 'mode'.
15518 (add_losym_<mode>): New.
15519 (ldr_got_small_<mode>): New, based on ldr_got_small.
15520 (ldr_got_small): Remove.
15521 (ldr_got_small_sidi): New.
15522 * config/aarch64/iterators.md (P): New.
15523 (PTR): Change to 'ptr_mode' in the condition.
15525 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15527 * config.gcc (aarch64*-*-*): Support --with-abi.
15528 (aarch64*-*-elf): Support --with-multilib-list.
15529 (aarch64*-*-linux*): Likewise.
15530 (supported_defaults): Add abi to aarch64*-*-*.
15531 * configure.ac: Mention AArch64 for --with-multilib-list.
15532 * configure: Re-generated.
15533 * config/aarch64/biarchilp32.h: New file.
15534 * config/aarch64/biarchlp64.h: New file.
15535 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15537 (MULTILIB_DEFAULTS): Ditto.
15538 (DRIVER_SELF_SPECS): Ditto.
15539 (ASM_SPEC): Update to also substitute -mabi.
15540 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15541 file whose name depends on -mabi= and -mbig-endian.
15542 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15544 (POINTER_SIZE): New define.
15545 (POINTERS_EXTEND_UNSIGNED): Ditto.
15546 (enum aarch64_abi_type): New enumeration tag.
15547 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15548 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15549 (TARGET_ILP32): New define.
15550 * config/aarch64/aarch64.opt (mabi): New.
15551 (aarch64_abi): New.
15552 (ilp32, lp64): New values for -mabi.
15553 * config/aarch64/t-aarch64 (comma): New define.
15554 (MULTILIB_OPTIONS): Ditto.
15555 (MULTILIB_DIRNAMES): Ditto.
15556 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15557 * doc/invoke.texi: Document -mabi for AArch64.
15559 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15561 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15563 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15564 Anton Blanchard <anton@au1.ibm.com>
15566 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15567 endianness when selecting field to splat.
15569 2013-07-22 Eric Christopher <echristo@gmail.com>
15571 * dwarf2out.c (die_odr_checksum): New function to use
15572 CHECKSUM_ macros and ULEB128 for DIE tag.
15573 (generate_type_signature): Use.
15575 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15577 * config.gcc (sparc*-*-*): Accept leon3 processor.
15578 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15579 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15580 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15581 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15582 (mfix-ut699): Adjust comment.
15583 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15584 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15585 (CPP_CPU_SPEC): Likewise.
15586 (ASM_CPU_SPEC): Likewise.
15587 * config/sparc/sparc.c (leon3_cost): New constant.
15588 (sparc_option_override): Add leon3 support.
15589 (mem_ref): New function.
15590 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15591 (sparc_do_work_around_errata): Look into the instruction in the delay
15592 slot and adjust accordingly. Add fix for the data cache nullify issues
15593 of the UT699. Change insertion position for the NOP.
15594 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15595 (leon3_load): New reservation.
15596 (leon_store): Bump latency to 2.
15597 (grfpu): New automaton.
15598 (grfpu_alu): New unit.
15599 (grfpu_ds): Likewise.
15600 (leon_fp_alu): Adjust.
15601 (leon_fp_mult): Delete.
15602 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15603 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15604 * config/sparc/sparc.md (cpu): Add leon3.
15605 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15606 (swapsi): Likewise.
15607 (atomic_test_and_set): Likewise.
15608 (ldstub): Likewise.
15610 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15612 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15613 default for R5900 targets.
15614 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15615 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15616 * config/mips/mips.c (mips_option_override): Report an error for
15617 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15618 for -march=r5900 -mhard-float.
15620 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15622 * df-problems.c (can_move_insns_across): Exit loop once we
15623 find a non-fixed, non-global register.
15625 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15628 * omega.c (omega_eliminate_red): Break after setting red_found.
15629 (omega_problem_has_red_equations): Similarly after setting found.
15630 (omega_query_variable): Similarly after setting coupled.
15632 2013-07-22 Marek Polacek <polacek@redhat.com>
15634 * gimplify.c: Don't include gimple.h twice.
15636 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15638 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15639 instead of TARGET_THUMB1.
15640 (Pz): New constraint.
15641 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15643 (compare_negsi_si): Likewise.
15644 (compare_addsi2_op0): Likewise.
15645 (compare_addsi2_op1): Likewise.
15646 (addsi3_carryin_<optab>): Likewise.
15647 (addsi3_carryin_alt2_<optab>): Likewise.
15648 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15649 for arm_restrict_it.
15650 (subsi3_carryin): Likewise.
15651 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15652 (minmax_arithsi): Disable for arm_restrict_it.
15653 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15654 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15655 (satsi_<SAT:code>_shift): Likewise.
15656 (arm_shiftsi3): Add alternative for 16-bit encoding.
15657 (arm32_movhf): Disable for arm_restrict_it.
15658 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15659 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15661 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15663 * config/arm/arm.md (attribute "insn"): Delete.
15664 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15665 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15666 (not_shiftsi): Update for attribute change.
15667 (not_shiftsi_compare0): Likewise.
15668 (not_shiftsi_compare0_scratch): Likewise.
15669 (arm_one_cmplsi2): Likewise.
15670 (thumb1_one_cmplsi2): Likewise.
15671 (notsi_compare0): Likewise.
15672 (notsi_compare0_scratch): Likewise.
15673 (thumb1_movdi_insn): Likewise.
15674 (arm_movsi_insn): Likewise.
15675 (movhi_insn_arch4): Likewise.
15676 (movhi_bytes): Likewise.
15677 (arm_movqi_insn): Likewise.
15678 (thumb1_movqi_insn): Likewise.
15679 (arm32_movhf): Likewise.
15680 (thumb1_movhf): Likewise.
15681 (arm_movsf_soft_insn): Likewise.
15682 (thumb1_movsf_insn): Likewise.
15683 (thumb_movdf_insn): Likewise.
15684 (movsicc_insn): Likewise.
15685 (movsfcc_soft_insn): Likewise.
15686 (and_scc): Likewise.
15687 (cond_move): Likewise.
15688 (if_move_not): Likewise.
15689 (if_not_move): Likewise.
15690 (if_shift_move): Likewise.
15691 (if_move_shift): Likewise.
15692 (if_shift_shift): Likewise.
15693 (if_not_arith): Likewise.
15694 (if_arith_not): Likewise.
15695 (cond_move_not): Likewise.
15696 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15697 (neon_mov<mode>): Likewise.
15698 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15699 (thumb2_movsi_vfp): Likewise.
15700 (movsf_vfp): Likewise.
15701 (thumb2_movsf_vfp): Likewise.
15702 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15704 (cortexa7_older_only): Likewise.
15705 (cortexa7_younger): Likewise.
15706 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15707 (1020alu_shift_op): Likewise.
15708 (1020alu_shift_reg_op): Likewise.
15709 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15710 (alu_shift_op): Likewise.
15711 (alu_shift_reg_op): Likewise.
15712 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15713 (11_alu_shift_op): Likewise.
15714 (11_alu_shift_reg_op): Likewise.
15715 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15716 (9_alu_shift_reg_op): Likewise.
15717 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15719 (cortex_a15_alu_shift): Likewise.
15720 (cortex_a15_alu_shift_reg): Likewise.
15721 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15722 (cortex_a5_alu_shift): Likewise.
15723 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15725 (cortex_a53_alu_shift): Likewise.
15726 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15728 (cortex_a7_alu_reg): Likewise.
15729 (cortex_a7_alu_shift): Likewise.
15730 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15731 (cortex_a8_alu_shift): Likewise.
15732 (cortex_a8_alu_shift_reg): Likewise.
15733 (cortex_a8_mov): Likewise.
15734 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15735 (cortex_a9_dp_shift): Likewise.
15736 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15737 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15738 (cortex_r4_mov): Likewise.
15739 (cortex_r4_alu_shift): Likewise.
15740 (cortex_r4_alu_shift_reg): Likewise.
15741 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15742 (526_alu_shift_op): Likewise.
15743 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15744 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15745 (626te_alu_shift_op): Likewise.
15746 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15747 (726te_alu_op): Likewise.
15748 (726te_alu_shift_op): Likewise.
15749 (726te_alu_shift_reg_op): Likewise.
15750 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15751 (mp626_alu_shift_op): Likewise.
15752 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15753 (pj4_alu_e1_conds): Likewise.
15754 (pj4_alu): Likewise.
15755 (pj4_alu_conds): Likewise.
15756 (pj4_shift): Likewise.
15757 (pj4_shift_conds): Likewise.
15758 (pj4_alu_shift): Likewise.
15759 (pj4_alu_shift_conds): Likewise.
15761 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15763 * config/arm/predicates.md (shiftable_operator_strict_it):
15765 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15766 Disable cond_exec version for arm_restrict_it.
15767 (thumb2_smaxsi3): Convert to generate cond_exec.
15768 (thumb2_sminsi3): Likewise.
15769 (thumb32_umaxsi3): Likewise.
15770 (thumb2_uminsi3): Likewise.
15771 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15772 (thumb2_neg_abssi2): Likewise.
15773 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15774 (thumb2_movsicc_insn): Adjust alternatives.
15775 (thumb2_mov_negscc): Disable for arm_restrict_it.
15776 (thumb2_mov_negscc_strict_it): New pattern.
15777 (thumb2_mov_notscc_strict_it): New pattern.
15778 (thumb2_mov_notscc): Disable for arm_restrict_it.
15779 (thumb2_ior_scc): Likewise.
15780 (thumb2_ior_scc_strict_it): New pattern.
15781 (thumb2_cond_move): Adjust for arm_restrict_it.
15782 (thumb2_cond_arith): Disable for arm_restrict_it.
15783 (thumb2_cond_arith_strict_it): New pattern.
15784 (thumb2_cond_sub): Adjust for arm_restrict_it.
15785 (thumb2_movcond): Likewise.
15786 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15787 (thumb2_zero_extendhisi2_v6): Likewise.
15788 (thumb2_zero_extendqisi2_v6): Likewise.
15789 (orsi_notsi_si): Likewise.
15790 (orsi_not_shiftsi_si): Likewise.
15792 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15794 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15795 instruction sequence is 1 byte shorter.
15797 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15799 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15800 it is not needed after split.
15802 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15805 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15806 second label for nonlocal goto receivers. Don't output pic base labels
15807 unless we're producing PIC; mark that action unreachable().
15808 (ix86_save_reg): If the function contains a nonlocal label, save the
15810 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15811 * config/darwin.c (emitted_pic_label_num): New GTY.
15812 (update_pic_label_number_if_needed): New.
15813 (machopic_output_function_base_name): Adjust for nonlocal receiver
15815 (machopic_should_output_picbase_label): New.
15816 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15817 (nonlocal_goto_receiver): New insn and split.
15819 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15821 * config/aarch64/aarch64-builtins.c
15822 (aarch64_fold_builtin): Fold abs in all modes.
15823 * config/aarch64/aarch64-simd-builtins.def
15824 (abs): Enable for all modes.
15825 * config/aarch64/arm_neon.h
15826 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15827 (vabs_f64): Add missing intrinsic.
15829 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15831 * config/aarch64/arm_neon.h (vabs_s64): New function
15833 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15836 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15837 * config/avr/avr.md (adjust_len): Add `round'.
15838 * config/avr/avr-protos.h (avr_out_round): New prototype.
15839 (avr_out_plus): Add `out_label' argument.
15840 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15841 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15842 Handle the case where `insn' is just a pattern.
15843 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15844 (avr_out_round): New function.
15845 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15847 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15849 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15850 Add check for flag_stack_usage to handle -fstack-usage support
15852 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15854 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15855 interaction for new Power8 flags and VSX.
15857 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15859 PR middle-end/57698
15860 * tree-inline.c (expand_call_inline): Emit errors during
15861 early_inlining only if optimization is not turned on.
15863 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15867 * passes.c (init_optimization_passes): Move the construction of
15868 the pass hierarchy into a new passes.def file.
15870 * Makefile.in (passes.o): Add dependency on passes.def.
15872 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15874 * passes.c (init_optimization_passes): Introduce macros for
15875 constructing the tree of passes (INSERT_PASSES_AFTER,
15876 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15877 TERMINATE_PASS_LIST).
15879 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15880 Wei Mi <wmi@google.com>
15882 PR rtl-optimization/57878
15883 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15885 (reload_pseudo_compare_func): Check nregs first for reload
15888 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15890 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15892 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15894 * read-rtl.c (validate_const_int): Once an invalid character is
15895 seen, quit the loop.
15897 * gengtype.c (write_roots): Similarly once we find the "deletable"
15898 or "if_marked" option.
15900 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15902 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15903 "xtab" and "sat". Move value "clz" from here to ...
15904 (attriubte "type"): ... here.
15905 (satsi_<SAT:code>): Delete "insn" attribute.
15906 (satsi_<SAT:code>_shift): Likewise.
15907 (arm_zero_extendqisi2addsi): Likewise.
15908 (arm_extendqisi2addsi): Likewise.
15909 (clzsi2): Update for attribute changes.
15910 (rbitsi2): Likewise.
15911 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15913 (arm_usatsihi): Likewise.
15914 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15916 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15918 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15919 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15920 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15921 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15922 in alphabetical order.
15923 (attribute "core_cycles"): Update for attribute changes.
15924 (arm_addsi3): Likewise.
15925 (addsi3_compare0): Likewise.
15926 (addsi3_compare0_scratch): Likewise.
15927 (addsi3_compare_op1): Likewise.
15928 (addsi3_compare_op2): Likewise.
15929 (compare_addsi2_op0): Likewise.
15930 (compare_addsi2_op1): Likewise.
15931 (addsi3_carryin_shift_<optab>): Likewise.
15932 (subsi3_carryin_shift): Likewise.
15933 (rsbsi3_carryin_shift): Likewise.
15934 (arm_subsi3_insn): Likewise.
15935 (subsi3_compare0): Likewise.
15936 (subsi3_compare): Likewise.
15937 (arm_andsi3_insn): Likewise.
15938 (thumb1_andsi3_insn): Likewise.
15939 (andsi3_compare0): Likewise.
15940 (andsi3_compare0_scratch): Likewise.
15941 (zeroextractsi_compare0_scratch
15942 (andsi_not_shiftsi_si): Likewise.
15943 (iorsi3_insn): Likewise.
15944 (iorsi3_compare0): Likewise.
15945 (iorsi3_compare0_scratch): Likewise.
15946 (arm_xorsi3): Likewise.
15947 (thumb1_xorsi3_insn): Likewise.
15948 (xorsi3_compare0): Likewise.
15949 (xorsi3_compare0_scratch): Likewise.
15950 (satsi_<SAT:code>_shift): Likewise.
15952 (arm_shiftsi3): Likewise.
15953 (shiftsi3_compare0): Likewise.
15954 (not_shiftsi): Likewise.
15955 (not_shiftsi_compare0): Likewise.
15956 (not_shiftsi_compare0_scratch): Likewise.
15957 (arm_one_cmplsi2): Likewise.
15958 (thumb_one_complsi2): Likewise.
15959 (notsi_compare0): Likewise.
15960 (notsi_compare0_scratch): Likewise.
15961 (thumb1_zero_extendhisi2): Likewise.
15962 (arm_zero_extendhisi2): Likewise.
15963 (arm_zero_extendhisi2_v6): Likewise.
15964 (arm_zero_extendhisi2addsi): Likewise.
15965 (thumb1_zero_extendqisi2): Likewise.
15966 (thumb1_zero_extendqisi2_v6): Likewise.
15967 (arm_zero_extendqisi2): Likewise.
15968 (arm_zero_extendqisi2_v6): Likewise.
15969 (arm_zero_extendqisi2addsi): Likewise.
15970 (thumb1_extendhisi2): Likewise.
15971 (arm_extendhisi2): Likewise.
15972 (arm_extendhisi2_v6): Likewise.
15973 (arm_extendqisi): Likewise.
15974 (arm_extendqisi_v6): Likewise.
15975 (arm_extendqisi2addsi): Likewise.
15976 (thumb1_extendqisi2): Likewise.
15977 (thumb1_movdi_insn): Likewise.
15978 (arm_movsi_insn): Likewise.
15979 (movsi_compare0): Likewise.
15980 (movhi_insn_arch4): Likewise.
15981 (movhi_bytes): Likewise.
15982 (arm_movqi_insn): Likewise.
15983 (thumb1_movqi_insn): Likewise.
15984 (arm32_movhf): Likewise.
15985 (thumb1_movhf): Likewise.
15986 (arm_movsf_soft_insn): Likewise.
15987 (thumb1_movsf_insn): Likewise.
15988 (movdf_soft_insn): Likewise.
15989 (thumb_movdf_insn): Likewise.
15990 (arm_cmpsi_insn): Likewise.
15991 (cmpsi_shiftsi): Likewise.
15992 (cmpsi_shiftsi_swp): Likewise.
15993 (arm_cmpsi_negshiftsi_si): Likewise.
15994 (movsicc_insn): Likewise.
15995 (movsfcc_soft_insn): Likewise.
15996 (arith_shiftsi): Likewise.
15997 (arith_shiftsi_compare0
15998 (arith_shiftsi_compare0_scratch
15999 (sub_shiftsi): Likewise.
16000 (sub_shiftsi_compare0
16001 (sub_shiftsi_compare0_scratch
16002 (and_scc): Likewise.
16003 (cond_move): Likewise.
16004 (if_plus_move): Likewise.
16005 (if_move_plus): Likewise.
16006 (if_move_not): Likewise.
16007 (if_not_move): Likewise.
16008 (if_shift_move): Likewise.
16009 (if_move_shift): Likewise.
16010 (if_shift_shift): Likewise.
16011 (if_not_arith): Likewise.
16012 (if_arith_not): Likewise.
16013 (cond_move_not): Likewise.
16014 (thumb1_ashlsi3): Set type attribute.
16015 (thumb1_ashrsi3): Likewise.
16016 (thumb1_lshrsi3): Likewise.
16017 (thumb1_rotrsi3): Likewise.
16018 (shiftsi3_compare0_scratch): Likewise.
16019 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
16020 (neon_mov<mode>): Likewise.
16021 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
16023 (thumb2_movsi_insn): Likewise.
16024 (thumb2_cmpsi_neg_shiftsi): Likewise.
16025 (thumb2_extendqisi_v6): Likewise.
16026 (thumb2_zero_extendhisi2_v6): Likewise.
16027 (thumb2_zero_extendqisi2_v6): Likewise.
16028 (thumb2_shiftsi3_short): Likewise.
16029 (thumb2_addsi3_compare0_scratch): Likewise.
16030 (orsi_not_shiftsi_si): Likewise.
16031 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
16032 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
16034 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
16035 (1020alu_shift_op): Likewise.
16036 (1020alu_shift_reg_op): Likewise.
16037 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
16038 (alu_shift_op): Likewise.
16039 (alu_shift_reg_op): Likewise.
16040 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
16041 (11_alu_shift_op): Likewise.
16042 (11_alu_shift_reg_op): Likewise.
16043 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
16044 (9_alu_shift_reg_op): Likewise.
16045 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16047 (cortex_a15_alu_shift): Likewise.
16048 (cortex_a15_alu_shift_reg): Likewise.
16049 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
16051 (cortex_a5_alu_shift): Likewise.
16052 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
16054 (cortex_a53_alu_shift): Likewise.
16055 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16057 (cortex_a7_alu_reg): Likewise.
16058 (cortex_a7_alu_shift): Likewise.
16059 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
16061 (cortex_a8_alu_shift): Likewise.
16062 (cortex_a8_alu_shift_reg): Likewise.
16063 (cortex_a8_mov): Likewise.
16064 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
16065 (cortex_a9_dp_shift): Likewise.
16066 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
16068 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
16070 (cortex_r4_mov): Likewise.
16071 (cortex_r4_alu_shift): Likewise.
16072 (cortex_r4_alu_shift_reg): Likewise.
16073 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
16074 (526_alu_shift_op): Likewise.
16075 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
16076 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
16077 (626te_alu_shift_op): Likewise.
16078 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
16079 (726te_alu_op): Likewise.
16080 (726te_alu_shift_op): Likewise.
16081 (726te_alu_shift_reg_op): Likewise.
16082 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
16083 (mp626_alu_shift_op): Likewise.
16084 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
16085 (pj4_alu_e1_conds): Likewise.
16086 (pj4_alu): Likewise.
16087 (pj4_alu_conds): Likewise.
16088 (pj4_shift): Likewise.
16089 (pj4_shift_conds): Likewise.
16090 (pj4_alu_shift): Likewise.
16091 (pj4_alu_shift_conds): Likewise.
16092 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16094 (cortexa7_older_only): Likewise.
16095 (cortexa7_younger): Likewise.
16097 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16099 * ipa-pure-const.c (generate_summary): Rename to...
16100 (pure_const_generate_summary): ... this.
16102 2013-07-17 Iain Sandoe <iain@codesourcery.com>
16104 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
16106 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
16109 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
16112 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16114 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
16115 enabled without -march=zEC12.
16116 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
16119 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16121 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
16122 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
16123 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
16124 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
16125 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
16126 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
16127 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
16129 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
16130 and ISA_HAS_NMADD3_NMSUB3.
16131 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
16132 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
16133 (nmsub4<mode>, nmsub3<mode>): Likewise.
16134 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
16136 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16138 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
16139 TARGET_MIPS5400 checking.
16141 2013-07-16 Jakub Jelinek <jakub@redhat.com>
16142 Peter Bergner <bergner@vnet.ibm.com>
16144 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
16145 registers in the comment.
16146 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
16147 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
16148 rather than FIRST_PSEUDO_REGISTERS.
16150 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
16152 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
16153 enable extra ISA flags with TARGET_HTM.
16155 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16157 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
16160 2013-07-15 Cong Hou <congh@google.com>
16162 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
16163 in compare function for sorting.
16165 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
16167 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
16168 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
16169 * config/rs6000/rs6000.opt: Add -mhtm option.
16170 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
16171 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
16172 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16173 __HTM__ if the HTM instructions are available.
16174 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
16175 htm_spr_reg_operand): New define_predicates.
16176 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
16177 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
16179 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
16180 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
16181 HTM builtin functions.
16182 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
16183 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
16184 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
16185 (rs6000_builtin_mask_calculate): Likewise.
16186 (rs6000_option_override_internal): Likewise.
16187 (bdesc_htm): Add new HTM builtin support.
16188 (htm_spr_num): New function.
16189 (htm_spr_regno): Likewise.
16190 (rs6000_htm_spr_icode): Likewise.
16191 (htm_expand_builtin): Likewise.
16192 (htm_init_builtins): Likewise.
16193 (rs6000_expand_builtin): Add support for HTM builtin functions.
16194 (rs6000_init_builtins): Likewise.
16195 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
16197 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
16198 (TARGET_HTM, MASK_HTM): Define macros.
16199 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
16200 (FIXED_REGISTERS): Likewise.
16201 (CALL_USED_REGISTERS): Likewise.
16202 (CALL_REALLY_USED_REGISTERS): Likewise.
16203 (REG_ALLOC_ORDER): Likewise.
16204 (enum reg_class): Likewise.
16205 (REG_CLASS_NAMES): Likewise.
16206 (REG_CLASS_CONTENTS): Likewise.
16207 (REGISTER_NAMES): Likewise.
16208 (ADDITIONAL_REGISTER_NAMES): Likewise.
16209 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
16210 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
16211 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
16212 * config/rs6000/htm.md: New file.
16213 * config/rs6000/htmintrin.h: New file.
16214 * config/rs6000/htmxlintrin.h: New file.
16216 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
16218 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16219 Define SYMBOL_TINY_GOT, update comment.
16220 * config/aarch64/aarch64.c
16221 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
16222 (aarch64_expand_mov_immediate): Likewise.
16223 (aarch64_print_operand): Likewise.
16224 (aarch64_classify_symbol): Likewise.
16225 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
16226 (ldr_got_tiny): Define.
16228 2013-07-13 Tobias Grosser <tobias@grosser.es>
16230 PR tree-optimization/54094
16231 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
16232 scheduling dimension for the parallelism check from the polyhedral
16233 information in the AST.
16234 * graphite-dependences.c (carries_deps): Do not assume the schedule is
16237 2013-07-13 Jason Merrill <jason@redhat.com>
16239 * print-tree.c (debug_vec_tree): Use debug_raw.
16240 (debug_raw (vec<tree, va_gc> &)): New.
16241 (debug_raw (vec<tree, va_gc> *)): New.
16242 * tree.h: Declare them.
16244 2013-07-13 Bin Cheng <bin.cheng@arm.com>
16246 * ifcvt.c (ifcvt_after_combine): New static variable.
16247 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
16249 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
16250 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
16251 rest_of_handle_if_after_reload): Pass new argument for if_convert.
16253 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
16255 * config/mips/mips.c (mips_expand_call): Remove empty statement.
16257 2013-07-12 Michael Matz <matz@suse.de>
16259 PR middle-end/55771
16260 * convert.c (convert_to_real): Reject non-float inner types.
16262 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
16264 * config/aarch64/aarch64-protos.h
16265 (aarch64_simd_immediate_valid_for_move): Remove.
16266 * config/aarch64/aarch64.c (simd_immediate_info): New member.
16267 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
16269 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
16271 2013-07-11 Steve Ellcey <sellcey@mips.com>
16273 * config/mips/mips.c (mips_conditional_register_usage): Do not
16274 use t[0-7] registers in MIPS16 mode when optimizing for size.
16276 2013-07-11 Sriraman Tallam <tmsriram@google.com>
16278 * config/i386/i386.c (dispatch_function_versions): Fix array
16279 indexing of function_version_info to match actual_versions.
16281 2013-07-11 Teresa Johnson <tejohnson@google.com>
16283 * vec.h (struct va_gc): Move release out-of-line.
16284 (va_gc::release): Call ggc_free on released vec.
16286 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16288 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
16289 Require GOT register as additional operand in UNSPEC.
16290 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
16291 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16292 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16293 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
16294 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
16295 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
16296 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
16298 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16301 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16302 name seen by assembler/linker rather if available.
16304 2013-07-11 Andreas Schwab <schwab@suse.de>
16306 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16308 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16310 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16312 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16314 * doc/tm.texi.in: Move hook documentation to ....
16315 * target.def: ... here.
16317 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16318 text on @hook line.
16319 * doc/tm.texi: Regenerate.
16321 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16324 * doc/invoke.texi: Document -Wconditionally-supported.
16326 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16329 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16332 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16335 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16336 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16337 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16338 Remove duplicate devices.
16339 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16340 * config/avr/t-multilib: Regenerate.
16341 * config/avr/avr-tables.opt: Regenerate.
16342 * doc/avr-mmcu.texi: Regenerate.
16344 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16347 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16349 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16351 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16352 the cost of MULT when optimizing for size.
16354 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16356 * config/cr16/cr16-protos.h: Don't include target.h.
16358 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16360 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16361 adjust register size for TDmode and TFmode for VSX registers.
16363 2013-07-08 Kai Tietz <ktietz@redhat.com>
16366 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16367 hook_bool_const_tree_true.
16369 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16371 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16372 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16373 * config/s390/s390.md: Define FPR*_REGNUM constants.
16374 Fix FPR2_REGNUM constant (18 -> 17).
16375 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16376 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16377 ("trunc<BFP:mode><DFP_ALL:mode>2")
16378 ("trunc<DFP_ALL:mode><BFP:mode>2")
16379 ("*extend<BFP:mode><DFP_ALL:mode>2")
16380 ("*extend<DFP_ALL:mode><BFP:mode>2")
16381 ("extend<BFP:mode><DFP_ALL:mode>2")
16382 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16385 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16387 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16389 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16391 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16392 and cfun_fpr_bit_p to cfun_fpr_save_p.
16393 (s390_frame_area, s390_register_info, s390_frame_info)
16394 (s390_emit_prologue, s390_emit_epilogue)
16395 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16397 * config/s390/s390.h: Define *_REGNUM macros for floating point
16400 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16402 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16404 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16406 PR rtl-optimization/57786
16407 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16408 and break out of the loop when it is set to false.
16410 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16413 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16414 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16415 (const_int 63)) 0)).
16416 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16417 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16418 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16420 PR rtl-optimization/57829
16421 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16422 mask bits outside of mode are just sign-extension from mode to HWI.
16424 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16426 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16427 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16428 adjust_address instead of change_address to keep info about alignment.
16429 (emit_strmov): Remove.
16430 (emit_memmov): New function.
16431 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16432 (expand_movmem_epilogue): Likewise and return updated rtx for
16434 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16435 destination and source.
16436 (decide_alignment): Refactor, handle vector_loop.
16437 (ix86_expand_movmem): Likewise.
16438 (ix86_expand_setmem): Likewise.
16439 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16441 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16443 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16444 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16446 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16448 * config/i386/sse.md (sse_movlhps): Change alternative 3
16449 of operand 2 to "m".
16451 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16454 * config/i386/sse.md (iptr): New mode attribute.
16455 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16456 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16457 (<sse>_vmmul<mode>3): Ditto.
16458 (<sse>_vmdiv<mode>3): Ditto.
16459 (sse_vmrcpv4sf2): Ditto.
16460 (<sse>_vmsqrt<mode>2): Ditto.
16461 (sse_vmrsqrtv4sf2): Ditto.
16462 (<sse>_vm<code><mode>3): Ditto.
16463 (avx_vmcmp<mode>3): Ditto.
16464 (<sse>_vmmaskcmp<mode>3): Ditto.
16465 (<sse>_comi): Ditto.
16466 (<sse>_ucomi): Ditto.
16467 (*xop_vmfrcz_<mode>): Ditto.
16468 (*fmai_fmadd_<mode>): Ditto.
16469 (*fmai_fmsub_<mode>): Ditto.
16470 (*fmai_fnmadd_<mode>): Ditto.
16471 (*fmai_fnmsub_<mode>): Ditto.
16472 (*fma4i_vmfmadd_<mode>): Ditto.
16473 (*fma4i_vmfmsub_<mode>): Ditto.
16474 (*fma4i_vmfnmadd_<mode>): Ditto.
16475 (*fma4i_vmfnmsub_<mode>): Ditto.
16476 (*xop_vmfrcz_<mode>): Ditto.
16477 (sse_cvtps2pi): Ditto.
16478 (sse_cvttps2pi): Ditto.
16479 (sse_cvtss2si): Ditto.
16480 (sse_cvtss2si_2): Ditto.
16481 (sse_cvtss2siq_2): Ditto.
16482 (sse_cvttss2si): Ditto.
16483 (sse_cvttss2siq): Ditto.
16484 (sse_cvtsd2si): Ditto.
16485 (sse_cvtsd2si_2): Ditto.
16486 (sse_cvtsd2siq_2): Ditto.
16487 (sse_cvttsd2si): Ditto.
16488 (sse_cvttsd2siq): Ditto.
16489 (sse_cvtsd2ss): Ditto.
16490 (sse_cvtss2sd): Ditto.
16491 (avx2_pbroadcast<mode>): Ditto.
16492 (avx2_pbroadcast<mode>_1): Ditto.
16493 (*avx_vperm_broadcast_v4sf): Ditto.
16495 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16496 (sse_movlhps): Ditto.
16497 (sse_storehps): Ditto.
16498 (sse_loadhps): Ditto.
16499 (sse_storelps): Ditto.
16500 (sse_loadlps): Ditto.
16501 (*vec_concatv4sf): Ditto.
16502 (*vec_interleave_highv2df): Ditto.
16503 (*vec_interleave_lowv2df): Ditto.
16504 (*vec_extractv2df_1_sse): Ditto.
16505 (*vec_extractv2df_0_sse): Ditto.
16506 (sse2_storelpd): Ditto.
16507 (sse2_loadlpd): Ditto.
16508 (sse2_movsd): Ditto.
16509 (*vec_concatv4si): Ditto.
16510 (vec_concatv2di): Ditto.
16512 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16513 for Intel asm dialect.
16514 (mmx_punpcklwd): Ditto.
16515 (mmx_punpckldq): Ditto.
16517 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16518 for intel assembler dialect.
16520 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16523 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16524 for BUILT_IN_C{LZ,LRSB}*.
16525 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16526 * tree-vrp.c (extract_range_basic): Handle
16527 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16528 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16529 fall thru to code calling set_value*.
16530 * builtins.c (expand_builtin): Remove *IMAX cases.
16531 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16532 if width is bigger than 2*HWI.
16534 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16536 PR rtl-optimization/55342
16537 * lra-int.h (lra_subreg_reload_pseudos): New.
16538 * lra.c: Add undoing optional reloads to the block diagram.
16539 (lra_subreg_reload_pseudos): New.
16540 (lra_optional_reload_pseudos): Change comments.
16541 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16542 lra_optional_reload_pseudos after undo transformations.
16543 * lra-assigns.c (pseudo_prefix_title): New.
16544 (lra_setup_reg_renumber): Use it.
16545 (spill_for): Ditto. Check subreg reload pseudos too.
16546 (assign_by_spills): Consider subreg reload pseudos too.
16547 * lra-constraints.c (simplify_operand_subreg): Use
16548 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16549 (curr_insn_transform): Recognize and do optional reloads.
16550 (undo_optional_reloads): New.
16551 (lra_undo_inheritance): Call undo_optional_reloads.
16553 2013-07-05 Thomas Quinot <quinot@adacore.com>
16555 * tree-complex.c (expand_complex_operations_1): Fix typo.
16557 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16559 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16560 (tune_params): New member 'const vec_costs'.
16561 * config/aarch64/aarch64.c (generic_vector_cost): New.
16562 (generic_tunings): New member 'generic_vector_cost'.
16563 (aarch64_builtin_vectorization_cost): New.
16564 (aarch64_add_stmt_cost): New.
16565 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16566 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16568 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16571 * config/i386/predicates.md (vsib_address_operand): Disallow
16572 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16574 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16576 PR middle-end/55030
16577 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16578 expand_builtin_setjmp_receiver.
16579 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16580 with NULL for the label parameter.
16581 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16582 the frame-pointer. Adjust comments.
16583 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16584 only if LABEL is non-NULL.
16586 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16588 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16589 (ARM_ABI_AAPCS64): Ditto.
16591 (ARM_DEFAULT_ABI): Ditto.
16593 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16595 * config/aarch64/aarch64-builtins.c
16596 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16597 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16599 * config/aarch64/aarch64-simd.md
16600 (aarch64_ld1<VALL:mode>): New.
16601 (aarch64_st1<VALL:mode>): Likewise.
16602 * config/aarch64/arm_neon.h
16603 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16605 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16607 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16609 (ix86_option_override_internal):Turn on all -mavx target flags by
16610 default as they are dependent on AVX anyway.
16612 2013-07-02 Cary Coutant <ccoutant@google.com>
16614 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16615 deterministic hash.
16616 (loc_checksum_ordered): Likewise.
16617 (hash_loc_operands): Remove inline keyword.
16619 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16621 PR tree-optimization/57741
16622 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16623 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16624 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16625 Allow REAL_CST step_exprs if flag_associative_math.
16626 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16628 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16630 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16632 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16634 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16636 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16638 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16640 (iorsi3_insn): Likewise.
16641 (arm_xorsi3): Likewise.
16643 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16645 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16647 (attribute "type"): ... here, and prefix with "wmmx_".
16648 (attribute "core_cycles"): Update for attribute changes.
16649 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16650 (tbcstv4hi): Likewise.
16651 (tbcstv2si): Likewise.
16652 (iwmmxt_iordi3): Likewise.
16653 (iwmmxt_xordi3): Likewise.
16654 (iwmmxt_anddi3): Likewise.
16655 (iwmmxt_nanddi3): Likewise.
16656 (iwmmxt_arm_movdi): Likewise.
16657 (iwmmxt_movsi_insn): Likewise.
16658 (mov<mode>_internal): Likewise.
16659 (and<mode>3_iwmmxt): Likewise.
16660 (ior<mode>3_iwmmxt): Likewise.
16661 (xor<mode>3_iwmmxt): Likewise.
16662 (add<mode>3_iwmmxt): Likewise.
16663 (ssaddv8qi3): Likewise.
16664 (ssaddv4hi3): Likewise.
16665 (ssaddv2si3): Likewise.
16666 (usaddv8qi3): Likewise.
16667 (usaddv4hi3): Likewise.
16668 (usaddv2si3): Likewise.
16669 (sub<mode>3_iwmmxt): Likewise.
16670 (sssubv8qi3): Likewise.
16671 (sssubv4hi3): Likewise.
16672 (sssubv2si3): Likewise.
16673 (ussubv8qi3): Likewise.
16674 (ussubv4hi3): Likewise.
16675 (ussubv2si3): Likewise.
16676 (mulv4hi3_iwmmxt): Likewise.
16677 (smulv4hi3_highpart): Likewise.
16678 (umulv4hi3_highpart): Likewise.
16679 (iwmmxt_wmacs): Likewise.
16680 (iwmmxt_wmacsz): Likewise.
16681 (iwmmxt_wmacu): Likewise.
16682 (iwmmxt_wmacuz): Likewise.
16683 (iwmmxt_clrdi): Likewise.
16684 (iwmmxt_clrv8qi): Likewise.
16685 (iwmmxt_clr4hi): Likewise.
16686 (iwmmxt_clr2si): Likewise.
16687 (iwmmxt_uavgrndv8qi3): Likewise.
16688 (iwmmxt_uavgrndv4hi3): Likewise.
16689 (iwmmxt_uavgv8qi3): Likewise.
16690 (iwmmxt_uavgv4hi3): Likewise.
16691 (iwmmxt_tinsrb): Likewise.
16692 (iwmmxt_tinsrh): Likewise.
16693 (iwmmxt_tinsrw): Likewise.
16694 (iwmmxt_textrmub): Likewise.
16695 (iwmmxt_textrmsb): Likewise.
16696 (iwmmxt_textrmuh): Likewise.
16697 (iwmmxt_textrmsh): Likewise.
16698 (iwmmxt_textrmw): Likewise.
16699 (iwmxxt_wshufh): Likewise.
16700 (eqv8qi3): Likewise.
16701 (eqv4hi3): Likewise.
16702 (eqv2si3): Likewise.
16703 (gtuv8qi3): Likewise.
16704 (gtuv4hi3): Likewise.
16705 (gtuv2si3): Likewise.
16706 (gtv8qi3): Likewise.
16707 (gtv4hi3): Likewise.
16708 (gtv2si3): Likewise.
16709 (smax<mode>3_iwmmxt): Likewise.
16710 (umax<mode>3_iwmmxt): Likewise.
16711 (smin<mode>3_iwmmxt): Likewise.
16712 (umin<mode>3_iwmmxt): Likewise.
16713 (iwmmxt_wpackhss): Likewise.
16714 (iwmmxt_wpackwss): Likewise.
16715 (iwmmxt_wpackdss): Likewise.
16716 (iwmmxt_wpackhus): Likewise.
16717 (iwmmxt_wpackwus): Likewise.
16718 (iwmmxt_wpackdus): Likewise.
16719 (iwmmxt_wunpckihb): Likewise.
16720 (iwmmxt_wunpckihh): Likewise.
16721 (iwmmxt_wunpckihw): Likewise.
16722 (iwmmxt_wunpckilb): Likewise.
16723 (iwmmxt_wunpckilh): Likewise.
16724 (iwmmxt_wunpckilw): Likewise.
16725 (iwmmxt_wunpckehub): Likewise.
16726 (iwmmxt_wunpckehuh): Likewise.
16727 (iwmmxt_wunpckehuw): Likewise.
16728 (iwmmxt_wunpckehsb): Likewise.
16729 (iwmmxt_wunpckehsh): Likewise.
16730 (iwmmxt_wunpckehsw): Likewise.
16731 (iwmmxt_wunpckelub): Likewise.
16732 (iwmmxt_wunpckeluh): Likewise.
16733 (iwmmxt_wunpckeluw): Likewise.
16734 (iwmmxt_wunpckelsb): Likewise.
16735 (iwmmxt_wunpckelsh): Likewise.
16736 (iwmmxt_wunpckelsw): Likewise.
16737 (ror<mode>3): Likewise.
16738 (ashr<mode>3_iwmmxt): Likewise.
16739 (lshr<mode>3_iwmmxt): Likewise.
16740 (ashl<mode>3_iwmmxt): Likewise.
16741 (ror<mode>3_di): Likewise.
16742 (ashr<mode>3_di): Likewise.
16743 (lshr<mode>3_di): Likewise.
16744 (ashl<mode>3_di): Likewise.
16745 (iwmmxt_wmadds): Likewise.
16746 (iwmmxt_wmaddu): Likewise.
16747 (iwmmxt_tmia): Likewise.
16748 (iwmmxt_tmiaph): Likewise.
16749 (iwmmxt_tmiabb): Likewise.
16750 (iwmmxt_tmiatb): Likewise.
16751 (iwmmxt_tmiabt): Likewise.
16752 (iwmmxt_tmiatt): Likewise.
16753 (iwmmxt_tmovmskb): Likewise.
16754 (iwmmxt_tmovmskh): Likewise.
16755 (iwmmxt_tmovmskw): Likewise.
16756 (iwmmxt_waccb): Likewise.
16757 (iwmmxt_wacch): Likewise.
16758 (iwmmxt_waccw): Likewise.
16759 (iwmmxt_waligni): Likewise.
16760 (iwmmxt_walignr): Likewise.
16761 (iwmmxt_walignr0): Likewise.
16762 (iwmmxt_walignr1): Likewise.
16763 (iwmmxt_walignr2): Likewise.
16764 (iwmmxt_walignr3): Likewise.
16765 (iwmmxt_wsadb): Likewise.
16766 (iwmmxt_wsadh): Likewise.
16767 (iwmmxt_wsadbz): Likewise.
16768 (iwmmxt_wsadhz): Likewise.
16769 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16770 (iwmmxt_wabsdiffb): Likewise.
16771 (iwmmxt_wabsdiffh): Likewise.
16772 (iwmmxt_wabsdiffw): Likewise.
16773 (iwmmxt_waddsubhx): Likewise
16774 (iwmmxt_wsubaddhx): Likewise.
16775 (addc<mode>3): Likewise.
16776 (iwmmxt_avg4): Likewise.
16777 (iwmmxt_avg4r): Likewise.
16778 (iwmmxt_wmaddsx): Likewise.
16779 (iwmmxt_wmaddux): Likewise.
16780 (iwmmxt_wmaddsn): Likewise.
16781 (iwmmxt_wmaddun): Likewise.
16782 (iwmmxt_wmulwsm): Likewise.
16783 (iwmmxt_wmulwum): Likewise.
16784 (iwmmxt_wmulsmr): Likewise.
16785 (iwmmxt_wmulumr): Likewise.
16786 (iwmmxt_wmulwsmr): Likewise.
16787 (iwmmxt_wmulwumr): Likewise.
16788 (iwmmxt_wmulwl): Likewise.
16789 (iwmmxt_wqmulm): Likewise.
16790 (iwmmxt_wqmulwm): Likewise.
16791 (iwmmxt_wqmulmr): Likewise.
16792 (iwmmxt_wqmulwmr): Likewise.
16793 (iwmmxt_waddbhusm): Likewise.
16794 (iwmmxt_waddbhusl): Likewise.
16795 (iwmmxt_wqmiabb): Likewise.
16796 (iwmmxt_wqmiabt): Likewise.
16797 (iwmmxt_wqmiatb): Likewise.
16798 (iwmmxt_wqmiatt): Likewise.
16799 (iwmmxt_wqmiabbn): Likewise.
16800 (iwmmxt_wqmiabtn): Likewise.
16801 (iwmmxt_wqmiatbn): Likewise.
16802 (iwmmxt_wqmiattn): Likewise.
16803 (iwmmxt_wmiabb): Likewise.
16804 (iwmmxt_wmiabt): Likewise.
16805 (iwmmxt_wmiatb): Likewise.
16806 (iwmmxt_wmiatt): Likewise.
16807 (iwmmxt_wmiabbn): Likewise.
16808 (iwmmxt_wmiabtn): Likewise.
16809 (iwmmxt_wmiatbn): Likewise.
16810 (iwmmxt_wmiattn): Likewise.
16811 (iwmmxt_wmiawbb): Likewise.
16812 (iwmmxt_wmiawbt): Likewise.
16813 (iwmmxt_wmiawtb): Likewise.
16814 (iwmmxt_wmiawtt): Likewise.
16815 (iwmmxt_wmiawbbn): Likewise.
16816 (iwmmxt_wmiawbtn): Likewise.
16817 (iwmmxt_wmiawtbn): Likewise.
16818 (iwmmxt_wmiawttn): Likewise.
16819 (iwmmxt_wmerge): Likewise.
16820 (iwmmxt_tandc<mode>3): Likewise.
16821 (iwmmxt_torc<mode>3): Likewise.
16822 (iwmmxt_torvsc<mode>3): Likewise.
16823 (iwmmxt_textrc<mode>3): Likewise.
16824 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16825 (wmmxt_pack): Likewise.
16826 (wmmxt_mult_c1): Likewise.
16827 (wmmxt_mult_c2): Likewise.
16828 (wmmxt_alu_c1): Likewise.
16829 (wmmxt_alu_c2): Likewise.
16830 (wmmxt_alu_c3): Likewise.
16831 (wmmxt_transfer_c1): Likewise.
16832 (wmmxt_transfer_c2): Likewise.
16833 (wmmxt_transfer_c3): Likewise.
16834 (marvell_f_iwmmxt_wstr): Likewise.
16835 (marvell_f_iwmmxt_wldr): Likewise.
16837 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16839 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16841 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16844 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16845 * lra-constraints.c (need_for_split_p): Check call used hard regs
16846 living through calls.
16848 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16849 call used regs for call insn.
16851 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16854 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16855 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16857 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16859 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16861 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16863 * lra-constraints.c (need_for_split_p): Check call used hard regs
16864 living through calls.
16866 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16869 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16870 to tie with any other modes. Eliminate Altivec vector mode tests,
16871 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16872 code, to return 0 if testing MODE2 for a condition, if we've
16873 already tested MODE1 for the same condition.
16875 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16877 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16880 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16882 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16883 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16885 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16887 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16889 (aarch64_symbolic_constant_p): Remove.
16890 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16891 static. Fix line length and white space.
16892 (aarch64_symbolic_constant_p): Remove.
16893 * config/aarch64/predicates.md (aarch64_valid_symref):
16894 Use aarch64_classify_symbol_expression.
16896 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16898 * config/arm/constraints.md (Ts): New constraint.
16899 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16901 (compare_scc): Use "Ts" constraint for operand 0.
16902 (ior_scc_scc): Likewise.
16903 (and_scc_scc): Likewise.
16904 (and_scc_scc_nodom): Likewise.
16905 (ior_scc_scc_cmp): Likewise for operand 7.
16906 (and_scc_scc_cmp): Likewise.
16907 * config/arm/thumb2.md (thumb2_movsi_insn):
16908 Add alternatives for 16-bit encodings.
16909 (thumb2_movhi_insn): Likewise.
16910 (thumb2_movsicc_insn): Likewise.
16911 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16912 (thumb2_negscc): Use "Ts" constraint.
16913 Move mvn instruction outside cond_exec block.
16914 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16915 for 16-bit encodings.
16917 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16919 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16921 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16922 (mulsi3subsi): Likewise.
16923 (mulsidi3adddi): Likewise.
16924 (mulsidi3_v6): Likewise.
16925 (umulsidi3_v6): Likewise.
16926 (umulsidi3adddi_v6): Likewise.
16927 (smulsi3_highpart_v6): Likewise.
16928 (umulsi3_highpart_v6): Likewise.
16929 (mulhisi3tb): Likewise.
16930 (mulhisi3bt): Likewise.
16931 (mulhisi3tt): Likewise.
16932 (maddhisi4): Likewise.
16933 (maddhisi4tb): Likewise.
16934 (maddhisi4tt): Likewise.
16935 (maddhidi4): Likewise.
16936 (maddhidi4tb): Likewise.
16937 (maddhidi4tt): Likewise.
16938 (zeroextractsi_compare0_scratch): Likewise.
16939 (insv_zero): Likewise.
16940 (insv_t2): Likewise.
16941 (anddi_notzesidi_di): Likewise.
16942 (anddi_notsesidi_di): Likewise.
16943 (andsi_notsi_si): Likewise.
16944 (iordi_zesidi_di): Likewise.
16945 (xordi_zesidi_di): Likewise.
16946 (andsi_iorsi3_notsi): Likewise.
16947 (smax_0): Likewise.
16948 (smax_m1): Likewise.
16949 (smin_0): Likewise.
16950 (not_shiftsi): Likewise.
16951 (unaligned_loadsi): Likewise.
16952 (unaligned_loadhis): Likewise.
16953 (unaligned_loadhiu): Likewise.
16954 (unaligned_storesi): Likewise.
16955 (unaligned_storehi): Likewise.
16956 (extv_reg): Likewise.
16957 (extzv_t2): Likewise.
16958 (divsi3): Likewise.
16959 (udivsi3): Likewise.
16960 (arm_zero_extendhisi2addsi): Likewise.
16961 (arm_zero_extendqisi2addsi): Likewise.
16962 (compareqi_eq0): Likewise.
16963 (arm_extendhisi2_v6): Likewise.
16964 (arm_extendqisi2addsi): Likewise.
16965 (arm_movt): Likewise.
16966 (thumb2_ldrd): Likewise.
16967 (thumb2_ldrd_base): Likewise.
16968 (thumb2_ldrd_base_neg): Likewise.
16969 (thumb2_strd): Likewise.
16970 (thumb2_strd_base): Likewise.
16971 (thumb2_strd_base_neg): Likewise.
16972 (arm_negsi2): Add alternative for 16-bit encoding.
16973 (arm_one_cmplsi2): Likewise.
16975 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16977 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16978 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16979 (movdfcc): Likewise.
16980 * config/arm/vfp.md (*thumb2_movsf_vfp):
16981 Disable predication for arm_restrict_it.
16982 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16983 (*thumb2_movdfcc_vfp): Likewise.
16984 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16985 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16986 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16987 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16988 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16989 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16990 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16991 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16992 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16993 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16994 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16995 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16996 Disable predication for arm_restrict_it.
16998 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
17000 * config/i386/bmiintrin.h (_bextr_u32): New.
17001 (_bextr_u64): Ditto.
17003 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
17005 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
17006 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
17007 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
17008 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
17009 * config/mips/n32-elf.h: ...this new file.
17011 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
17014 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
17015 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
17017 2013-06-27 Catherine Moore <clm@codesourcery.com>
17019 * config/mips/mips-tables.opt: Regenerate.
17020 * config/mips/mips-cpus.def: Add m14ke and m14kec.
17021 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
17022 * doc/invoke.texi: Add -m14kc.
17024 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17027 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
17028 constraints of operand 1 and 2.
17031 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
17032 to match RTL canonicalization. Swap predicates and
17033 constraints of operand 1 and 2.
17035 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
17037 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
17038 Process OP_INOUT regs for splitting too.
17040 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17042 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
17043 decl before the loop, initialize to NULL.
17044 (vectorizable_load): Initialize ptr_incr to NULL.
17046 2013-06-27 Martin Jambor <mjambor@suse.cz>
17049 * ipa-ref.h (ipa_maybe_record_reference): Declare.
17050 * ipa-ref.c (ipa_maybe_record_reference): New function.
17051 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
17052 * ipa-cp.c (create_specialized_node): Record potential references from
17054 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
17056 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17058 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
17059 parameter 'mode' of type 'enum machine_mode mode'; change to pass
17060 'mode' to force_reg.
17061 (aarch64_add_offset): Update calls to aarch64_force_temporary.
17062 (aarch64_expand_mov_immediate): Likewise.
17064 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17066 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
17067 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
17069 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17071 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
17072 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17073 (struct machine_function): Add tbegin_p.
17074 (s390_canonicalize_comparison): Fold CC mode compares to
17075 conditional jump if possible.
17076 (s390_emit_jump): Return the emitted jump.
17077 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
17078 Handle CCRAWmode compares.
17079 (s390_option_override): Default to -mhtm if available.
17080 (s390_reg_clobbered_rtx): Handle floating point regs as well.
17081 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
17082 FPRs instead of df_regs_ever_live_p.
17083 (s390_optimize_nonescaping_tx): New function.
17084 (s390_init_frame_layout): Extend clobbered_regs array to cover
17086 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
17087 (s390_expand_tbegin): New function.
17088 (enum s390_builtin): New enum definition.
17089 (code_for_builtin): New array definition.
17090 (s390_init_builtins): New function.
17091 (s390_expand_builtin): New function.
17092 (TARGET_INIT_BUILTINS): Define.
17093 (TARGET_EXPAND_BUILTIN): Define.
17094 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
17095 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
17096 (s390_alc_comparison): Likewise.
17097 * config/s390/s390-modes.def: Add CCRAWmode.
17098 * config/s390/s390.h (processor_flags): Add PF_TX.
17099 (TARGET_CPU_HTM): Define macro.
17100 (TARGET_HTM): Define macro.
17101 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
17102 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
17103 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17104 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
17105 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
17106 (TBEGIN_MASK, TBEGINC_MASK): New constants.
17107 ("*cc_to_int"): Move up.
17108 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
17109 constants other than 0.
17110 ("*ccraw_to_int"): New insn and splitter definition.
17111 ("tbegin", "tbegin_nofloat", "tbegin_retry")
17112 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
17113 ("tx_assist"): New expander.
17114 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
17115 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
17116 * config/s390/s390.opt: Add -mhtm option.
17117 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
17118 * config/s390/htmxlintrin.h: New file.
17119 * config/s390/htmintrin.h: New file.
17120 * config/s390/s390intrin.h: New file.
17121 * doc/extend.texi: Document htm builtins.
17122 * config.gcc: Add the new header files to extra_headers.
17124 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
17126 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
17127 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
17129 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
17130 Pat Haugen <pthaugen@us.ibm.com>
17131 Peter Bergner <bergner@vnet.ibm.com>
17133 * config/rs6000/power8.md: New.
17134 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
17135 setting for power8 entry.
17136 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
17137 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
17138 test for Power4/Power5 only.
17139 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
17141 (force_new_group): Adjust comment.
17142 * config/rs6000/rs6000.md: Include power8.md.
17144 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
17146 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
17147 * config/arm/arm-protos.h (arm_max_conditional_execute): New
17149 (tune_params): Update comment.
17150 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
17151 (arm_max_conditional_execute): New function.
17152 (thumb2_final_prescan_insn): Use max_insn_skipped and
17153 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
17155 2013-06-25 Jakub Jelinek <jakub@redhat.com>
17157 PR tree-optimization/57705
17158 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
17159 SSA_NAME step, provided that it is not defined inside the loop.
17160 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
17161 (get_initial_def_for_induction): Handle SSA_NAME IV step.
17163 2013-06-25 Martin Jambor <mjambor@suse.cz>
17165 PR middle-end/57670
17166 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
17167 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
17169 (ipa_note_param_call): Initialize member_ptr flag.
17170 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
17171 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
17172 (ipa_write_indirect_edge_info): Stream member_ptr flag.
17173 (ipa_read_indirect_edge_info): Likewise.
17175 2013-06-25 Richard Biener <rguenther@suse.de>
17177 PR middle-end/56977
17178 * passes.c (init_optimization_passes): Move pass_fold_builtins
17179 and pass_dce earlier with -Og.
17181 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
17183 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
17184 <BIT_FIELD_REF>: Remove trailing TAB.
17185 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
17188 2013-06-24 Martin Jambor <mjambor@suse.cz>
17190 PR tree-optimization/57358
17191 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
17192 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
17193 (ipa_analyze_params_uses): Generate pessimistic info when true.
17195 2013-06-24 Martin Jambor <mjambor@suse.cz>
17197 PR tree-optimization/57539
17198 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
17199 global.inlined_to of the new node to it. All callers changed.
17200 * ipa-inline-transform.c (clone_inlined_nodes): New variable
17201 inlining_into, pass it to cgraph_clone_node.
17202 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
17203 ipa_free_edge_args_substructures.
17204 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
17205 rdesc linked list. Do not assert rdesc edges have inlined caller.
17206 Assert we have found an rdesc in the rdesc list.
17208 2013-06-24 Richard Biener <rguenther@suse.de>
17210 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
17211 (pointer_set_lookup): Declare.
17212 (class pointer_map): New template class implementing a
17213 generic pointer to T map.
17214 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
17215 pointer_map<T>::contains, pointer_map<T>::insert,
17216 pointer_map<T>::traverse): New functions.
17217 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
17218 (pointer_set_lookup): New function.
17219 (pointer_set_contains): Use pointer_set_lookup.
17220 (pointer_set_insert): Likewise.
17221 (insert_aux): Remove.
17222 (struct pointer_map_t): Embed a pointer_set_t.
17223 (pointer_map_create): Adjust.
17224 (pointer_map_destroy): Likewise.
17225 (pointer_map_contains): Likewise.
17226 (pointer_map_insert): Likewise.
17227 (pointer_map_traverse): Likewise.
17228 * tree-streamer.h (struct streamer_tree_cache_d): Use a
17229 pointer_map<unsigned> instead of a pointer_map_t.
17230 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
17231 (streamer_tree_cache_lookup): Likewise.
17232 (streamer_tree_cache_create): Likewise.
17233 (streamer_tree_cache_delete): Likewise.
17234 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
17235 pointer_map<unsigned> instead of a pointer_map_t.
17236 (lto_init_tree_ref_encoder): Adjust.
17237 (lto_destroy_tree_ref_encoder): Likewise.
17238 * lto-section-out.c (lto_output_decl_index): Likewise.
17239 (lto_record_function_out_decl_state): Likewise.
17240 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
17242 2013-06-24 Richard Biener <rguenther@suse.de>
17244 PR tree-optimization/57488
17245 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
17247 2013-06-24 Alan Modra <amodra@gmail.com>
17249 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
17250 (gen_easy_altivec_constant): Likewise.
17251 * config/rs6000/predicates.md (easy_vector_constant_add_self,
17252 easy_vector_constant_msb): Likewise.
17254 2013-06-23 Jakub Jelinek <jakub@redhat.com>
17257 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
17258 add missing return true.
17260 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
17263 * config/sh/predicates.md (general_extend_operand): Invoke
17264 general_movsrc_operand for memory operands.
17265 (general_movsrc_operand): Allow reg+reg addressing, do not use
17266 general_operand for memory operands.
17268 2013-06-23 Sriraman Tallam <tmsriram@google.com>
17270 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
17271 when current target options does not apply.
17272 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
17273 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
17274 * config/i386/bmiintrin.h: Pass appropriate target
17275 attributes to header.
17276 * config/i386/mmintrin.h: Ditto.
17277 * config/i386/nmmintrin.h: Ditto.
17278 * config/i386/avx2intrin.h: Ditto.
17279 * config/i386/fxsrintrin.h: Ditto.
17280 * config/i386/tbmintrin.h: Ditto.
17281 * config/i386/xsaveintrin.h: Ditto.
17282 * config/i386/f16cintrin.h: Ditto.
17283 * config/i386/xtestintrin.h: Ditto.
17284 * config/i386/xsaveoptintrin.h: Ditto.
17285 * config/i386/bmi2intrin.h: Ditto.
17286 * config/i386/lzcntintrin.h: Ditto.
17287 * config/i386/smmintrin.h: Ditto.
17288 * config/i386/wmmintrin.h: Ditto.
17289 * config/i386/x86intrin.h: Remove all header include guards.
17290 * config/i386/prfchwintrin.h: Ditto.
17291 * config/i386/pmmintrin.h: Ditto.
17292 * config/i386/tmmintrin.h: Ditto.
17293 * config/i386/xmmintrin.h: Ditto.
17294 * config/i386/popcntintrin.h: Ditto.
17295 * config/i386/rdseedintrin.h: Ditto.
17296 * config/i386/ammintrin.h: Ditto.
17297 * config/i386/emmintrin.h: Ditto.
17298 * config/i386/immintrin.h: Remove all header include guards.
17299 * config/i386/fma4intrin.h: Ditto.
17300 * config/i386/lwpintrin.h: Ditto.
17301 * config/i386/xopintrin.h: Ditto.
17302 * config/i386/ia32intrin.h: Ditto.
17303 * config/i386/avxintrin.h: Ditto.
17304 * config/i386/rtmintrin.h: Ditto.
17305 * config/i386/fmaintrin.h: Ditto.
17306 * config/i386/mm3dnow.h: Ditto.
17308 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17310 * common/config/i386/i386-common.c: Handle LZCNT.
17312 2013-06-22 Andi Kleen <ak@linux.intel.com>
17314 * doc/extend.texi: Use __atomic_store_n instead of
17315 __atomic_store in HLE example.
17317 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17319 * config/sh/sh.c: Remove <cstdlib> workaround.
17321 2013-06-21 Andi Kleen <ak@linux.intel.com>
17323 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17325 2013-06-21 Andi Kleen <ak@linux.intel.com>
17327 * doc/extend.texi: Document that __atomic_clear and
17328 __atomic_test_and_set should only be used with bool.
17330 2013-06-20 Jan Hubicka <jh@suse.cz>
17332 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17333 types_same_for_odr.
17334 * tree.c (decls_same_for_odr): New function.
17335 (same_for_edr): New function.
17336 (types_same_for_odr): New function.
17337 (get_binfo_at_offset): Use it.
17338 * tree.h (types_same_for_odr): Declare.
17340 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17341 Jason Merrill <jason@redhat.com>
17343 * system.h: Include <cstdlib> as well as <stdlib.h>.
17345 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17348 * config/i386/i386.c (construct_container): Report error if
17349 long double is used with disabled x87 float returns.
17351 2013-06-20 Jan Hubicka <jh@suse.cz>
17353 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17355 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17357 PR rtl-optimization/57425
17358 PR rtl-optimization/57569
17359 * alias.c (write_dependence_p): Remove parameters mem_mode and
17360 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17361 Changed all callers.
17362 (canon_anti_dependence): Get comments and semantics in sync.
17363 Add parameter mem_canonicalized. Changed all callers.
17364 * rtl.h (canon_anti_dependence): Update prototype.
17366 2013-06-20 Richard Biener <rguenther@suse.de>
17368 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17369 byte case, inline streamer_read_uchar and defer section
17372 2013-06-20 Richard Biener <rguenther@suse.de>
17374 PR tree-optimization/57584
17375 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17376 SSA names into the expanded expression that take part in
17377 abnormal coalescing.
17379 2013-06-19 Sharad Singhai <singhai@google.com>
17381 * gcov.c (print_usage): Handle new option.
17382 (process_args): Ditto.
17383 (get_gcov_intermediate_filename): New function.
17384 (output_intermediate_file): New function.
17385 (output_gcov_file): New function
17386 (generate_results): Handle new option.
17387 (release_function): Relase demangled name.
17388 (read_graph_file): Handle demangled name.
17389 (output_lines): Ditto.
17390 * doc/gcov.texi: Document gcov intermediate format.
17392 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17395 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17396 (lra_emit_add): Use the functions. Add comment about Y as an
17399 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17402 * collect2.c (collect_atexit): New.
17403 (collect_exit): Delete.
17404 (main): Register collect_atexit with atexit.
17405 (collect_wait): Change collect_exit to exit.
17407 * collect2.h (collect_exit): Delete.
17408 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17410 2013-06-19 Wei Mi <wmi@google.com>
17412 PR rtl-optimization/57518
17413 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17414 if regno is used in paradoxical subreg.
17415 (update_equiv_regs): Check pdx_subregs[regno] before
17416 set a reg to be equivalent with a mem.
17418 2013-06-19 Matthias Klose <doko@ubuntu.com>
17421 * file-find.h (find_a_file): Add a mode parameter.
17422 * file-find.c (find_a_file): Likewise.
17423 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17424 with X_OK for the executables.
17425 * collect2.c (main): Call find_a_file with X_OK.
17427 2013-06-19 Steve Ellcey <sellcey@mips.com>
17430 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17431 Use NEXT_INSN instead of next_real_insn.
17433 2013-06-19 Jan Hubicka <jh@suse.cz>
17435 * cgraph.h (const_value_known_p): Replace by ...
17436 (ctor_for_folding): .. this one.
17437 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17438 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17439 * expr.c (expand_expr_real_1): Likewise.
17440 (string_constant): Likewise.
17441 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17442 * ipa.c (process_references): Likewise.
17443 (symtab_remove_unreachable_nodes): Likewise.
17444 * ipa-inline-analysis.c (param_change_prob): Likewise.
17445 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17446 (get_base_constructor): Likwise.
17447 * varpool.c (varpool_remove_node): Likewise.
17448 (varpool_remove_initializer): LIkewise.
17449 (dump_varpool_node): LIkwise.
17450 (const_value_known_p): Rewrite to ...
17451 (ctor_for_folding): ... this one.
17453 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17456 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17457 PERSONALITY in $PATH derived prefixes.
17459 2013-06-19 Jeff Law <law@redhat.com>
17461 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17464 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17465 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17468 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17470 * config/arm/vfp.md: Move VFP instruction classification documentation
17472 * config/arm/arm.md: ... here. Update instruction classification
17475 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17477 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17478 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17479 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17481 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17484 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17485 with NEXT_INSN. Use tablejump_p to check for jump table data
17488 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17491 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17492 that now in C++ the value is correct per the C++ standards.
17494 2013-06-19 Richard Biener <rguenther@suse.de>
17496 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17497 for global context.
17499 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17502 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17505 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17506 with next_active_insn.
17508 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17510 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17511 functions are inlined during failures to flag an error.
17512 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17513 in early inline pass.
17515 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17517 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17520 2013-06-18 Julian Brown <julian@codesourcery.com>
17522 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17523 Permit virtual register pre-reload if !strict.
17524 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17526 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17528 * config/arm/neon.md (movmisalign<mode>): Use
17529 neon_perm_struct_or_reg_operand instead of
17530 neon_struct_or_register_operand.
17531 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17532 neon_permissive_struct_operand instead of neon_struct_operand.
17533 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17534 neon_vector_mem_operand.
17535 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17536 neon_vector_mem_operand.
17537 (neon_permissive_struct_operand): New.
17538 (neon_struct_or_register_operand): Rename to...
17539 (neon_perm_struct_or_reg_operand): This. Adjust call to
17540 neon_vector_mem_operand.
17542 2013-06-18 Richard Biener <rguenther@suse.de>
17544 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17545 * lto-streamer.h: Include pointer-set.h.
17546 (struct lto_decl_slot): Remove.
17547 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17548 Remove next_index entry.
17549 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17550 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17551 (lto_init_tree_ref_encoder): Adjust.
17552 (lto_destroy_tree_ref_encoder): Likewise.
17553 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17554 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17555 (lto_output_decl_index): Adjust.
17556 (lto_new_out_decl_state): Likewise.
17557 (lto_record_function_out_decl_state): Likewise.
17558 * lto-streamer-out.c (copy_function): Likewise.
17560 2013-06-18 Richard Biener <rguenther@suse.de>
17562 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17563 * cgraphunit.c: Include cfgloop.h.
17564 (init_lowered_empty_function): Initialize the loop tree.
17565 (assemble_thunk): Insert new BBs into loops.
17567 2013-06-18 Richard Biener <rguenther@suse.de>
17569 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17570 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17571 the map from cache entry to cache index optional.
17572 (streamer_tree_cache_replace_tree): Adjust accordingly.
17573 (streamer_tree_cache_append): Likewise.
17574 (streamer_tree_cache_delete): Likewise.
17575 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17576 streamer cache map from cache entry to cache index.
17577 * lto-streamer-out.c (create_output_block): Adjust.
17579 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17581 * config/arm/arm.md (attribute "insn"): Move multiplication and
17582 division attributes to...
17583 (attribute "type"): ... here. Remove mult.
17584 (attribute "mul32"): New attribute.
17585 (attribute "mul64"): Add umaal.
17586 (*arm_mulsi3): Update attributes.
17587 (*arm_mulsi3_v6): Likewise.
17588 (*thumb_mulsi3): Likewise.
17589 (*thumb_mulsi3_v6): Likewise.
17590 (*mulsi3_compare0): Likewise.
17591 (*mulsi3_compare0_v6): Likewise.
17592 (*mulsi_compare0_scratch): Likewise.
17593 (*mulsi_compare0_scratch_v6): Likewise.
17594 (*mulsi3addsi): Likewise.
17595 (*mulsi3addsi_v6): Likewise.
17596 (*mulsi3addsi_compare0): Likewise.
17597 (*mulsi3addsi_compare0_v6): Likewise.
17598 (*mulsi3addsi_compare0_scratch): Likewise.
17599 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17600 (*mulsi3subsi): Likewise.
17601 (*mulsidi3adddi): Likewise.
17602 (*mulsi3addsi_v6): Likewise.
17603 (*mulsidi3adddi_v6): Likewise.
17604 (*mulsidi3_nov6): Likewise.
17605 (*mulsidi3_v6): Likewise.
17606 (*umulsidi3_nov6): Likewise.
17607 (*umulsidi3_v6): Likewise.
17608 (*umulsidi3adddi): Likewise.
17609 (*umulsidi3adddi_v6): Likewise.
17610 (*smulsi3_highpart_nov6): Likewise.
17611 (*smulsi3_highpart_v6): Likewise.
17612 (*umulsi3_highpart_nov6): Likewise.
17613 (*umulsi3_highpart_v6): Likewise.
17614 (mulhisi3): Likewise.
17615 (*mulhisi3tb): Likewise.
17616 (*mulhisi3bt): Likewise.
17617 (*mulhisi3tt): Likewise.
17618 (maddhisi4): Likewise.
17619 (*maddhisi4tb): Likewise.
17620 (*maddhisi4tt): Likewise.
17621 (maddhidi4): Likewise.
17622 (*maddhidi4tb): Likewise.
17623 (*maddhidi4tt): Likewise.
17624 (divsi3): Likewise.
17625 (udivsi3): Likewise.
17626 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17627 (thumb2_mulsi_short_compare0): Likewise.
17628 (thumb2_mulsi_short_compare0_scratch): Likewise.
17629 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17630 (1020mult2): Likewise.
17631 (1020mult3): Likewise.
17632 (1020mult4): Likewise.
17633 (1020mult5): Likewise.
17634 (1020mult6): Likewise.
17635 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17637 (cortex_a15_mult64): Likewise.
17638 (cortex_a15_sdiv): Likewise.
17639 (cortex_a15_udiv): Likewise.
17640 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17646 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17647 (pj4_ir_div): Likewise.
17648 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17649 (11_mult2): Likewise.
17650 (11_mult3): Likewise.
17651 (11_mult4): Likewise.
17652 (11_mult5): Likewise.
17653 (11_mult6): Likewise.
17654 (11_mult7): Likewise.
17655 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17656 (cortex_a8_mla): Likewise.
17657 (cortex_a8_mull): Likewise.
17658 (cortex_a8_smulwy): Likewise.
17659 (cortex_a8_smlald): Likewise.
17660 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17661 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17662 (cortex_r4_mul_3): Likewise.
17663 (cortex_r4_mla_4): Likewise.
17664 (cortex_r4_mla_3): Likewise.
17665 (cortex_r4_smlald): Likewise.
17666 (cortex_r4_mull): Likewise.
17667 (cortex_r4_sdiv): Likewise.
17668 (cortex_r4_udiv): Likewise.
17669 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17670 (cortex_a7_idiv): Likewise.
17671 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17672 (9_mult2): Likewise.
17673 (9_mult3): Likewise.
17674 (9_mult4): Likewise.
17675 (9_mult5): Likewise.
17676 (9_mult6): Likewise.
17677 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17678 (cortex_a53_sdiv): Likewise.
17679 (cortex_a53_udiv): Likewise.
17680 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17681 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17682 (mp626_mult2): Likewise.
17683 (mp626_mult3): Likewise.
17684 (mp626_mult4): Likewise.
17685 * config/arm/fa526.md (526_mult1): Update attribute change.
17686 (526_mult2): Likewise.
17687 * config/arm/arm-generic.md (mult): Update attribute change.
17688 (mult_ldsched_strongarm): Likewise.
17689 (mult_ldsched): Likewise.
17690 (multi_cycle): Likewise.
17691 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17692 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17693 (606te_mult2): Likewise.
17694 (606te_mult3): Likewise.
17695 (606te_mult4): Likewise.
17696 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17697 (cortex_a9_mac16): Likewise.
17698 (cortex_a9_multiply): Likewise.
17699 (cortex_a9_mac): Likewise.
17700 (cortex_a9_multiply_long): Likewise.
17701 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17702 (626te_mult2): Likewise.
17703 (626te_mult3): Likewise.
17704 (626te_mult4): Likewise.
17706 2013-06-18 Richard Biener <rguenther@suse.de>
17709 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17711 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17714 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17715 with next_active_insn.
17717 2013-06-18 Alan Modra <amodra@gmail.com>
17719 * config/rs6000/rs6000.h (enum data_align): New.
17720 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17721 (DATA_ABI_ALIGNMENT): Define.
17722 (CONSTANT_ALIGNMENT): Correct comment.
17723 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17724 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17726 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17728 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17729 ATTRIBUTE_UNUSED marking.
17731 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17733 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17734 alternative and update.
17735 (aarch64_dup_lanedi): Delete.
17736 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17737 * config/aarch64/aarch64-simd-builtins.def: Update.
17739 2013-06-17 Richard Biener <rguenther@suse.de>
17741 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17742 (lto_input_scc): Declare.
17743 (lto_input_tree_1): Likewise.
17744 (struct lto_stats_d): Add num_tree_bodies_output and
17745 num_pickle_refs_output.
17746 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17747 (lto_read_tree_1): Split out from ...
17748 (lto_read_tree): ... this.
17749 (lto_input_scc): New function.
17750 (lto_input_tree_1): Split out from ...
17751 (lto_input_tree): ... this. Handle LTO_tree_scc.
17752 (lto_data_in_create): Create the streamer cache without hashes.
17753 * lto-streamer-out.c (create_output_block): Create the streamer
17754 cache with hashes when not doing WPA.
17755 (lto_write_tree_1): Split out from ...
17756 (lto_write_tree): ... this.
17757 (get_symbol_initial_value): New function.
17758 (lto_output_tree_1): Split out from ...
17759 (lto_output_tree): ... this. Write trees as series of SCCs
17760 using a DFS walk via DFS_write_tree.
17761 (struct sccs, struct scc_entry): New types.
17762 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17763 (DFS_write_tree_body): New function.
17764 (DFS_write_tree): Likewise.
17765 (hash_tree): Likewise.
17766 (scc_entry_compare): Likewise.
17767 (hash_scc): Likewise.
17768 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17769 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17770 TREE_CHAIN as regular reference.
17771 (streamer_read_integer_cst): Remove.
17772 (streamer_get_pickled_tree): Adjust.
17773 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17774 of DECL_EXTERNALs in BLOCK_VARS for now.
17775 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17777 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17778 Add hash value argument and record that if hashes are recorded
17780 (streamer_tree_cache_insert_1): Adjust.
17781 (streamer_tree_cache_insert): Likewise.
17782 (streamer_tree_cache_insert_at): Rename to ...
17783 (streamer_tree_cache_replace_tree): ... this and adjust.
17784 (streamer_tree_cache_append): Adjust.
17785 (record_common_node): Likewise.
17786 (streamer_tree_cache_create): Add argument whether to
17787 record hash values together with trees.
17788 (streamer_tree_cache_delete): Adjust.
17789 * tree-streamer.h (struct streamer_tree_cache_d): Add
17791 (streamer_read_integer_cst): Remove.
17792 (streamer_tree_cache_insert): Adjust.
17793 (streamer_tree_cache_append): Likewise.
17794 (streamer_tree_cache_insert_at): Rename to ...
17795 (streamer_tree_cache_replace_tree): ... this and adjust.
17796 (streamer_tree_cache_create): Add argument whether to record hashes.
17797 (streamer_tree_cache_get): Rename to ...
17798 (streamer_tree_cache_get_tree): ... this.
17799 (streamer_tree_cache_get_hash): New function.
17800 * tree.c (cache_integer_cst): New function.
17801 * tree.h (cache_integer_cst): Declare.
17802 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17803 * lto-symtab.c (lto_varpool_replace_node): Only release
17804 DECL_INITIAL of non-prevailing decls.
17805 * varpool.c (varpool_remove_initializer): Do not release
17806 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17808 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17810 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17811 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17812 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17813 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17814 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17815 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17816 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17817 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17818 instead of TARGET_64BIT.
17819 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17820 Require ISA_HAS_<D>DIV.
17822 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17824 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17825 (mips*-*-linux*): Move default with_llsc setting to where other
17827 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17829 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17830 Likewise. Remove default with_tune setting. Move default float
17831 setting to its own block. Handle with_llsc in the same block as above.
17833 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17835 PR rtl-optimization/57425
17836 PR rtl-optimization/57569
17837 * alias.c (write_dependence_p): Add new parameters mem_mode,
17838 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17839 Changed all callers.
17840 (canon_anti_dependence): New function.
17841 * cse.c (check_dependence): Use canon_anti_dependence.
17842 * cselib.c (cselib_invalidate_mem): Likewise.
17843 * rtl.h (canon_anti_dependence): Declare.
17845 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17847 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17848 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17849 ".set mips3" for 64-bit targets.
17851 2013-06-15 Dehao Chen <dehao@google.com>
17853 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17854 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17855 (gimple_check_call_args): Likewise.
17856 * value-prof.c (check_ic_target): Likewise.
17857 * ipa-inline.c (early_inliner): Likewise.
17858 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17859 * cgraph.c (cgraph_create_edge_1): Likewise.
17860 (cgraph_make_edge_direct): Likewise.
17862 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17865 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17866 rs6000_output_move_128bit to handle emitting quad memory
17867 operations. Set attribute length to 8 bytes.
17869 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17871 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17873 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17874 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17875 (aarch64_<su>mlsl<mode>): Likewise.
17877 2013-06-14 Mike Stump <mikestump@comcast.net>
17879 * Makefile.in (TARGET_H): Add insn-codes.h.
17881 2013-06-14 Alan Modra <amodra@gmail.com>
17883 PR middle-end/57134
17884 PR middle-end/57586
17885 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17886 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17887 bitfield expansion when EXPAND_MEMORY.
17888 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17890 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17892 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17893 test for clearing quad memory on 32-bit later.
17895 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17897 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17898 (fold_negate_expr): Likewise.
17899 (fold_real_zero_addition_p): Handle vectors.
17900 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17902 2013-06-14 Alan Modra <amodra@gmail.com>
17904 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17906 2013-06-13 Jan Hubicka <jh@suse.cz>
17908 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17909 Local comdats are not externally visible.
17910 * symtab.c (dump_symtab_base): Dump externally visible.
17911 (verify_symtab_base): Verify back links in the symtab hash.
17913 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17915 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17916 CONVERT_EXPR as equal nodes.
17918 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17920 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17922 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17924 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17925 Generalize to complex and vector.
17926 * tree.c (build_all_ones_cst): New function.
17927 * tree.h (build_all_ones_cst): Declare it.
17929 2013-06-13 Alan Modra <amodra@gmail.com>
17931 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17932 * config/rs6000/rs6000.md (signbittf2): New insn.
17933 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17934 (abstf2_internal, cmptf_internal2): Likewise.
17935 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17937 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17938 Pat Haugen <pthaugen@us.ibm.com>
17939 Peter Bergner <bergner@vnet.ibm.com>
17941 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17942 power8 byte, half-word, and quad-word atomic instructions.
17943 (emit_store_conditional): Likewise.
17944 (rs6000_expand_atomic_compare_and_swap): Likewise.
17945 (rs6000_expand_atomic_op): Likewise.
17947 * config/rs6000/sync.md (larx): Add new modes for power8.
17949 (AINT): New mode iterator to include TImode as well as normal
17950 integer modes on power8.
17951 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17952 that VSX registers are not considered. Use AINT mode iterator
17953 instead of INT1 to allow inclusion of quad word atomic operations
17955 (load_locked<mode>): Likewise.
17956 (store_conditional<mode>): Likewise.
17957 (atomic_compare_and_swap<mode>): Likewise.
17958 (atomic_exchange<mode>): Likewise.
17959 (atomic_nand<mode>): Likewise.
17960 (atomic_fetch_<fetchop_name><mode>): Likewise.
17961 (atomic_nand_fetch<mode>): Likewise.
17962 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17964 (ATOMIC): On power8, add QImode, HImode modes.
17965 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17966 modes that promote to SImode.
17967 (load_lockedti): Convert TImode arguments to PTImode, so that we
17968 get a guaranteed even/odd register pair.
17969 (load_lockedpti): Likewise.
17970 (store_conditionalti): Likewise.
17971 (store_conditionalpti): Likewise.
17973 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17974 atomic load/store instructions.
17977 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17979 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17981 (insn_count): New attribute, with most cases extracted from...
17982 (length): ...here. Redefine most cases in terms of insn_count.
17983 (single_insn): Delete.
17984 (can_delay): Use insn_count to check for single instructions.
17985 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17986 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17987 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17988 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17989 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17990 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17991 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17992 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17993 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17994 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17995 rather than "length".
17996 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17997 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17998 Use "insn_count" rather than "length".
17999 * config/mips/mips-dsp.md
18000 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
18001 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
18004 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
18006 PR tree-optimization/57361
18007 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
18009 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
18011 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
18013 (aarch64_simd_combine<mode>): New instruction expansion.
18014 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
18015 function prototype.
18016 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
18017 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
18019 2013-06-12 Jan Hubicka <jh@suse.cz>
18021 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
18022 decl has when in streaming stage.
18023 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
18024 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
18026 2013-06-12 Roland Stigge <stigge@antcom.de>
18029 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
18031 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18033 PR tree-optimization/57537
18034 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
18035 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
18037 2013-06-12 Richard Biener <rguenther@suse.de>
18039 * data-streamer.h (streamer_write_char_stream): CSE
18040 obs->current_pointer.
18041 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
18042 streamer_write_char_stream manually and optimize the resulting loop.
18043 (streamer_write_hwi_stream): Likewise.
18045 2013-06-12 Jan Hubicka <jh@suse.cz>
18047 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
18048 * cgraph.h (varpool_create_empty_node): Declare.
18049 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
18051 * symtab.c (symtab_unregister_node): Be lax about missin entries
18053 (symtab_get_node): Update comment.
18054 * varpool.c (varpool_create_empty_node): Break out from ...
18055 (varpool_node_for_decl): ... here.
18056 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
18058 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
18060 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
18061 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
18062 part. Use straight-line flow at the end.
18063 <COMPONENT_REF>: Remove superfluous else.
18064 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
18066 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18069 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
18070 target hook even for !TREE_PUBLIC decls. If no resolution info
18071 is available, return false for common and external decls.
18073 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18075 * config/rl78/constraints.md (U): New constraint.
18076 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
18079 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18082 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
18083 to allow returning address to AT_PLATFORM name.
18085 2013-06-11 Jan Hubicka <jh@suse.cz>
18087 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
18088 * cgraph.h (symtab_node_base): Add weakref flag.
18089 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
18090 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
18091 (output_weakrefs): Use weakref flag.
18092 * fold-const.c (simple_operand_p): Handle WEAK.
18093 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
18094 * ipa.c (varpool_externally_visible_p): Drop weakref.
18095 (function_and_variable_visibility): Update comment; fix weakref
18096 sanity checks; do not clear DECL_WEAK on them.
18097 * lto-cgraph.c (lto_output_node): update.
18098 (lto_output_varpool_node): Update.
18099 (input_overwrite_node): Update.
18100 (input_node): Update.
18101 (input_varpool_node): Update.
18102 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
18103 (lto_symtab_merge_symbols): Add sanity check.
18104 (lto_symtab_prevailing_decl): Do not special case weakrefs.
18105 * passes.c (rest_of_decl_compilation): Set static flag, too.
18106 * symtab.c (dump_symtab_base): Dump weakref.
18107 (verify_symtab_base): Sanity check weakrefs.
18108 (symtab_make_decl_local): Remove duplicated code.
18109 (symtab_alias_ultimate_target): Simplify.
18110 * varpool.c (varpool_create_variable_alias): Set weakref flag.
18112 2013-06-11 Tom de Vries <tom@codesourcery.com>
18114 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
18115 sequence_vect == NULL.
18117 2013-06-11 DJ Delorie <dj@redhat.com>
18119 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
18120 (rl78_unwind_word_mode): New.
18122 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18124 * final.c (debug_prefix_maps): Make static.
18126 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18128 * function.c (initial_trampoline): Remove stray copy.
18130 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
18132 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
18134 2013-06-11 Martin Jambor <mjambor@suse.cz>
18136 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
18137 within bounds at the beginning of the function.
18139 2013-06-11 Alan Modra <amodra@gmail.com>
18141 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
18143 (get_named_section): Don't NULL !DECL_P decl.
18145 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
18147 * doc/invoke.texi (core-avx2): Document.
18149 (atom): Updated with MOVBE.
18151 2013-06-11 Richard Biener <rguenther@suse.de>
18153 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
18155 2013-06-11 Anton Blanchard <anton@samba.org>
18157 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
18158 correct shift value in little-endian mode.
18160 2013-06-11 Jakub Jelinek <jakub@redhat.com>
18163 * varasm.c (get_variable_align): Move #endif to the right place.
18165 2013-06-10 Cary Coutant <ccoutant@google.com>
18167 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
18168 for hash so that hash table traversal order is deterministic.
18170 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
18171 Pat Haugen <pthaugen@us.ibm.com>
18172 Peter Bergner <bergner@vnet.ibm.com>
18174 * config/rs6000/vector.md (GPR move splitter): Do not split moves
18175 of vectors in GPRS if they are direct moves or quad word load or
18178 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
18180 (direct_move_p): Likewise.
18181 (quad_load_store_p): Likewise.
18183 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
18184 classes into bins based on the physical register type.
18185 (reg_class_to_reg_type): Likewise.
18186 (IS_STD_REG_TYPE): Likewise.
18187 (IS_FP_VECT_REG_TYPE): Likewise.
18188 (reload_fpr_gpr): Arrays to determine what insn to use if we can
18189 use direct move instructions.
18190 (reload_gpr_vsx): Likewise.
18191 (reload_vsx_gpr): Likewise.
18192 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
18193 information that is a simplification of register classes. Also
18194 precalculate direct move reload helpers.
18195 (direct_move_p): New function to return true if the operation can
18196 be done as a direct move instruciton.
18197 (quad_load_store_p): New function to return true if the operation
18198 is a quad memory operation.
18199 (rs6000_legitimize_address): If quad memory, only allow register
18200 indirect for TImode addresses.
18201 (rs6000_legitimate_address_p): Likewise.
18202 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
18203 (rs6000_reload_register_type): Likewise.
18204 (register_to_reg_type): Return register type.
18205 (rs6000_secondary_reload_simple_move): New helper function for
18206 secondary reload and secondary memory needed to identify anything
18207 that is a simple move, and does not need reloading.
18208 (rs6000_secondary_reload_direct_move): New helper function for
18209 secondary reload to identify cases that can be done with several
18210 instructions via the direct move instructions.
18211 (rs6000_secondary_reload_move): New helper function for secondary
18212 reload to identify moves between register types that can be done.
18213 (rs6000_secondary_reload): Add support for quad memory operations
18214 and for direct move.
18215 (rs6000_secondary_memory_needed): Likewise.
18216 (rs6000_debug_secondary_memory_needed): Change argument names.
18217 (rs6000_output_move_128bit): New function to return the move to
18218 use for 128-bit moves, including knowing about the various
18219 limitations of quad memory operations.
18221 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
18222 memory operations. call rs6000_output_move_128bit for the actual
18223 instruciton(s) to generate.
18224 (vsx_movti_64bit): Likewise.
18226 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
18227 (UNSPEC_P8V_MTVSRWZ): Likewise.
18228 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
18229 (UNSPEC_P8V_MTVSRD): Likewise.
18230 (UNSPEC_P8V_XXPERMDI): Likewise.
18231 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
18232 (UNSPEC_FUSION_GPR): Likewise.
18233 (FMOVE128_GPR): New iterator for direct move.
18234 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
18235 (f32_sv): Likewise.
18236 (f32_dm): Likewise.
18237 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
18238 loads and direct move instructions.
18239 (zero_extendsidi2_lfiwzx): Likewise.
18240 (extendsidi2_lfiwax): Likewise.
18241 (extendsidi2_nocell): Likewise.
18242 (floatsi<mode>2_lfiwax): Likewise.
18243 (lfiwax): Likewise.
18244 (floatunssi<mode>2_lfiwzx): Likewise.
18245 (lfiwzx): Likewise.
18246 (fix_trunc<mode>_stfiwx): Likewise.
18247 (fixuns_trunc<mode>_stfiwx): Likewise.
18248 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
18249 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
18250 (parity<mode>2_cmpb): Set length/type attr.
18251 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
18252 for 'mr.' to fast_compare.
18253 (bpermd_<mode>): Change type attr to popcnt.
18254 (p8_fmrgow_<mode>): New insns for power8 direct move support.
18255 (p8_mtvsrwz_1): Likewise.
18256 (p8_mtvsrwz_2): Likewise.
18257 (reload_fpr_from_gpr<mode>): Likewise.
18258 (p8_mtvsrd_1): Likewise.
18259 (p8_mtvsrd_2): Likewise.
18260 (p8_xxpermdi_<mode>): Likewise.
18261 (reload_vsx_from_gpr<mode>): Likewise.
18262 (reload_vsx_from_gprsf): Likewise.
18263 (p8_mfvsrd_3_<mode>): LIkewise.
18264 (reload_gpr_from_vsx<mode>): Likewise.
18265 (reload_gpr_from_vsxsf): Likewise.
18266 (p8_mfvsrd_4_disf): Likewise.
18267 (multi-word GPR splits): Do not split direct moves or quad memory
18270 2013-06-10 David Malcolm <dmalcolm@redhat.com>
18272 * tree-into-ssa.c (interesting_blocks): Make static.
18274 2013-06-10 Jakub Jelinek <jakub@redhat.com>
18277 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
18278 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
18279 Use DATA_ABI_ALIGNMENT for that case instead if defined.
18280 (get_variable_align): New function.
18281 (get_variable_section, emit_bss, emit_common,
18282 assemble_variable_contents, place_block_symbol): Use
18283 get_variable_align instead of DECL_ALIGN.
18284 (assemble_noswitch_variable): Add align argument, use it
18285 instead of DECL_ALIGN.
18286 (assemble_variable): Adjust caller. Use get_variable_align
18287 instead of DECL_ALIGN.
18288 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
18290 (DATA_ABI_ALIGNMENT): Define.
18291 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
18292 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
18293 opt is false, only return the psABI mandated alignment increase.
18294 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
18295 (DATA_ABI_ALIGNMENT): ... this.
18296 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
18297 (DATA_ABI_ALIGNMENT): ... this.
18298 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
18299 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18300 (DATA_ABI_ALIGNMENT): ... this.
18301 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18302 * doc/tm.texi: Regenerated.
18304 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18306 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18307 cmp_code to construct REG_EQUAL note.
18309 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18312 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18313 that operands[2] doesn't overlap with operands[0].
18315 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18316 Jan Hubicka <jh@suse.cz>
18318 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18319 hack to mark symbols as used.
18321 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18323 PR rtl-optimization/57559
18324 * lra-constraints.c (process_alt_operands): Don't discourage
18325 memory with known offset for offsetable memory constraint.
18326 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18328 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18330 * varasm.c (struct oc_local_state): Reorder fields.
18331 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18332 and adjust accordingly.
18333 (output_constructor): Reorder initialization code and adjust call to
18334 output_constructor_bitfield.
18336 2013-06-07 Jan Hubicka <jh@suse.cz>
18338 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18340 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18342 * tree-object-size.c (unknown): Make const.
18344 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18346 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18347 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18348 for last alternative in the cpu_facility attribute.
18350 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18353 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18354 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18355 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18356 * config/arm/constraints.md (Dg): New constraint.
18357 * config/arm/neon.md (xordi3_neon): Remove.
18358 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18359 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18361 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18363 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18364 Clean up alternatives.
18366 2013-06-07 Alan Modra <amodra@gmail.com>
18368 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18371 2013-06-07 Alan Modra <amodra@gmail.com>
18373 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18375 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18377 * config/arm/constraints.md (Df): New constraint.
18378 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18379 Correct length attribute for last two alternatives.
18381 2013-06-07 Alan Modra <amodra@gmail.com>
18383 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18384 override user -mfp-in-toc.
18385 (offsettable_ok_by_alignment): Consider just the current access
18386 rather than the whole object, unless BLKmode. Handle
18387 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18388 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18389 for -mcmodel=medium.
18390 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18391 override user -mfp-in-toc or -msum-in-toc. Default to
18392 -mno-fp-in-toc for -mcmodel=medium.
18394 2013-06-06 DJ Delorie <dj@redhat.com>
18396 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18397 TARGET_VALID_POINTER_MODE.
18399 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18400 Pat Haugen <pthaugen@us.ibm.com>
18401 Peter Bergner <bergner@vnet.ibm.com>
18403 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18404 Document new power8 builtins.
18406 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18407 condition code register, to allow 128-bit logical operations to be
18408 done in the VSX or GPR registers.
18409 (nor<mode>3): Use the canonical form for nor.
18410 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18411 vclz*, and vpopcnt* vector instructions.
18412 (nand<mode>3): Likewise.
18413 (orc<mode>3): Likewise.
18414 (clz<mode>2): LIkewise.
18415 (popcount<mode>2): Likewise.
18417 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18418 that only the GPRs are recognized.
18420 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18421 support for new power8 builtins.
18423 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18425 (xscvdpspn): Likewise.
18431 (vpopcnt): Likewise.
18432 (vpopcntb): Likewise.
18433 (vpopcnth): Likewise.
18434 (vpopcntw): Likewise.
18435 (vpopcntd): Likewise.
18437 (vmrgew): Likewise.
18438 (vmrgow): Likewise.
18440 (eqv_v16qi3): Likewise.
18441 (eqv_v8hi3): Likewise.
18442 (eqv_v4si3): Likewise.
18443 (eqv_v2di3): Likewise.
18444 (eqv_v4sf3): Likewise.
18445 (eqv_v2df3): Likewise.
18447 (nand_v16qi3): Likewise.
18448 (nand_v8hi3): Likewise.
18449 (nand_v4si3): Likewise.
18450 (nand_v2di3): Likewise.
18451 (nand_v4sf3): Likewise.
18452 (nand_v2df3): Likewise.
18454 (orc_v16qi3): Likewise.
18455 (orc_v8hi3): Likewise.
18456 (orc_v4si3): Likewise.
18457 (orc_v2di3): Likewise.
18458 (orc_v4sf3): Likewise.
18459 (orc_v2df3): Likewise.
18461 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18462 allow power8 quad mode in 64-bit.
18463 (rs6000_builtin_vectorized_function): Add support to vectorize
18464 ISA 2.07 count leading zeros, population count builtins.
18465 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18466 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18467 (builtin_function_type): Add vgbbd builtin function which takes an
18469 (altivec_expand_vec_perm_const): Add support for new power8 merge
18472 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18473 that does not include TImdoe for use with 32-bit.
18474 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18476 (UNSPEC_VSX_CVDPSPN): Likewise.
18477 (vsx_xscvdpspn): Likewise.
18478 (vsx_xscvspdpn): Likewise.
18479 (vsx_xscvdpspn_scalar): Likewise.
18480 (vsx_xscvspdpn_directmove): Likewise.
18481 (vsx_and<mode>3): Split logical operations into 32-bit and
18482 64-bit. Add support to do logical operations on TImode as well as
18483 VSX vector types. Allow logical operations to be done in either
18484 VSX registers or in general purpose registers in 64-bit mode. Add
18485 splitters if GPRs were used. For AND, add clobber of CCmode to
18486 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18488 (vsx_and<mode>3_32bit): Likewise.
18489 (vsx_and<mode>3_64bit): Likewise.
18490 (vsx_ior<mode>3): Likewise.
18491 (vsx_ior<mode>3_32bit): Likewise.
18492 (vsx_ior<mode>3_64bit): Likewise.
18493 (vsx_xor<mode>3): Likewise.
18494 (vsx_xor<mode>3_32bit): Likewise.
18495 (vsx_xor<mode>3_64bit): Likewise.
18496 (vsx_one_cmpl<mode>2): Likewise.
18497 (vsx_one_cmpl<mode>2_32bit): Likewise.
18498 (vsx_one_cmpl<mode>2_64bit): Likewise.
18499 (vsx_nor<mode>3): Likewise.
18500 (vsx_nor<mode>3_32bit): Likewise.
18501 (vsx_nor<mode>3_64bit): Likewise.
18502 (vsx_andc<mode>3): Likewise.
18503 (vsx_andc<mode>3_32bit): Likewise.
18504 (vsx_andc<mode>3_64bit): Likewise.
18505 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18506 and xxlorc instructions.
18507 (vsx_eqv<mode>3_64bit): Likewise.
18508 (vsx_nand<mode>3_32bit): Likewise.
18509 (vsx_nand<mode>3_64bit): Likewise.
18510 (vsx_orc<mode>3_32bit): Likewise.
18511 (vsx_orc<mode>3_64bit): Likewise.
18513 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18515 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18517 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18518 (p8_vmrgow): Likewise.
18519 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18520 GPRs to be split under VSX.
18521 (p8v_clz<mode>2): Add power8 count leading zero support.
18522 (p8v_popcount<mode>2): Add power8 population count support.
18523 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18526 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18529 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18531 (vec_nand): Likewise.
18532 (vec_vclz): Likewise.
18533 (vec_vclzb): Likewise.
18534 (vec_vclzd): Likewise.
18535 (vec_vclzh): Likewise.
18536 (vec_vclzw): Likewise.
18537 (vec_vgbbd): Likewise.
18538 (vec_vmrgew): Likewise.
18539 (vec_vmrgow): Likewise.
18540 (vec_vpopcnt): Likewise.
18541 (vec_vpopcntb): Likewise.
18542 (vec_vpopcntd): Likewise.
18543 (vec_vpopcnth): Likewise.
18544 (vec_vpopcntw): Likewise.
18546 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18548 PR rtl-optimization/57468
18549 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18552 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18554 PR rtl-optimization/57459
18555 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18556 type when setting live regs.
18558 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18560 * config/s390/s390.opt (mlra): New option.
18561 * config/s390/s390.c (s390_decompose_address): Check displacement
18562 for all registers for LRA.
18563 (s390_secondary_reload): Don't used secondary reloads for LRA.
18564 (s390_lra_p): New function.
18565 (TARGET_LRA_P): Define.
18566 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18567 of attribute cpu_facility to zarch for the last alternative.
18568 (*cmpmem_short): Ditto.
18570 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18572 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18573 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18574 (arm_expand_prologue): Likewise.
18576 2013-06-06 Teresa Johnson <tejohnson@google.com>
18579 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18580 as this is now done by redirect_edge_and_branch_force.
18581 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18582 barriers, and fix interaction with splitting.
18583 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18584 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18585 reflect changes made in the routine.
18586 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18587 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18588 since this is called in cfglayout mode, and replace partition fixup
18589 with assert as that is now done by force_nonfallthru_and_redirect.
18590 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18591 already be marked with region crossing note.
18592 (insert_section_boundary_note): Make non-static, gate on flag
18593 has_bb_partition, rewrite to also check for multiple partitions.
18594 (rest_of_handle_reorder_blocks): Remove call to
18595 insert_section_boundary_note, now done later during free_cfg.
18596 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18597 * bb-reorder.h (insert_section_boundary_note): Declare.
18598 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18599 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18600 invoke insert_section_boundary_note.
18601 (try_redirect_by_replacing_jump): Remove unnecessary
18602 check for region crossing note.
18603 (fixup_partition_crossing): New function.
18604 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18605 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18606 in non-cfglayout mode.
18607 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18608 remove old code that tried to do this. Emit barrier correctly
18609 when we are in cfglayout mode.
18610 (last_bb_in_partition): New function.
18611 (rtl_split_edge): Correctly fixup partition boundaries.
18612 (commit_one_edge_insertion): Remove old code that tried to
18613 fixup region crossing edge since this is now handled in
18614 split_block, and set up insertion point correctly since
18615 block may now end in a jump.
18616 (verify_hot_cold_block_grouping): Guard against checking when not in
18617 linearized RTL mode.
18618 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18620 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18621 rtl_verify_flow_info, so not called in cfglayout mode.
18622 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18623 (fixup_reorder_chain): Remove old code that attempted to fixup region
18624 crossing note as this is now handled in force_nonfallthru_and_redirect.
18625 (duplicate_insn_chain): Don't duplicate switch section notes.
18626 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18628 * basic-block.h (emit_barrier_after_bb): Declare.
18630 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18632 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18633 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18634 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18636 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18638 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18640 * config/arm/ldmstm.md: Regenerate.
18642 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18644 * config/arm/sync.md (atomic_loaddi_1):
18645 Disable predication for arm_restrict_it.
18646 (arm_load_exclusive<mode>): Likewise.
18647 (arm_load_exclusivesi): Likewise.
18648 (arm_load_exclusivedi): Likewise.
18649 (arm_load_acquire_exclusive<mode>): Likewise.
18650 (arm_load_acquire_exclusivesi): Likewise.
18651 (arm_load_acquire_exclusivedi): Likewise.
18652 (arm_store_exclusive<mode>): Likewise.
18653 (arm_store_exclusive<mode>): Likewise.
18654 (arm_store_release_exclusivedi): Likewise.
18655 (arm_store_release_exclusive<mode>): Likewise.
18657 2013-06-06 Richard Biener <rguenther@suse.de>
18659 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18661 (lto_tag_is_tree_code_p): Adjust.
18662 (lto_tag_is_gimple_code_p): Likewise.
18663 (lto_gimple_code_to_tag): Likewise.
18664 (lto_tag_to_gimple_code): Likewise.
18665 (lto_tree_code_to_tag): Likewise.
18666 (lto_tag_to_tree_code): Likewise.
18667 * data-streamer.h (streamer_write_hwi_in_range): Use
18668 uhwi streaming to stream the normalized range.
18669 (streamer_read_hwi_in_range): Likewise.
18671 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18673 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18674 (predicable_short_it): Likewise.
18675 (predicated): Likewise.
18676 (enabled): Handle above.
18677 (define_cond_exec): Set predicated attribute to yes.
18679 2013-06-05 Mike Stump <mikestump@comcast.net>
18681 * gdbinit.in (__FUNCTION__): Add.
18683 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18685 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18686 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18688 2013-06-05 Jan Hubicka <jh@suse.cz>
18690 * varasm.c (mark_decl_referenced): Revert the removal until targets
18693 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18695 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18696 instead of mark_decl_referenced.
18698 2013-06-05 Jan Hubicka <jh@suse.cz>
18700 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18701 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18702 and symtab_used_from_object_file_p.
18703 (cgraph_make_node_local_1): Clear forced_by_abi.
18704 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18705 * cgraph.h (symtab_node_base): Add forced_by_abi.
18706 (decide_is_variable_needed): Remove.
18707 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18708 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18709 (decide_is_symbol_needed): ... this one; handle symbols in general;
18710 always analyze virtuals; honnor forced_by_abi.
18711 (cgraph_finalize_function): Update.
18712 (varpool_finalize_decl): Update.
18713 (symbol_defined_and_needed): Remove.
18714 (analyze_functions): Update.
18715 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18716 output_refs, input_overwrite_node): Handle forced_by_abi.
18717 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18718 (address_taken_from_non_vtable_p): ... this one.
18719 (comdat_can_be_unshared_p_1): New function.
18720 (cgraph_comdat_can_be_unshared_p): Rename to ...
18721 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18722 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18723 (function_and_variable_visibility): Clear forced_by_abi as needed.
18724 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18725 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18726 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18727 * varpool.c (decide_is_variable_needed): Remove.
18729 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18731 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18732 (arm_option_override): Override arm_restrict_it where appropriate.
18733 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18734 * config/arm/arm.opt (mrestrict-it): New command-line option.
18735 * doc/invoke.texi: Document -mrestrict-it.
18737 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18739 * tsan.c (tsan_atomic_table): Make const.
18741 2013-06-05 Richard Biener <rguenther@suse.de>
18743 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18744 index associated with the tree we are supposed to replace.
18745 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18746 TREE_ASM_WRITTEN as zero for everything but SSA names.
18748 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18750 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18752 2013-06-04 Jan Hubicka <jh@suse.cz>
18754 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18755 (input_node, input_varpool_node): Handle correctly external same
18757 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18758 nodes at ltrans stage.
18760 2013-06-04 Jan Hubicka <jh@suse.cz>
18762 * ipa-inline.c (update_caller_keys): Fix availability test.
18763 (update_callee_keys): Likewise.
18764 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18765 to follow ELF standard.
18767 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18769 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18770 (mips64r5900el-*-elf*): New configurations.
18771 * config/mips/mips-cpus.def (r5900): New processor.
18772 * config/mips/mips-tables.opt: Regenerate.
18773 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18774 (mips_issue_rate): Handle PROCESSOR_R5900.
18775 (mips_reorg_process_insns): Force reorder mode for the R5900.
18776 * config/mips/mips.h (TARGET_MIPS5900): Define.
18777 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18779 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18781 * config/mips/mips.md (processor): Add r5900.
18782 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18784 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18786 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18787 into function to generate MOVI instruction.
18788 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18789 (aarch64_preferred_simd_mode): Turn into wrapper.
18790 (aarch64_output_scalar_simd_mov_immediate): New function.
18791 * config/aarch64/aarch64-protos.h: Add prototype for above.
18793 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18795 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18796 element_char member.
18797 (sizetochar): Return signed char.
18798 (aarch64_simd_valid_immediate): Remove elchar and other
18799 unnecessary variables.
18800 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18801 Calculate element_char as required.
18802 * config/aarch64/aarch64-protos.h: Update and move prototype
18803 for aarch64_output_simd_mov_immediate.
18804 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18807 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18809 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18810 information completed by aarch64_simd_valid_immediate.
18811 (aarch64_legitimate_constant_p): Update arguments.
18812 (aarch64_simd_valid_immediate): Work with struct rather than many
18814 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18815 (aarch64_simd_make_constant): Update arguments.
18816 (aarch64_output_simd_mov_immediate): Work with struct rather than
18817 many pointers. Output immediate directly rather than as operand.
18818 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18820 * config/aarch64/constraints.md (Dn): Update arguments.
18822 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18824 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18826 (aarch64_simd_immediate_valid_for_move): Remove.
18827 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18828 (aarch64_simd_make_constant): Update call.
18829 (aarch64_output_simd_mov_immediate): Update call.
18830 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18832 * config/aarch64/constraints.md (Dn): Update call.
18834 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18836 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18837 return type to bool for prototype.
18838 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18839 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18840 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18842 2013-06-04 Catherine Moore <clm@codesourcery.com>
18844 * config/mips/mips.opt (meva): New.
18845 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18846 (ASM_SPEC): Handle -meva.
18847 * doc/invoke.texi (meva): Document.
18849 2013-06-04 Alan Modra <amodra@gmail.com>
18851 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18854 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18856 * rtl.def: Add extra fourth optional field to define_cond_exec.
18857 * gensupport.c (process_one_cond_exec): Process attributes from
18859 * doc/md.texi: Document fourth field in define_cond_exec.
18861 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18863 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18864 out the processing order as in store_bit_field_1.
18866 2013-06-04 Jan Hubicka <jh@suse.cz>
18868 PR middle-end/57500
18869 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18870 non-VAR_DECL node if it does not exist yet.
18872 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18874 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18875 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18876 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18877 target_cpu_default setting.
18879 2013-06-03 Teresa Johnson <tejohnson@google.com>
18881 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18882 default to -fopt-info=optimized instead of all.
18883 * doc/invoke.texi: Ditto.
18884 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18885 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18886 (execute_vect_slp): Emit BB vectorization success under
18887 MSG_OPTIMIZED_LOCATIONS.
18888 * tree-vect-slp.c (vect_slp_transform_bb): Change
18889 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18890 * tree-vect-loop.c (vect_transform_loop): Ditto.
18892 2013-06-03 Jason Merrill <jason@redhat.com>
18895 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18896 Use TARGET_EXPR for C++.
18898 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18900 PR rtl-optimization/57268
18901 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18902 if DEBUG_INSN_P (insn).
18905 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18907 PR rtl-optimization/57268
18908 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18909 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18911 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18913 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18914 (ix86_avoid_lea_for_addr): Likewise.
18915 (exact_dependency_1): Likewise.
18916 (ix86_adjust_cost): Likewise.
18917 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18919 (do_reorder_for_imul): Fix typo, formatting and
18920 !reload_completed check removed.
18921 (ix86_sched_reorder): Fix typo and formatting.
18922 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18925 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18927 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18929 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18931 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18932 <STRING_CST>: Likewise.
18933 <VECTOR_CST>: Likewise.
18935 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18936 Mikael Morin <mikael@gcc.gnu.org>
18938 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18939 * config.in: Regenerated.
18940 * configure: Regenerated.
18942 2013-06-01 Jan Hubicka <jh@suse.cz>
18944 PR middle-end/57366
18945 * cgraphunit.c (compile): When weakref is not supported,
18946 set up transparent aliases before final output pass.
18947 * varasm.c (assemble_alias): Do not try to do it here.
18949 2013-06-01 Jan Hubicka <jh@suse.cz>
18951 PR middle-end/57467
18952 * passes.c (for_per_function): Skip unanalyzed functions.
18954 2013-06-01 Jan Hubicka <jh@suse.cz>
18956 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18957 (lto_symtab_merge_symbols_1): ... this one.
18958 (lto_symtab_merge_cgraph_nodes): Rename to ...
18959 (lto_symtab_merge_symbols): ... this one; simplify.
18960 * cgraph.c (same_body_aliases_done): Rename to ...
18961 (cpp_implicit_aliases_done): ... this one.
18962 (cgraph_create_function_alias): Update.
18963 (cgraph_same_body_alias): Update.
18964 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18965 (verify_edge_corresponds_to_fndecl): Simplify.
18966 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18967 (cgraph_node): Remove same_body_alias.
18968 (varpool_node): Remove alias_of and extra_name_alias.
18969 (same_body_aliases_done): Rename to ..
18970 (cpp_implicit_aliases_done): ... this one.
18971 (symtab_alias_ultimate_target): Add default parameter.
18972 (symtab_resolve_alias): New function.
18973 (fixup_same_cpp_alias_visibility): Declare.
18974 (cgraph_function_node): Add default parameter.
18975 (cgraph_node_asm_name): Likewise.
18976 (cgraph_function_or_thunk_node): Add default parameter; do
18977 not ICE when it is NULL.
18978 (varpool_variable_node): Likewise.
18979 * tree-emutls.c (create_emultls_var): Update.
18980 (ipa_lower_emutls): Update.
18981 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18982 (cgraph_reset_node): Reset alias info.
18983 (cgraph_finalize_function): Update.
18984 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18985 (analyze_function): Simplify.
18986 (cgraph_process_same_body_aliases): Simplify.
18987 (analyze_functions): Fixup same body aliases.
18988 (handle_alias_pairs): Simplify.
18989 (assemble_thunk): Update.
18990 (assemble_thunks_and_aliases): Update.
18991 (output_weakrefs): Rewrite.
18992 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18993 (lto_output_varpool_node): Likewise.
18994 (compute_ltrans_boundary): Remve assert.
18995 (get_alias_symbol): New functoin.
18996 (input_node): Rewrite alias handling.
18997 (input_varpool_node): Likewise.
18998 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18999 * ipa.c (process_references): Handle weakrefs correctly.
19000 (symtab_remove_unreachable_nodes): Likewise.
19001 * trans-mem.c (get_cg_data): Update.
19002 (ipa_tm_create_version_alias): Update.
19003 (ipa_tm_execute): Update.
19004 * symtab.c (dump_symtab_base): Dump aliases.
19005 (verify_symtab_base): Verify aliases.
19006 (symtab_node_availability): New function.
19007 (symtab_alias_ultimate_target): Simplify.
19008 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
19009 handle all the fixup cases.
19010 (symtab_resolve_alias): New function.
19011 * passes.c (ipa_write_summaries): Handle weakrefs.
19012 * varpool.c (varpool_analyze_node): Simplify.
19013 (assemble_aliases): Update.
19014 (varpool_create_variable_alias): Simplify.
19015 (varpool_extra_name_alias): Simplify.
19016 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
19017 (lto_symtab_merge_symbols): ... this one.
19019 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
19022 PR rtl-optimization/57268
19023 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19024 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19026 2013-06-01 Tobias Burnus <burnus@net-b.de>
19028 Partially reverted:
19029 2013-05-31 Tobias Burnus <burnus@net-b.de>
19031 PR middle-end/57073
19032 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
19035 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19037 PR rtl-optimization/57268
19038 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19039 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19041 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19043 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
19044 unordered comparison operators when -fno-trapping-math is in effect
19046 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
19047 and implement unordered comparison operators properly on the e500.
19049 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19051 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
19052 for constant scalar integers.
19053 (simplify_relational_operation_1): Likewise.
19055 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
19057 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
19058 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
19061 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19062 Igor Zamyatin <igor.zamyatin@intel.com>
19064 Silvermont (SLM) architecture performance tuning.
19065 * config/i386/i386.h (enum ix86_tune_indices): Add
19066 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
19067 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
19069 * config/i386/i386.c (initial_ix86_tune_features)
19070 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
19071 (ix86_lea_outperforms): Handle Silvermont tuning.
19072 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
19074 (ix86_use_lea_for_mov): Likewise.
19075 (ix86_avoid_lea_for_addr): Likewise.
19076 (ix86_lea_for_add_ok): Likewise.
19077 (exact_dependency_1): New function.
19078 (exact_store_load_dependency): Likewise.
19079 (ix86_adjust_cost): Handle Silvermont tuning.
19080 (do_reoder_for_imul): Likewise.
19081 (swap_top_of_ready_list): New function.
19082 (ix86_sched_reorder): Changed to handle Silvermont tuning.
19084 * config/i386/i386.md (peepholes that split memory operand in fp
19087 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19089 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
19090 Remove un-necessary braces.
19092 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19094 * config/aarch64/aarch64.c (aarch64_classify_symbol):
19095 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
19097 2013-05-31 Tobias Burnus <burnus@net-b.de>
19099 PR middle-end/57073
19100 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
19102 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19105 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
19106 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
19107 * config/arm/neon.md (iordi3_neon): Remove.
19108 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
19109 * config/arm/predicates.md (imm_for_neon_logic_operand):
19110 Move to earlier in the file.
19111 (neon_logic_op2): Likewise.
19112 (arm_iordi_operand_neon): New predicate.
19114 2013-05-31 Richard Biener <rguenther@suse.de>
19116 PR tree-optimization/57478
19117 PR tree-optimization/57453
19118 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
19121 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19123 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
19124 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
19126 2013-05-30 Tobias Burnus <burnus@net-b.de>
19127 Thomas Koenig <tkoenig@gcc.gnu.org>
19129 PR middle-end/57073
19130 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
19131 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
19133 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
19135 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
19137 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
19139 * target.def (register_usage_leveling_p): New hook.
19140 * targhooks.c (default_register_usage_leveling_p): New.
19141 * targhooks.h (default_register_usage_leveling_p): New prototype.
19142 * lra-assigns.c (register_usage_leveling_p): Use the hook.
19143 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
19144 * doc/tm.texi: Update.
19145 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
19147 2013-05-30 Ian Bolton <ian.bolton@arm.com>
19149 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
19150 (*insv_reg<mode>): New define_insn.
19152 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
19154 PR rtl-optimization/57439
19155 * postreload.c (move2add_valid_value_p): Check that we have
19156 a zero subreg_regno_offset when accessing the register in
19157 the requested mode.
19159 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19160 Igor Zamyatin <igor.zamyatin@intel.com>
19162 Silvermont (SLM) architecture pipeline model, tuning and
19164 * config.gcc: Add slm config options and target.
19166 * config/i386/slm.md: New.
19168 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
19170 * config/i386/i386-c.c (ix86_target_macros_internal): New case
19172 (ix86_target_macros_internal): Likewise.
19174 * config/i386/i386.c (slm_cost): New cost.
19175 (m_SLM): New macro flag.
19176 (initial_ix86_tune_features): Set m_SLM.
19177 (x86_accumulate_outgoing_args): Likewise.
19178 (x86_arch_always_fancy_math_387): Likewise.
19179 (processor_target_table): Add slm cost.
19180 (cpu_names): Add slm cpu name.
19181 (x86_option_override_internal): Set SLM ISA.
19182 (ix86_issue_rate): New case PROCESSOR_SLM.
19183 (ia32_multipass_dfa_lookahead): Likewise.
19184 (fold_builtin_cpu): Add slm.
19186 * config/i386/i386.h (TARGET_SLM): New target macro.
19187 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
19188 (processor_type): Add PROCESSOR_SLM.
19190 * config/i386/i386.md (cpu): Add new value "slm".
19191 (slm.md): Include slm.md.
19193 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
19194 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19196 * config/arm/arm-protos.h: Add and update function protos.
19197 * config/arm/arm.c (use_simple_return_p): New added.
19198 (thumb2_expand_return): Check simple_return flag.
19199 * config/arm/arm.md: Add simple_return and conditional simple_return.
19200 * config/arm/iterators.md: Add iterator for return and simple_return.
19202 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19204 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
19205 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
19206 (arm_emit_vfp_multi_reg_pop): Likewise.
19207 (thumb2_emit_ldrd_pop): Likewise.
19208 (arm_expand_epilogue): Add misc REG_CFA notes.
19209 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
19211 2013-05-29 Lawrence Crowl <crowl@google.com>
19213 * config/arm/t-arm: Update for below.
19215 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
19216 Change type to hash_table. Update dependent calls and types.
19218 * config/i386/t-cygming: Update for below.
19220 * config/i386/t-interix: Update for below.
19222 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
19223 Change type to hash_table. Update dependent calls and types.
19224 (i386_find_on_wrapper_list::wrappers): Likewise.
19226 * config/ia64/t-ia64: Update for below.
19228 * config/ia64/ia64.c (bundle_state_table):
19229 Change type to hash_table. Update dependent calls and types.
19231 * config/mips/mips.c (mips_reorg_process_insns::htab):
19232 Change type to hash_table. Update dependent calls and types.
19234 * config/sol2.c (solaris_comdat_htab):
19235 Change type to hash_table. Update dependent calls and types.
19237 * config/t-sol2: Update for above.
19239 2013-05-29 Teresa Johnson <tejohnson@google.com>
19241 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
19242 functions are not yet marked as defined.
19244 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
19245 Pat Haugen <pthaugen@us.ibm.com>
19246 Peter Bergner <bergner@vnet.ibm.com>
19248 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
19252 (vrotl<mode>3): Likewise.
19253 (vashl<mode>3): Likewise.
19254 (vlshr<mode>3): Likewise.
19255 (vashr<mode>3): Likewise.
19257 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19258 support for power8 V2DI builtins.
19260 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
19261 power8 V2DI builtins.
19262 (vupkhsw): Likewise.
19263 (vupklsw): Likewise.
19264 (vaddudm): Likewise.
19265 (vminsd): Likewise.
19266 (vmaxsd): Likewise.
19267 (vminud): Likewise.
19268 (vmaxud): Likewise.
19269 (vpkudum): Likewise.
19270 (vpksdss): Likewise.
19271 (vpkudus): Likewise.
19272 (vpksdus): Likewise.
19277 (vsubudm): Likewise.
19278 (vcmpequd): Likewise.
19279 (vcmpgtsd): Likewise.
19280 (vcmpgtud): Likewise.
19281 (vcmpequd_p): Likewise.
19282 (vcmpgtsd_p): Likewise.
19283 (vcmpgtud_p): Likewise.
19284 (vupkhsw): Likewise.
19285 (vupklsw): Likewise.
19286 (vaddudm): Likewise.
19287 (vmaxsd): Likewise.
19288 (vmaxud): Likewise.
19289 (vminsd): Likewise.
19290 (vminud): Likewise.
19291 (vpksdss): Likewise.
19292 (vpksdus): Likewise.
19293 (vpkudum): Likewise.
19294 (vpkudus): Likewise.
19299 (vsubudm): Likewise.
19301 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19302 support for power8 V2DI instructions.
19304 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19305 power8 V2DI instructions. Combine pack and unpack insns to use an
19306 iterator for each mode. Check whether a particular mode supports
19307 Altivec instructions instead of just checking TARGET_ALTIVEC.
19308 (UNSPEC_VPKUWUM): Likewise.
19309 (UNSPEC_VPKSHSS): Likewise.
19310 (UNSPEC_VPKSWSS): Likewise.
19311 (UNSPEC_VPKUHUS): Likewise.
19312 (UNSPEC_VPKSHUS): Likewise.
19313 (UNSPEC_VPKUWUS): Likewise.
19314 (UNSPEC_VPKSWUS): Likewise.
19315 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19316 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19317 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19318 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19319 (UNSPEC_VUPKHSB): Likewise.
19320 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19321 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19322 (UNSPEC_VUPKHSH): Likewise.
19323 (UNSPEC_VUPKLSB): Likewise.
19324 (UNSPEC_VUPKLSH): Likewise.
19326 (VI_char): Likewise.
19327 (VI_scalar): Likewise.
19328 (VI_unit): Likewise.
19330 (VP_small): Likewise.
19331 (VP_small_lc): Likewise.
19332 (VU_char): Likewise.
19333 (add<mode>3): Likewise.
19334 (altivec_vaddcuw): Likewise.
19335 (altivec_vaddu<VI_char>s): Likewise.
19336 (altivec_vadds<VI_char>s): Likewise.
19337 (sub<mode>3): Likewise.
19338 (altivec_vsubcuw): Likewise.
19339 (altivec_vsubu<VI_char>s): Likewise.
19340 (altivec_vsubs<VI_char>s): Likewise.
19341 (altivec_vavgs<VI_char>): Likewise.
19342 (altivec_vcmpbfp): Likewise.
19343 (altivec_eq<mode>): Likewise.
19344 (altivec_gt<mode>): Likewise.
19345 (altivec_gtu<mode>): Likewise.
19346 (umax<mode>3): Likewise.
19347 (smax<mode>3): Likewise.
19348 (umin<mode>3): Likewise.
19349 (smin<mode>3): Likewise.
19350 (altivec_vpkuhum): Likewise.
19351 (altivec_vpkuwum): Likewise.
19352 (altivec_vpkshss): Likewise.
19353 (altivec_vpkswss): Likewise.
19354 (altivec_vpkuhus): Likewise.
19355 (altivec_vpkshus): Likewise.
19356 (altivec_vpkuwus): Likewise.
19357 (altivec_vpkswus): Likewise.
19358 (altivec_vpks<VI_char>ss): Likewise.
19359 (altivec_vpks<VI_char>us): Likewise.
19360 (altivec_vpku<VI_char>us): Likewise.
19361 (altivec_vpku<VI_char>um): Likewise.
19362 (altivec_vrl<VI_char>): Likewise.
19363 (altivec_vsl<VI_char>): Likewise.
19364 (altivec_vsr<VI_char>): Likewise.
19365 (altivec_vsra<VI_char>): Likewise.
19366 (altivec_vsldoi_<mode>): Likewise.
19367 (altivec_vupkhsb): Likewise.
19368 (altivec_vupkhs<VU_char>): Likewise.
19369 (altivec_vupkls<VU_char>): Likewise.
19370 (altivec_vupkhsh): Likewise.
19371 (altivec_vupklsb): Likewise.
19372 (altivec_vupklsh): Likewise.
19373 (altivec_vcmpequ<VI_char>_p): Likewise.
19374 (altivec_vcmpgts<VI_char>_p): Likewise.
19375 (altivec_vcmpgtu<VI_char>_p): Likewise.
19376 (abs<mode>2): Likewise.
19377 (vec_unpacks_hi_v16qi): Likewise.
19378 (vec_unpacks_hi_v8hi): Likewise.
19379 (vec_unpacks_lo_v16qi): Likewise.
19380 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19381 (vec_unpacks_lo_v8hi): Likewise.
19382 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19383 (vec_pack_trunc_v8h): Likewise.
19384 (vec_pack_trunc_v4si): Likewise.
19385 (vec_pack_trunc_<mode>): Likewise.
19387 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19389 (vec_vmaxsd): Likewise.
19390 (vec_vmaxud): Likewise.
19391 (vec_vminsd): Likewise.
19392 (vec_vminud): Likewise.
19393 (vec_vpksdss): Likewise.
19394 (vec_vpksdus): Likewise.
19395 (vec_vpkudum): Likewise.
19396 (vec_vpkudus): Likewise.
19397 (vec_vrld): Likewise.
19398 (vec_vsld): Likewise.
19399 (vec_vsrad): Likewise.
19400 (vec_vsrd): Likewise.
19401 (vec_vsubudm): Likewise.
19402 (vec_vupkhsw): Likewise.
19403 (vec_vupklsw): Likewise.
19405 2013-05-29 Jan Hubicka <jh@suse.cz>
19407 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19408 flags; reorder rest of fields in more consistent way.
19409 (varpool_node): Remove analyzed, finalized and alias.
19410 (cgraph_ndoe): Likewise.
19411 (symtab_alias_ultimate_target): New function.
19412 (cgraph_function_node): Move offline.
19413 (cgraph_reset_node): Declare.
19414 (cgraph_comdat_can_be_unshared_p): Remove.
19415 (varpool_remove_initializer): Declare.
19416 (varpool_first_defined_variable, varpool_next_defined_variable
19417 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19418 (cgraph_function_with_gimple_body_p): Update.
19419 (varpool_all_refs_explicit_p): Update.
19420 (symtab_alias_target): New function.
19421 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19422 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19423 (cgraph_function_or_thunk_node): Simplify using
19424 symtab_alias_ultimate_target.
19425 (varpool_variable_node): Likewise.
19426 * cgraph.c (cgraph_create_function_alias): Update.
19427 (cgraph_add_thunk): Update.
19428 (cgraph_remove_node): Update.
19429 (dump_cgraph_node): Do not dump removed flags.
19430 (cgraph_function_body_availability): Update.
19431 (cgraph_propagate_frequency): Update.
19432 (verify_cgraph_node): Check sanity of local flag.
19433 (cgraph_function_node): Move here from cgraph.h; revamp for
19434 cgraph_function_or_thunk_node.
19435 * lto-symtab.c (lto_varpool_replace_node): Update.
19436 (lto_symtab_resolve_can_prevail_p): Update.
19437 (lto_symtab_merge_cgraph_nodes): Update.
19438 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19439 propagate_constants_accross_call, devirtualization_time_bonus,
19440 ipcp_propagate_stage): Update.
19441 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19442 * ipa-inline-transform.c (clone_inlined_nodes,
19443 preserve_function_body_p): Update.
19444 * ipa-reference.c (propagate): Update.
19445 (write_node_summary_p): Update.
19446 * toplev.c (wrapup_global_declaration_2): Update.
19447 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19448 (analyze_function) ... this one.
19449 (cgraph_process_new_functions): Update.
19450 (cgraph_reset_node): Export.
19451 (cgraph_finalize_function): Update.
19452 (cgraph_add_new_function): Update.
19453 (process_function_and_variable_attributes): Update.
19454 (varpool_finalize_decl): Update.
19455 (symbol_finalized): Remove.
19456 (symbol_finalized_and_needed): Rename to ...
19457 (symbol_defined_and_needed): ... update.
19458 (cgraph_analyze_functions): Update.
19459 (handle_alias_pairs): Update.
19460 (mark_functions_to_output): Update.
19461 (assemble_thunk): Update.
19462 (output_in_order): Update.
19463 (output_weakrefs): Update.
19464 (finalize_compilation_unit): Update.
19465 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19466 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19467 input_node, input_varpool_node): Update.
19468 * dbxout.c (dbxout_expand_expr): Update.
19469 * cgraphclones.c (cgraph_clone_node): Update.
19470 (cgraph_copy_node_for_versioning): Update.
19471 (cgraph_materialize_clone): Update.
19472 (cgraph_materialize_all_clones): Update.
19473 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19474 propagate_pure_const, propagate_nothrow): Update.
19475 * lto-streamer-out.c (lto_output, write_symbol): Update.
19476 * ipa-utils.c (ipa_reverse_postorder): Update.
19477 * ipa-inline.c (can_inline_edge_p): Update.
19478 (update_caller_keys, ipa_inline): Update.
19479 * dwarf2out.c (reference_to_unused,
19480 premark_types_used_by_global_vars_helper): Update.
19481 * tree-eh.c (tree_could_trap_p): Update.
19482 * ipa-split.c (consider_split, execute_split_functions): Update.
19483 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19484 has_addr_references_p): Update; move ahead in file for better
19486 (process_references): Simplify.
19487 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19488 bodies are removed.
19489 (cgraph_comdat_can_be_unshared_p): Make static.
19490 (cgraph_externally_visible_p): Update.
19491 (varpool_externally_visible_p): Update.
19492 (function_and_variable_visibility): Update.
19493 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19494 ipa_tm_mark_force_output_node): Update.
19495 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19496 estimate_edge_devirt_benefit, inline_generate_summary,
19497 inline_write_summary): Update.
19498 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19499 * ipa-prop.c (ipa_compute_jump_functions): Update.
19500 (ipa_print_node_params, ipa_prop_read_section,
19501 ipa_update_after_lto_read, read_replacements_section): Update.
19502 * varasm.c (mark_decl_referenced): Update.
19503 (assemble_alias, dump_tm_clone_pairs): Update.
19504 * tree-inline.c (copy_bb): Update.
19505 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19507 * symtab.c (dump_symtab_base): Print new flags.
19508 (verify_symtab_base): Verify new flags.
19509 (symtab_alias_ultimate_target): New function.
19510 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19511 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19513 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19515 * i386.c (ix86_get_function_versions_dispatcher,
19516 ix86_generate_version_dispatcher_body): Update.
19517 (fold_builtin_cpu): Use varpool_add_new_variable.
19518 * varpool.c (varpool_remove_initializer): Break out from ...
19519 (varpool_remove_node): ... this one.
19520 (dump_varpool_node, varpool_node_for_asm,
19521 cgraph_variable_initializer_availability, varpool_analyze_node,
19522 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19523 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19526 2013-05-29 Jan Hubicka <jh@suse.cz>
19528 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19530 2013-05-29 Easwaran Raman <eraman@google.com>
19532 PR tree-optimization/57442
19533 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19534 when control exits the main loop.
19536 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19538 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19540 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19541 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19542 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19545 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19547 PR tree-optimization/57441
19548 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19549 Don't limit size of incr_vec to number of candidates.
19551 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19553 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19554 and mips16 directories.
19555 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19556 (MULTILIB_DIRNAMES): Ditto.
19557 (MULTILIB_EXCEPTIONS): Add new exceptions.
19558 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19559 (MULTILIB_DIRNAMES): Ditto.
19560 (MULTILIB_EXCEPTIONS): Add new exceptions.
19562 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19563 Marcus Shawcroft <marcus.shawcroft@arm.com>
19565 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19566 SYMBOL_TINY_ABSOLUTE.
19567 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19568 SYMBOL_TINY_ABSOLUTE.
19569 (aarch64_expand_mov_immediate): Likewise.
19570 (aarch64_classify_symbol): Likewise.
19571 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19572 Permit SYMBOL_TINY_ABSOLUTE.
19573 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19575 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19576 Marcus Shawcroft <marcus.shawcroft@arm.com>
19578 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19579 Refactor if/switch. Replace gcc_assert with if.
19581 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19583 * config/i386/i386.c (initial_ix86_tune_features): Enable
19584 FP Reassociation for AMD bdver1 and bdver2.
19586 2013-05-29 Martin Jambor <mjambor@suse.cz>
19588 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19589 and IMAGPART_EXPR do not occur within other handled_components.
19591 2013-05-29 Richard Biener <rguenther@suse.de>
19593 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19594 access on whether the use is in the BB we currently try to
19596 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19597 vectorize to vect_bb_slp_scalar_cost.
19599 2013-05-29 Richard Biener <rguenther@suse.de>
19601 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19602 computing scalar cost offsetted by stmts that are kept live
19604 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19605 for computation of scalar cost.
19607 2013-05-28 Steve Ellcey <sellcey@mips.com>
19609 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19611 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19613 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19614 array notation built-in reduction functions.
19615 * doc/passes.texi (Passes): Added documentation about changes done
19617 * doc/invoke.texi (C Dialect Options): Added documentation about
19618 the -fcilkplus flag.
19619 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19620 (BUILTINS_DEF): Depend on cilkplus.def.
19621 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19622 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19623 * cilkplus.def: New file.
19625 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19627 PR rtl-optimization/57439
19628 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19630 2013-05-28 Easwaran Raman <eraman@google.com>
19632 PR tree-optimization/57337
19633 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19634 (find_insert_point): Correctly identify the insertion point
19635 when two statements with the same UID is compared.
19637 2013-05-28 Richard Biener <rguenther@suse.de>
19639 PR tree-optimization/56787
19640 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19641 from the list of data references.
19642 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19644 (vect_analyze_loop_operations): Likewise.
19645 (vect_transform_loop): Remove clobbers.
19647 2013-05-28 Martin Jambor <mjambor@suse.cz>
19649 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19650 and REALPART_EXPRs have scalar type.
19652 2013-05-28 Richard Biener <rguenther@suse.de>
19654 PR tree-optimization/57411
19655 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19657 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19658 virtual operand propagation.
19660 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19662 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19663 destination register for bmasksi_vis.
19664 (vector_init_bshuffle): Likewise.
19665 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19667 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19669 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19670 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19671 mode if the instruction isn't available in the original mode.
19672 * config/sparc/sparc.opt (mfix-ut699): New option.
19673 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19674 (divdf3): Turn into expander.
19675 (divdf3_nofix): New insn.
19676 (divdf3_fix): Likewise.
19677 (divsf3): Disable if -mfix-ut699.
19678 (sqrtdf2): Turn into expander.
19679 (sqrtdf2_nofix): New insn.
19680 (sqrtdf2_fix): Likewise.
19681 (sqrtsf2): Disable if -mfix-ut699.
19683 2013-05-27 Richard Biener <rguenther@suse.de>
19685 PR middle-end/57412
19686 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19687 block for the new loop.
19689 2013-05-27 Richard Biener <rguenther@suse.de>
19691 PR tree-optimization/57343
19692 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19693 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19694 (number_of_iterations_cond): Do not build the folded tree.
19696 2013-05-27 Richard Biener <rguenther@suse.de>
19699 PR middle-end/57381
19700 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19701 OEP_CONSTANT_ADDRESS_OF retained.
19703 PR tree-optimization/57417
19704 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19705 for unchanged base.
19706 (set_ssa_val_to): Compare addresses using
19707 get_addr_base_and_unit_offset.
19709 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19711 PR rtl-optimization/56833
19712 * postreload.c (move2add_record_mode): New function.
19713 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19714 (move2add_use_add2_insn): Use move2add_record_sym_value.
19715 (move2add_use_add3_insn): Likewise.
19716 (reload_cse_move2add): Use move2add_valid_value_p and
19717 move2add_record_mode. Invalidate call-clobbered and REG_INC
19718 affected regs by setting reg_mode to VOIDmode.
19719 (move2add_note_store): Don't pretend the inside of a SUBREG is
19720 the actual destination. Invalidate single/leading registers by
19721 setting reg_mode to VOIDmode.
19722 Use move2add_record_sym_value, move2add_valid_value_p and
19723 move2add_record_mode.
19725 2013-05-27 Richard Biener <rguenther@suse.de>
19727 PR tree-optimization/57396
19728 * tree-affine.c (double_int_constant_multiple_p): Properly
19729 return false for val == 0 and div != 0.
19731 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19733 * config/mips/mips.h: Use #elif in preprocessor conditions.
19735 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19738 * config/mips/constraints.md (kl): New constraint.
19739 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19740 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19741 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19742 from LO for MIPS16.
19743 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19745 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19748 * config/mips/mips.c (mips_can_inline_p): New function.
19749 (TARGET_CAN_INLINE_P): Define.
19751 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19753 * sched-int.h (ds_t, dw_t): Make unsigned int.
19754 Fix documentation that describes how all the ds_t bits are used.
19755 Reserve the last bit for delayed-branch scheduling.
19756 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19757 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19758 (gen_dep_weak_1): Remove prototype.
19759 * sched-deps.c (get_dep_weak_1): Make static.
19760 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19761 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19762 * doc/tm.texi: Regenerate.
19763 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19765 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19768 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19770 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19771 Sandra Loosemore <sandra@codesourcery.com>
19773 * config.gcc (powerpc-*): Allow native for with-cpu.
19775 2013-05-24 Jeff Law <law@redhat.com>
19777 PR tree-optimization/57124
19778 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19779 conversion feeding a condition if the range has an overflow
19780 if -fstrict-overflow. Add warnings for when we do make the
19783 2013-05-24 Dehao Chen <dehao@google.com>
19785 * tree-cfg.c (locus_discrim_map): Fix the typo.
19786 (locus_discrim_hasher): Likewise.
19787 (locus_discrim_hasher::hash): Likewise.
19788 (locus_discrim_hasher::equal): Likewise.
19790 2013-05-24 Martin Jambor <mjambor@suse.cz>
19792 PR tree-optimization/57294
19793 * cgraph.h (ipa_record_stmt_references): Declare.
19794 * cgraphbuild.c (ipa_record_stmt_references): New function.
19795 (build_cgraph_edges): Use ipa_record_stmt_references.
19796 (rebuild_cgraph_edges): Likewise.
19797 (cgraph_rebuild_references): Likewise.
19798 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19799 associated with the old statement and build references from the
19800 newly built statements.
19801 * ipa-ref.c (ipa_remove_stmt_references): New function.
19802 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19804 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19806 * lra-constraints.c (emit_spill_move): Use smaller mode for
19808 (check_and_process_move): Consider mem-reg moves for secondary
19810 (curr_insn_transform): Don't lose insns emitted before for
19811 secondary memory moves.
19812 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19813 reg set up in the current insn.
19815 2013-05-24 Dehao Chen <dehao@google.com>
19817 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19819 (locus_descrim_hasher::equal): Likewise.
19820 (build_gimple_cfg): New discriminator assignment algorithm.
19821 (make_edges): Likewise.
19822 (next_discriminator_for_locus): Likewise.
19823 (same_line_p): Likewise.
19824 (assign_discriminators): Likewise.
19825 (make_cond_expr_edges): Likewise.
19826 (make_gimple_switch_edges): Likewise.
19827 (make_goto_expr_edges): Likewise.
19828 (make_gimple_asm_edges): Likewise.
19830 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19832 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19833 X format specifier to only display bottom 16 bits.
19834 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19835 immediate to match for operand 2, since it will be masked.
19837 2013-05-24 Richard Biener <rguenther@suse.de>
19839 PR tree-optimization/57287
19840 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19841 all SSA names that occur in abnormal PHIs.
19843 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19846 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19847 that index is not negative.
19849 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19851 PR rtl-optimization/55177
19852 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19853 (simplify_byte_swapping_operation): New.
19854 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19855 (simplify_relational_operation_1): Deal with BSWAP.
19857 2013-05-23 Richard Henderson <rth@redhat.com>
19860 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19861 (ix86_reorg): Call it.
19863 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19866 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19867 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19868 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19870 2013-05-23 Christian Bruel <christian.bruel@st.com>
19873 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19875 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19876 Marcus Shawcroft <marcus.shawcroft@arm.com>
19878 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19879 * config/aarch64/constraints.md (Usa): Remove.
19880 * doc/md.texi (AArch64 Usa): Remove.
19882 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19883 Marcus Shawcroft <marcus.shawcroft@arm.com>
19885 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19886 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19887 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19888 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19890 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19892 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19893 instruction (AdvSIMD).
19894 * config/aarch64/aarch64-builtins.c
19895 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19896 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19898 2013-05-23 Martin Jambor <mjambor@suse.cz>
19900 PR middle-end/57347
19901 * tree.h (contains_bitfld_component_ref_p): Declare.
19902 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19903 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19905 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19906 not access a bit-field. Assert all final offsets are byte-aligned.
19908 2013-05-23 Richard Biener <rguenther@suse.de>
19910 PR tree-optimization/57380
19911 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19912 least one invariant or re-used load.
19913 * passes.c (init_optimization_passes): Move pass_phiprop before
19916 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19918 * config/aarch64/aarch64-simd.md
19919 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19921 2013-05-23 Richard Biener <rguenther@suse.de>
19923 PR middle-end/57381
19924 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19925 OEP_CONSTANT_ADDRESS_OF retained.
19927 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19929 PR middle-end/57344
19930 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19931 don't lower unit. Handle unit not being always BITS_PER_WORD.
19933 2013-05-23 Richard Biener <rguenther@suse.de>
19935 PR rtl-optimization/57341
19936 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19937 instead of true_dependence.
19939 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19941 * bb-reorder.c (branch_threshold): Make const.
19942 (exec_threshold): Ditto.
19944 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19945 Pat Haugen <pthaugen@us.ibm.com>
19946 Peter Bergner <bergner@vnet.ibm.com>
19948 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19949 documentation for the power8 crypto builtins.
19951 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19953 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19954 macros for defining power8 builtin functions.
19955 (BU_P8V_AV_2): Likewise.
19956 (BU_P8V_AV_P): Likewise.
19957 (BU_P8V_VSX_1): Likewise.
19958 (BU_P8V_OVERLOAD_1): Likewise.
19959 (BU_P8V_OVERLOAD_2): Likewise.
19960 (BU_CRYPTO_1): Likewise.
19961 (BU_CRYPTO_2): Likewise.
19962 (BU_CRYPTO_3): Likewise.
19963 (BU_CRYPTO_OVERLOAD_1): Likewise.
19964 (BU_CRYPTO_OVERLOAD_2): Likewise.
19965 (XSCVSPDP): Fix typo, point to the correct instruction.
19966 (VCIPHER): Add power8 crypto builtins.
19967 (VCIPHERLAST): Likewise.
19968 (VNCIPHER): Likewise.
19969 (VNCIPHERLAST): Likewise.
19970 (VPMSUMB): Likewise.
19971 (VPMSUMH): Likewise.
19972 (VPMSUMW): Likewise.
19973 (VPERMXOR_V2DI): Likewise.
19974 (VPERMXOR_V4SI: Likewise.
19975 (VPERMXOR_V8HI: Likewise.
19976 (VPERMXOR_V16QI: Likewise.
19977 (VSHASIGMAW): Likewise.
19978 (VSHASIGMAD): Likewise.
19979 (VPMSUM): Likewise.
19980 (VPERMXOR): Likewise.
19981 (VSHASIGMA): Likewise.
19983 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19984 __CRYPTO__ if the crypto instructions are available.
19985 (altivec_overloaded_builtins): Add support for overloaded power8
19988 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19989 support for power8 crypto builtins.
19990 (builtin_function_type): Likewise.
19991 (altivec_init_builtins): Add support for builtins that take vector
19992 long long (V2DI) arguments.
19994 * config/rs6000/crypto.md: New file, define power8 crypto
19997 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19998 Pat Haugen <pthaugen@us.ibm.com>
19999 Peter Bergner <bergner@vnet.ibm.com>
20001 * doc/invoke.texi (Option Summary): Add power8 options.
20002 (RS/6000 and PowerPC Options): Likewise.
20004 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
20005 constraints.md instead of rs6000.h. Reorder w* constraints. Add
20006 wm, wn, wr documentation.
20008 * config/rs6000/constraints.md (wm): New constraint for VSX
20009 registers if direct move instructions are enabled.
20010 (wn): New constraint for no registers.
20011 (wq): New constraint for quad word even GPR registers.
20012 (wr): New constraint if 64-bit instructions are enabled.
20013 (wv): New constraint if power8 vector instructions are enabled.
20014 (wQ): New constraint for quad word memory locations.
20016 * config/rs6000/predicates.md (const_0_to_15_operand): New
20017 constraint for 0..15 for crypto instructions.
20018 (gpc_reg_operand): If VSX allow registers in VSX registers as well
20019 as GPR and floating point registers.
20020 (int_reg_operand): New predicate to match only GPR registers.
20021 (base_reg_operand): New predicate to match base registers.
20022 (quad_int_reg_operand): New predicate to match even GPR registers
20023 for quad memory operations.
20024 (vsx_reg_or_cint_operand): New predicate to allow vector logical
20025 operations in both GPR and VSX registers.
20026 (quad_memory_operand): New predicate for quad memory operations.
20027 (reg_or_indexed_operand): New predicate for direct move support.
20029 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
20030 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
20031 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
20032 (POWERPC_MASKS): Add power8 options.
20033 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
20036 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
20037 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
20039 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
20040 (-mpower8-fusion): New power8 options.
20041 (-mpower8-fusion-sign): Likewise.
20042 (-mpower8-vector): Likewise.
20043 (-mcrypto): Likewise.
20044 (-mdirect-move): Likewise.
20045 (-mquad-memory): Likewise.
20047 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
20048 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
20050 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
20051 (rs6000_debug_vector_unit): Add p8_vector.
20052 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
20053 definitions. Also print fusion state.
20054 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
20055 (rs6000_builtin_mask_calculate): Add power8 builtin support.
20056 (rs6000_option_override_internal): Add support for power8.
20057 (rs6000_common_init_builtins): Add debugging for skipped builtins
20058 if -mdebug=builtin.
20059 (rs6000_adjust_cost): Add power8 support.
20060 (rs6000_issue_rate): Likewise.
20061 (insn_must_be_first_in_group): Likewise.
20062 (insn_must_be_last_in_group): Likewise.
20063 (force_new_group): Likewise.
20064 (rs6000_register_move_cost): Likewise.
20065 (rs6000_opt_masks): Likewise.
20067 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
20068 power8 capable assembler, default to power7 options.
20069 (TARGET_DIRECT_MOVE): Likewise.
20070 (TARGET_CRYPTO): Likewise.
20071 (TARGET_P8_VECTOR): Likewise.
20072 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
20073 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
20074 (VECTOR_MEM_P8_VECTOR_P): Likewise.
20075 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
20076 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
20077 (TARGET_XSCVDPSPN): Likewise.
20078 (TARGET_XSCVSPDPN): Likewsie.
20079 (TARGET_SYNC_HI_QI): Likewise.
20080 (TARGET_SYNC_TI): Likewise.
20081 (MASK_CRYPTO): Likewise.
20082 (MASK_DIRECT_MOVE): Likewise.
20083 (MASK_P8_FUSION): Likewise.
20084 (MASK_P8_VECTOR): Likewise.
20085 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
20086 temporary used by some of the direct move instructions to get two FP
20087 temporary registers does not force creation of a stack frame.
20088 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
20089 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
20090 that any VSX registers are tieable, even if they are also an
20091 Altivec vector mode.
20092 (r6000_reg_class_enum): Add wm, wr, wv constraints.
20093 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
20094 (RS6000_BTM_CRYPTO): Likewise.
20095 (RS6000_BTM_COMMON): Likewise.
20097 * config/rs6000/rs6000.md (cpu attribute): Add power8.
20098 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
20099 (enum rs6000_vector): Add power8 vector support.
20101 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20105 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
20106 (any_sibcall_could_use_r3): this and handle indirect calls.
20107 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
20109 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20111 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
20113 2013-05-22 Richard Biener <rguenther@suse.de>
20115 PR middle-end/57349
20116 * profile.c (branch_prob): Do not split blocks that are
20117 abnormally receiving from ECF_RETURNS_TWICE functions.
20119 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20121 * recog.c (offsettable_address_addr_space_p): Fix calculation of
20122 address mode. Move pointer mode initialization to the same place.
20124 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
20126 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
20127 while it has any effect.
20129 2013-05-21 Easwaran Raman <eraman@google.com>
20131 PR tree-optimization/57322
20132 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
20133 UID of the statement added to the BB to be 1.
20135 2013-05-21 Jakub Jelinek <jakub@redhat.com>
20137 PR tree-optimization/57331
20138 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
20139 of conversion from pointer type to integral type with integer.
20141 2013-05-21 Martin Jambor <mjambor@suse.cz>
20144 * ipa-prop.c (ipa_read_node_info): Process param_used and
20145 controlled_uses in the same order as when writing.
20147 2013-05-21 Magnus Granberg <baldrick@free.fr>
20150 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
20152 2013-05-21 Richard Biener <rguenther@suse.de>
20154 PR tree-optimization/57318
20155 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
20156 estimate stmts with side-effects as likely eliminated.
20158 2013-05-21 Richard Biener <rguenther@suse.de>
20160 PR tree-optimization/57330
20161 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
20162 preserve the call stmts fntype.
20164 2013-05-21 Richard Biener <rguenther@suse.de>
20166 PR tree-optimization/57303
20167 * tree-ssa-sink.c (statement_sink_location): Improve killing
20168 stmt detection and properly handle self-assignments.
20170 2013-05-21 Christian Bruel <christian.bruel@st.com>
20172 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
20173 spanning registers. LEAF_REG_REMAP is supported only for contiguous
20174 registers. Set register size out of the PARALLEL loop.
20176 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
20179 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
20180 (*fmasf4, *fmasf4_media): New insns.
20182 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20184 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
20185 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
20186 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
20187 (mips_idiv_insns): Update the comments to say that the returned
20188 instruction counts are in units of BASE_INSN_LENGTH.
20189 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
20190 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
20191 using 2 rather than 4 as the length of indirect MIPS16 and
20192 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
20193 length of a NOP. Don't divide MIPS16 lengths by 2.
20194 (mips16_split_long_branches): Assume a branch is long if the
20195 length is greater than 4 rather than 8.
20196 * config/mips/mips.md (length): Give MIPS16 lengths directly,
20197 rather than multiplying them by 2. Multiply instruction counts
20198 by BASE_INSN_LENGTH rather than 4.
20199 (*jump_mips16, tls_get_tp_mips16_<mode>)
20200 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
20202 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20204 * config/mips/mips.md (extended_mips16): Remove branch case.
20205 (length): Remove duplicated extended_mips16 test.
20207 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20209 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
20211 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
20213 * recog.h (Recog_data): Rename to...
20214 (recog_data_d): ...this.
20215 (recog_data): Update accordingly.
20216 * recog.c (recog_data): Likewise.
20217 * reload.c (save_recog_data): Likewise.
20218 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
20219 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
20221 2013-05-17 Julian Brown <julian@codesourcery.com>
20223 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
20224 found in a REG_EQUAL note, invalidate it.
20226 2013-05-17 Easwaran Raman <eraman@google.com>
20228 * tree-ssa-reassoc.c (find_insert_point): New function.
20229 (insert_stmt_after): Likewise.
20230 (get_def_stmt): Likewise.
20231 (ensure_ops_are_available): Likewise.
20232 (not_dominated_by): Likewise.
20233 (rewrite_expr_tree): Do not move statements beyond what is
20234 necessary. Remove call to swap_ops_for_binary_stmt...
20235 (reassociate_bb): ... and move it here.
20236 (build_and_add_sum): Assign UIDs for new statements.
20237 (linearize_expr): Likewise.
20238 (do_reassoc): Renumber gimple statement UIDs.
20240 2013-05-17 Jan Hubicka <jh@suse.cz>
20242 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
20244 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
20245 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
20247 (output_weakrefs): Update.
20249 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
20250 Martin Jambor <mjambor@suse.cz>
20252 PR middle-end/57276
20253 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
20254 value that corresponds to the given aggval is found in values vector.
20256 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
20258 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
20259 sse, sse2, sse3, ssse3 and sse4a flags to options.
20261 2013-05-17 David Malcolm <dmalcolm@redhat.com>
20263 * gengtype-state.c: (s_expr_writer): New class, to handle
20264 prettifying of output layout of s-expressions.
20265 (state_writer): New class, to write out gtype.state.
20266 (state_written_type_count): Move this variable into member data of
20268 (s_expr_writer::s_expr_writer): New code: constructor for new class
20269 (state_writer::state_writer(): ditto
20270 (s_expr_writer::write_new_line): New function
20271 (s_expr_writer::write_any_indent): ditto
20272 (s_expr_writer::begin_s_expr): ditto
20273 (s_expr_writer::end_s_expr): ditto
20274 (write_state_fileloc): convert to method of state_writer...
20275 (state_writer:: write_state_fileloc): ...and use methods of
20276 s_expr_writer to write indentation into the gtype.state output file
20277 to visually represent the hierarchical structure of the list
20279 (write_state_fields): ditto, renaming to...
20280 (state_writer::write_state_fields)
20281 (write_state_a_string): ditto, renaming to...
20282 (state_writer::write_state_a_string)
20283 (write_state_string_option): ditto, renaming to...
20284 (state_writer::write_state_string_option)
20285 (write_state_type_option): ditto, renaming to...
20286 (state_writer::write_state_type_option)
20287 (write_state_nested_option): ditto, renaming to...
20288 (state_writer::write_state_nested_option)
20289 (write_state_option): ditto, renaming to...
20290 (state_writer::write_state_option)
20291 (write_state_options): ditto, renaming to...
20292 (state_writer::write_state_options)
20293 (write_state_lang_bitmap): ditto, renaming to...
20294 (state_writer::write_state_lang_bitmap)
20295 (write_state_version): ditto, renaming to...
20296 (state_writer::write_state_version)
20297 (write_state_scalar_type): ditto, renaming to...
20298 (state_writer::write_state_scalar_type)
20299 (write_state_string_type): ditto, renaming to...
20300 (state_writer::write_state_string_type)
20301 (write_state_undefined_type): ditto, renaming to...
20302 (state_writer::write_state_undefined_type)
20303 (write_state_struct_union_type): ditto, renaming to...
20304 (state_writer::write_state_struct_union_type)
20305 (write_state_struct_type): ditto, renaming to...
20306 (state_writer::write_state_struct_type)
20307 (write_state_user_struct_type): ditto, renaming to...
20308 (state_writer::write_state_user_struct_type)
20309 (write_state_lang_struct_type): ditto, renaming to...
20310 (state_writer::write_state_lang_struct_type)
20311 (write_state_param_struct_type): ditto, renaming to...
20312 (state_writer::write_state_param_struct_type)
20313 (write_state_pointer_type): ditto, renaming to...
20314 (state_writer::write_state_pointer_type)
20315 (write_state_array_type): ditto, renaming to...
20316 (state_writer::write_state_array_type)
20317 (write_state_gc_used): ditto, renaming to...
20318 (state_writer::write_state_gc_used)
20319 (write_state_common_type_content): ditto, renaming to...
20320 (state_writer::write_state_common_type_content)
20321 (write_state_type): ditto, renaming to...
20322 (state_writer::write_state_type)
20323 (write_state_pair_list): ditto, renaming to...
20324 (state_writer::write_state_pair_list)
20325 (write_state_pair): ditto, renaming to...
20326 (state_writer::write_state_pair)
20327 (write_state_typedefs): ditto, renaming to...
20328 (state_writer::write_state_typedefs)
20329 (write_state_structures): ditto, renaming to...
20330 (state_writer::write_state_structures)
20331 (write_state_param_structs): ditto, renaming to...
20332 (state_writer::write_state_param_structs)
20333 (write_state_variables): ditto, renaming to...
20334 (state_writer::write_state_variables)
20335 (write_state_srcdir): ditto, renaming to...
20336 (state_writer::write_state_srcdir)
20337 (write_state_files_list): ditto, renaming to...
20338 (state_writer::write_state_files_list)
20339 (write_state_languages): ditto, renaming to...
20340 (state_writer::write_state_languages)
20341 (write_state): create a state_writer instance and use it when
20342 writing out the state file
20344 2013-05-17 Mike Stump <mikestump@comcast.net>
20346 PR rtl-optimization/57304
20347 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20348 accessing DF_REF_REAL_LOC.
20350 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20352 PR rtl-optimization/57281
20353 PR rtl-optimization/57300
20354 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20355 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20356 what the other splitter did if the registers are dead.
20358 2013-05-17 Richard Biener <rguenther@suse.de>
20360 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20363 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20365 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20368 2013-05-17 Marek Polacek <polacek@redhat.com>
20370 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20371 length when doing non-zero store of storing '\0' to '\0'.
20373 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20375 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20376 vect_external_def oprnd1 with loop_vinfo, try to emit
20377 optional cast, negation and and stmts on the loop preheader
20378 edge instead of into the pattern def seq.
20380 PR tree-optimization/57051
20381 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20382 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20384 2013-05-16 Nick Clifton <nickc@redhat.com>
20386 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20387 (rl78_is_naked_func): New function.
20388 (rl78_expand_prologue): Skip prologue generation for naked functions.
20389 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20390 * doc/extend.texi (naked): Add RL78 to the list of processors
20391 that supports this attribute.
20393 2013-05-16 Jeff Law <law@redhat.com>
20395 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20397 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20399 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20400 cache parameters using detect_caches_amd also for CYRIX,
20401 NSC and TM2 signatures.
20403 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20404 Dzianis Kahanovich <mahatma@eu.by>
20408 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20409 VIA/Centaur processors and determine their cache parameters
20410 using detect_caches_amd.
20412 2013-05-16 Teresa Johnson <tejohnson@google.com>
20414 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20415 (rtl_verify_edges): New function.
20416 (rtl_verify_bb_insns): Ditto.
20417 (rtl_verify_bb_pointers): Ditto.
20418 (rtl_verify_bb_insn_chain): Ditto.
20419 (rtl_verify_fallthru): Ditto.
20420 (rtl_verify_bb_layout): Ditto.
20421 (rtl_verify_flow_info_1): Outline checks into new functions.
20422 (rtl_verify_flow_info): Ditto.
20424 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20426 * cfghooks.c (copy_bbs): Add update_dominance argument.
20427 * cfghooks.h (copy_bbs): Update prototype.
20428 * tree-cfg.c (gimple_duplicate_sese_region):
20429 Add update_dominance argument.
20430 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20431 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20432 gimple_duplicate_sese_region call.
20433 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20434 Update copy_bbs call.
20435 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20436 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20438 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20440 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20441 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20442 vect_recog_rotate_pattern.
20443 (vect_recog_rotate_pattern): New function.
20445 2013-05-16 Jason Merrill <jason@redhat.com>
20447 * Makefile.in (LLINKER): New variable.
20448 (mostlyclean): Remove link mutex.
20449 * configure.ac: Handle --enable-link-mutex.
20450 * lock-and-run.sh: New script.
20452 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20455 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20458 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20460 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20462 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20464 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20465 * config/arm/arm.c (next_consecutive_mem): New function.
20466 (gen_movmem_ldrd_strd): Likewise.
20467 * config/arm/arm.md (movmemqi): Update condition and code.
20468 (unaligned_loaddi, unaligned_storedi): New patterns.
20470 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20472 * config.gcc: Obsolete *-*-solaris2.9*.
20473 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20475 2013-05-16 Richard Biener <rguenther@suse.de>
20477 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20478 earlier, after GRAPHITE transforms and IV canonicalization.
20480 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20482 * omp-low.c (extract_omp_for_data): For collapsed loops,
20483 if at least one of the loops is known at compile time to
20484 iterate zero times, set count to 0.
20485 (expand_omp_regimplify_p): New function.
20486 (expand_omp_for_generic): For collapsed loops, if at least
20487 one of the loops isn't known to iterate at least once,
20488 add runtime check with setting count to 0.
20489 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20490 For unsigned types if it isn't known at compile time that
20491 the loop will iterate at least once, add runtime check to bypass
20492 the whole loop if initial condition isn't true.
20494 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20496 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20498 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20500 PR middle-end/57286
20501 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20502 transformations to avoid an infinite loop.
20504 2013-05-16 Marek Polacek <polacek@redhat.com>
20506 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20508 2013-05-15 Leif Ekblad <leif@rdos.net>
20510 * config/i386/i386.c (ix86_decompose_address): Use
20511 DEFAULT_TLS_SEG_REG to access TLS segment register.
20512 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20513 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20514 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20516 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20519 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20520 sibling calls to functions that would normally be lazily bound,
20521 unless $gp is call-clobbered.
20523 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20525 * config/i386/i386.c (ix86_option_override_internal): Update
20526 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20527 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20528 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20529 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20531 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20533 2013-05-15 Andreas Schwab <schwab@suse.de>
20535 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20536 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20539 2013-05-15 Teresa Johnson <tejohnson@google.com>
20541 * loop-unroll.c (report_unroll_peel): Check decision before
20542 emitting unroll/peel message.
20544 2013-05-15 Teresa Johnson <tejohnson@google.com>
20546 * function.h (has_bb_partition): New rtl_data flag.
20547 (bb_reorder_complete): Ditto.
20548 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20549 instead of flag_reorder_blocks_and_partition.
20550 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20551 with some enhancements.
20552 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20553 * bb-reorder.c (connect_traces): Check for has_bb_partition
20554 instead of flag_reorder_blocks_and_partition.
20555 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20556 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20557 verify_hot_cold_block_grouping.
20558 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20560 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20563 * config/arm/predicates.md (call_insn_operand): New predicate.
20564 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20565 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20566 if insn is not a tail call.
20567 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20569 * config/arm/arm.h (enum reg_class): New caller save register class.
20570 (REG_CLASS_NAMES): Likewise.
20571 (REG_CLASS_CONTENTS): Likewise.
20572 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20575 2013-05-15 Richard Biener <rguenther@suse.de>
20577 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20578 of MSG_OPTIMIZED_LOCATIONS.
20579 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20580 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20582 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20583 of MSG_OPTIMIZED_LOCATIONS.
20584 (execute_vect_slp): Likewise.
20585 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20586 (vect_create_cond_for_alias_checks): Likewise.
20587 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20588 (vect_recog_widen_mult_pattern): Likewise.
20589 (vect_recog_widen_sum_pattern): Likewise.
20590 (vect_recog_over_widening_pattern): Likewise.
20591 (vect_recog_widen_shift_pattern): Likewise.
20592 (vect_recog_vector_vector_shift_pattern): Likewise.
20593 (vect_recog_divmod_pattern): Likewise.
20594 (vect_recog_mixed_size_cond_pattern): Likewise.
20595 (vect_recog_bool_pattern): Likewise.
20596 (vect_pattern_recog_1): Likewise.
20598 2013-05-15 Martin Jambor <mjambor@suse.cz>
20600 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20601 non-functions to builtin_unreachable.
20602 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20603 correct when new direct edges were discovered.
20605 2013-05-15 Martin Jambor <mjambor@suse.cz>
20607 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20608 header, print symbol order instead of node uid, print more information
20609 about indirect edge targets.
20610 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20612 (ipa_make_edge_direct_to_target): Likewise.
20613 (remove_described_reference): Likewise.
20614 (propagate_controlled_uses): Likewise.
20615 (ipa_print_node_params): Also print symbol order.
20616 (ipcp_transform_function): Print symbol order instead of node uids.
20617 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20618 (cgraph_get_create_real_symbol_node): Likewise.
20619 * ipa-cp.c (print_lattice): Likewise.
20620 (print_all_lattices): Likewise.
20621 (determine_versionability): Likewise.
20622 (initialize_node_lattices): Likewise.
20623 (estimate_local_effects): Likewise.
20624 (update_profiling_info): Likewise.
20625 (create_specialized_node): Likewise.
20626 (perhaps_add_new_callers): Likewise.
20627 (decide_about_value): Likewise.
20628 (decide_whether_version_node): Likewise.
20629 (identify_dead_nodes): Likewise.
20630 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20631 (dump_inline_summary): Likewise.
20632 (estimate_node_size_and_time): Likewise.
20633 (inline_analyze_function): Likewise.
20634 * ipa-inline.c (report_inline_failed_reason): Likewise.
20635 (want_early_inline_function_p): Likewise.
20636 (edge_badness): Likewise.
20637 (update_edge_key): Likewise.
20638 (inline_small_functions): Likewise. Add dumping of order to two other
20640 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20641 instead of node uids.
20642 (propagate_pure_const): Likewise.
20643 (propagate_pure_const): Likewise.
20644 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20645 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20646 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20648 * tree-pretty-print.c (dump_function_header): Likewise.
20649 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20650 Print symbol order instead of node uids.
20652 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20654 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20655 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20657 2013-05-15 Richard Biener <rguenther@suse.de>
20659 PR tree-optimization/57275
20660 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20661 return value for fail to do runtime alias checks for gather loads.
20663 2013-05-15 Jan Hubicka <jh@suse.cz>
20667 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20668 weakrefs are not external.
20669 (lto_symtab_merge_decls): Fix thinko when dealing with
20670 non-lto_symtab decls.
20671 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20672 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20673 * varpool.c (dump_varpool_node): Dump more flags.
20675 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20677 * config/i386/i386.c (processor_alias_table): Add instruction
20678 FSGSBASE for AMD bdver3 architecture.
20680 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20682 * tree.c (warn_deprecated_use): Print file:line using locus color.
20683 * diagnostic.c (diagnostic_report_current_module): Print file:line
20684 and file:line:column using locus color.
20686 2013-05-14 Mike Stump <mikestump@comcast.net>
20688 * gdbinit.in: Add __null.
20690 2013-05-14 Mike Stump <mikestump@comcast.net>
20692 * recog.h: Rename struct recog_data to Recog_data.
20693 * recog.c: Likewise.
20694 * reload.c (can_reload_into): Likewise.
20695 * config/picochip/picochip.c: Likewise.
20697 2013-05-14 Mike Stump <mikestump@comcast.net>
20699 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20701 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20703 * resource.h (struct resources): Remove unch_memory member.
20704 (CLEAR_RESOURCE): Don't clear unch_memory.
20705 * resource.c (mark_referenced_resources): Don't set it.
20706 (mark_set_resources): Likewise.
20707 (mark_target_live_regs): Don't clear it.
20708 (init_resource_info): Likewise.
20709 * reorg.c (resource_conflicts_p): Don't compare it.
20710 (redundant_insn): Don't set it.
20712 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20714 * emit-rtl.c (next_label): Remove unused function.
20715 (skip_consecutive_labels, link_cc0_insns): Move to ...
20716 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20717 only place where these functions are used, and make them static.
20719 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20721 * fold-const.c (fold_negate_expr): Handle vectors.
20722 (fold_truth_not_expr): Make it static.
20723 (fold_invert_truthvalue): New static function.
20724 (invert_truthvalue_loc): Handle vectors. Do not call
20725 fold_truth_not_expr directly.
20726 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20727 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20728 (fold_comparison): Handle vector constants.
20729 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20730 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20731 * tree.h (fold_truth_not_expr): Remove declaration.
20733 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20735 * config/aarch64/aarch64-simd.md
20736 (aarch64_vcond_internal<mode>): Rename to...
20737 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20738 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20739 float modes. Clarify all iterator modes.
20740 (vcond<mode><mode>): Use new name for vcond expanders.
20741 (vcond<v_cmp_result><mode>): Likewise.
20742 (vcondu<mode><mode>: Likewise.
20743 * config/aarch64/iterators.md (VDQF_COND): New.
20745 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20748 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20749 variable for the shift amount. Check that we shift by non-negative
20752 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20755 * config/arm/arm.h (EPILOGUE_USES): Only return true
20756 for LR_REGNUM after epilogue_completed.
20758 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20760 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20761 is error_mark_node.
20763 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20766 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20767 and Solaris 11+/x86 with gld.
20768 * configure: Regenerate.
20770 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20772 * expmed.c (expand_shift_1): Canonicalize rotates by
20773 constant bitsize / 2 to bitsize - 1.
20774 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20775 case ROTATERT>: Likewise.
20778 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20780 * config/i386/i386.md (rotateinv): New code attr.
20781 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20782 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20783 roll $31, %eax, etc.
20785 2013-05-14 Richard Biener <rguenther@suse.de>
20787 PR middle-end/57235
20788 * tree-eh.c (sink_clobbers): Give up for successors with
20789 multiple predecessors and no virtual uses.
20791 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20793 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20794 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20796 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20798 PR middle-end/57251
20799 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20800 the case when both op0 and op1 have VOIDmode.
20802 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20804 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20805 in multiply-accumulate mode.
20807 2013-05-13 Guozhi Wei <carrot@google.com>
20809 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20811 2013-05-13 Kai Tietz <ktietz@redhat.com>
20814 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20815 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20816 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20817 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20818 (ix86_expand_prologue): Likewise.
20819 (ix86_expand_split_stack_prologue): Likewise.
20820 (legitimate_pic_address_disp_p): Likewise.
20821 (legitimize_pic_address): Likewise.
20822 (legitimize_tls_address): Likewise.
20823 (legitimize_pe_coff_symbol): Likewise.
20824 (output_pic_addr_const): Likewise.
20825 (construct_plt_address): Likewise.
20826 (ix86_expand_call): Likewise.
20827 (x86_output_mi_thunk): Likewise.
20828 (x86_function_profiler): Likewise.
20830 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20832 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20833 similar switch cases.
20834 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20835 (aarch64_simd_mov_to_<mode>low): Delete.
20836 (aarch64_simd_mov_to_<mode>high): Delete.
20837 (move_lo_quad_<mode>): Add w<-r alternative.
20838 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20839 (aarch64_simd_mov_from_*): Update type attribute.
20840 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20843 2013-05-13 Jan Hubicka <jh@suse.cz>
20845 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20846 * config/i386/i386.c (ix86_compute_frame_layout,
20847 ix86_expand_epilogue, emit_i387_cw_initialization,
20848 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20849 ix86_local_alignment): Fix use of size/speed predicates.
20851 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20853 PR tree-optimization/45216
20854 PR tree-optimization/57157
20855 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20856 the (-Y) & (B - 1) variant if OP is |.
20857 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20858 return shifted. Use (-op1) & (prec - 1) as other_amount
20859 instead of prec - op1.
20861 2013-05-13 Martin Jambor <mjambor@suse.cz>
20863 PR middle-end/42371
20864 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20865 (ipa_constant_data): New type.
20866 (ipa_jump_func): Use ipa_constant_data to hold information about
20867 constant jump functions.
20868 (ipa_get_jf_constant): Adjust to jump function type changes.
20869 (ipa_get_jf_constant_rdesc): New function.
20870 (ipa_param_descriptor): New field controlled_uses.
20871 (ipa_get_controlled_uses): New function.
20872 (ipa_set_controlled_uses): Likewise.
20873 * ipa-ref.h (ipa_find_reference): Declare.
20874 * ipa-prop.c (ipa_cst_ref_desc): New type.
20875 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20877 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20878 New parameter cs. Adjust all callers.
20879 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20880 (remove_described_reference): New function.
20881 (jfunc_rdesc_usable): Likewise.
20882 (try_make_edge_direct_simple_call): Decrement controlled use count,
20883 attempt to remove reference if it hits zero.
20884 (combine_controlled_uses_counters): New function.
20885 (propagate_controlled_uses): Likewise.
20886 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20887 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20888 (ipa_print_node_params): Print described use counter.
20889 (ipa_write_jump_function): Adjust to jump function type changes.
20890 (ipa_read_jump_function): New parameter CS, pass it to
20891 ipa_set_jf_constant. Adjust caller.
20892 (ipa_write_node_info): Stream controlled use count
20893 (ipa_read_node_info): Likewise.
20894 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20896 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20897 count. Remove cloning-added reference if it reaches zero.
20898 * ipa-ref.c (ipa_find_reference): New function.
20900 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20902 * config/i386/i386.c (processor_target_table): Modified default
20903 alignment values for AMD BD and BT architectures.
20905 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20907 * tree-vect-generic.c (uniform_vector_p): Move ...
20908 * tree.c (uniform_vector_p): ... here.
20909 * tree.h (uniform_vector_p): Declare it.
20910 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20913 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20915 PR tree-optimization/57230
20916 * tree-ssa-strlen.c (handle_char_store): Record length for
20917 array store from STRING_CST.
20919 PR tree-optimization/57230
20920 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20923 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20925 * config/epiphany/epiphany.c (epiphany_init): Check size of
20926 NUM_MODES_FOR_MODE_SWITCHING.
20927 (epiphany_expand_prologue):
20928 Remove CONFIG_REGNUM initial value handling code.
20929 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20930 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20931 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20932 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20933 Don't return 1 for FP_MODE_NONE.
20934 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20935 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20936 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20937 * config/epiphany/epiphany.md (save_config): New pattern.
20939 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20941 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20943 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20945 * config/i386/i386.md (memory): Handle sseishft1.
20946 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20947 (*vec_extractv2di_1): Ditto.
20949 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20951 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20954 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20956 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20957 Add mthumb/march=armv7-a multilib.
20958 Add mthumb/march=armv7-r multilib.
20959 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20961 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20963 * config/v850/t-rtems: Add more multilibs.
20965 2013-05-10 Richard Biener <rguenther@suse.de>
20967 PR tree-optimization/57214
20968 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20969 not propagate from SSA names that occur in abnormal PHI nodes.
20971 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20973 * stor-layout.c (element_precision): New function.
20974 * machmode.h (element_precision): Declare it.
20975 * tree.c (build_minus_one_cst): New function.
20976 (element_precision): Likewise.
20977 * tree.h (build_minus_one_cst): Declare new function.
20978 (element_precision): Likewise.
20979 * fold-const.c (operand_equal_p): Use element_precision.
20980 (fold_binary_loc): Handle vector types.
20981 * convert.c (convert_to_integer): Use element_precision.
20982 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20985 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20987 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20988 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20989 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20990 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20991 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20992 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20993 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20994 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20995 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20996 (Uuw8): New constraints.
20997 (Usb4): Move into alphabetical order.
20998 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20999 (sd8_operand, ub8_operand, uw8_operand): New predicates.
21000 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
21001 previously unnamed patterns.
21002 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
21003 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
21004 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
21005 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
21006 of set_attr_alternative/if_then_else. Use extended_mips16 instead
21007 of specific lengths.
21009 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21011 * config/i386/i386.md (rotateinv): New code attr.
21012 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21013 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21014 roll $31, %eax, etc.
21016 PR tree-optimization/45216
21017 PR tree-optimization/57157
21018 * tree-ssa-forwprop.c (simplify_rotate): New function.
21019 (ssa_forward_propagate_and_combine): Call it.
21021 2013-05-10 Richard Biener <rguenther@suse.de>
21023 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
21024 disable peeling when we version for aliasing.
21025 (vector_alignment_reachable_p): Honor explicit user alignment.
21026 (vect_supportable_dr_alignment): Likewise.
21027 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
21028 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
21029 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
21030 then peeling to arrange for the cost-model check to come first.
21032 2013-05-10 Alan Modra <amodra@gmail.com>
21034 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
21035 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
21036 * configure: Regenerate.
21038 2013-05-10 Alan Modra <amodra@gmail.com>
21041 * varasm.c (default_elf_select_section): Move !DECL_P check..
21042 (get_named_section): ..to here before calling get_section_name.
21044 (default_section_type_flags): Add DECL_P check.
21045 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
21046 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
21048 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
21050 * config/epiphany/epiphany.c (epiphany_expand_prologue):
21051 When using gen_stack_adjust_str with a register offset, add a
21052 REG_FRAME_RELATED_EXPR note.
21054 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
21056 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
21057 (*vec_extractv4si_zext_mem): Ditto.
21058 (*vec_extractv2di): Add 0->x and x->x alternatives.
21059 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
21060 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
21062 2013-05-09 Jason Merrill <jason@redhat.com>
21064 N3639 C++1y VLA support
21065 * gimplify.c (gimplify_vla_decl): Don't touch an existing
21068 * tree.c (build_constructor_va): New.
21069 * tree.h: Declare it.
21071 2013-05-09 Martin Jambor <mjambor@suse.cz>
21074 * gimple-fold.c (canonicalize_constructor_val): Call
21075 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
21077 2013-05-09 Jan Hubicka <jh@suse.cz>
21078 Richard Biener <rguenther@suse.de>
21081 * symtab.c (symtab_make_decl_local): Do not add private names.
21083 2013-05-09 Jan Hubicka <jh@suse.cz>
21086 * symtab.c (insert_to_assembler_name_hash): Handle clones.
21087 (unlink_from_assembler_name_hash): Likewise.
21088 (symtab_prevail_in_asm_name_hash, symtab_register_node,
21089 symtab_unregister_node, symtab_initialize_asm_name_hash,
21090 change_decl_assembler_name): Update.
21092 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
21094 * config/aarch64/aarch64.md: New movtf split.
21095 (*movtf_aarch64): Update.
21096 (aarch64_movdi_tilow): Handle TF modes and rename to
21097 aarch64_movdi_<mode>low.
21098 (aarch64_movdi_tihigh): Handle TF modes and rename to
21099 aarch64_movdi_<mode>high
21100 (aarch64_movtihigh_di): Handle TF modes and rename to
21101 aarch64_mov<mode>high_di
21102 (aarch64_movtilow_di): Handle TF modes and rename to
21103 aarch64_mov<mode>low_di
21104 (aarch64_movtilow_tilow): Remove spurious whitespace.
21105 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
21107 (aarch64_print_operand): Update.
21109 2013-05-09 Alan Modra <amodra@gmail.com>
21111 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
21113 * configure: Regenerate.
21115 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21117 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
21118 splitter preparation statements.
21119 * config/i386/sse.md (*vec_extract* splitters): Ditto.
21120 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
21123 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21125 * gimple-ssa-strength-reduction.c (count_candidates): Change
21126 return value to int.
21127 (analyze_candidates_and_replace): Change type of length to int.
21129 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21131 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
21132 (*vec_extract<mode>): Use VI12_128 mode iterator.
21133 (*vec_extract<mode>_mem): Ditto.
21134 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
21137 2013-05-08 Diego Novillo <dnovillo@google.com>
21142 2012-08-17 Diego Novillo <dnovillo@google.com>
21145 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
21146 * config.in: Regenerate.
21147 * configure: Regenerate.
21148 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
21150 2013-05-08 Jan Hubicka <jh@suse.cz>
21153 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
21154 * cgraph.h (symtab_node_base): Add unique_name.
21155 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21156 input_overwrite_node, input_varpool_node): Stream unique_name.
21157 * cgraphclones.c (cgraph_create_virtual_clone,
21158 cgraph_function_versioning): Set unique_name.
21159 * ipa.c (function_and_variable_visibility): Set unique_name.
21161 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21163 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
21164 (alloc_cand_and_find_basis): Restrict conditional candidate
21165 processing to CAND_MULTs.
21167 2013-05-08 Jan Hubicka <jh@suse.cz>
21170 lto-symtab.c (lto_symtab_symbol_p): New function.
21171 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
21172 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
21173 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
21174 Skip static symbols.
21176 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
21178 PR tree-optimization/57200
21179 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
21180 Only call inform if the preceding warning_at returns true.
21182 2013-05-07 Han Shen <shenhan@google.com>
21184 * cfgexpand.c (record_or_union_type_has_array_p): New function.
21185 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
21186 * common.opt (fstack-protector-strong): New option.
21187 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
21188 * doc/invoke.texi (Optimization Options): Document
21189 "-fstack-protector-strong".
21190 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
21192 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21194 * config/mips/mips.c (mips_machine_reorg2): Return 0.
21196 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
21198 * ira.c (update_equiv_regs): Add insn having equiv memory even if
21199 it is not lhs of the insn.
21200 (setup_reg_equiv): Remove insn having equiv memory which it is not
21202 * lra-constraints.c (process_address): Try to improve generation
21203 code for address base + disp.
21204 (lra_constraints): Make correct the code for checking insn setting
21205 up backward equivalence. Remove insn only if it is in the init
21207 * lra-eliminations.c (update_reg_eliminate): Change return value.
21208 (lra_eliminate): Use the result.
21210 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
21212 * config/i386/sse.md (ssescalarnummask): New mode attribute.
21213 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
21214 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
21215 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
21216 register target operands.
21217 (*vec_extractv8hi_sse2): New pattern.
21218 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
21219 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
21220 (*vec_extract<mode>_mem): New insn and split pattern.
21222 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
21224 * config/arm/arm.c (arm_asan_shadow_offset): New function.
21225 (TARGET_ASAN_SHADOW_OFFSET): Define.
21226 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
21227 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
21229 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21231 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
21232 (incr_vec_index): Return -1 if increment not found.
21233 (create_add_on_incoming_edge): Assert if increment not found.
21234 (record_increment): Limit number of increments recorded.
21235 (all_phi_incrs_profitable): Return false if an increment not found.
21236 (replace_profitable_candidates): Don't process increments that were
21238 (analyze_candidates_and_replace): Limit size of incr_vec.
21240 2013-05-07 Richard Biener <rguenther@suse.de>
21242 * calls.c (special_function_p): setjmp-like functions are leaf.
21243 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
21244 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
21246 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
21248 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
21249 (aarch64_simd_mov<mode>): New expander.
21250 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
21251 (aarch64_simd_mov_to_<mode>high): Likewise.
21252 (aarch64_simd_mov_from_<mode>low): Likewise.
21253 (aarch64_simd_mov_from_<mode>high): Likewise.
21254 (aarch64_dup_lane<mode>): Update.
21255 (aarch64_dup_lanedi): New instruction pattern.
21256 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
21257 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
21259 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21261 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
21262 (replace_mult_candidate): Remove unnecessary argument; remove
21263 unnecessary parameter from call to introduce_cast_before_cand.
21264 (replace_unconditional_candidate): Remove unnecessary parameter
21265 from call to replace_mult_candidate.
21266 (replace_conditional_candidate): Likewise.
21267 (insert_initializers): Use make_temp_ssa_name.
21268 (introduce_cast_before_cand): Remove unnecessary argument; use
21269 make_temp_ssa_name.
21270 (replace_one_candidate): Remove unnecessary argument; remove
21271 unnecessary parameter from calls to introduce_cast_before_cand.
21272 (replace_profitable_candidates): Remove unnecessary parameters
21273 from calls to replace_one_candidate.
21275 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21277 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
21278 phi def as possibly hiding a basis for a CAND_ADD whose operands
21279 have been commuted in the analysis.
21280 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
21282 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21284 * config/aarch64/aarch64.md
21285 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
21286 shift value between 0-4.
21288 2013-05-07 Richard Biener <rguenther@suse.de>
21290 * double-int.h (rshift): New overload.
21291 * double-int.c (rshift): New function.
21292 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
21293 (create_reference_ops_from_ref): Remove.
21294 (vn_reference_insert): Use shared ops for constructing the
21295 reference and copy it.
21297 2013-05-07 Richard Biener <rguenther@suse.de>
21299 PR middle-end/57190
21300 * tree-eh.c (sink_clobbers): Properly propagate
21301 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21303 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21305 PR tree-optimization/57149
21306 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21307 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21308 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21309 uninit_undefined_value_p instead of ssa_undefined_value_p.
21312 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21313 for modifier == EXPAND_INITIALIZER.
21315 2013-05-07 Anton Blanchard <anton@samba.org>
21317 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21318 for powerpc64 little endian.
21319 * configure: Regenerate.
21321 2013-05-06 Graham Stott <grahams@btinternet.com>
21323 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21324 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21325 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21326 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21328 2013-05-06 Graham Stott <grahams@btinternet.com>
21330 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21331 codes which allow non-lvalues.
21333 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21335 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21336 components are all 1s.
21337 (integer_minus_onep): New function.
21338 * tree.h (integer_minus_onep): Declare it.
21339 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21340 integer_minus_onep instead of integer_all_onesp.
21342 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21345 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21346 variations of these patterns.
21348 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21350 * config/i386/i386.md (isa): Add x64_sse4 member.
21351 (enabled): Handle x64_sse4.
21352 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21353 instruction for 64bit SSE4_1 targets. Update insn attributes.
21354 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21355 instruction for SSE4_1 targets. Update insn attributes.
21356 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21357 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21358 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21360 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21361 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21362 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21364 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21367 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21369 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21371 * final.c (do_assembler_dialects): Don't handle curly braces and
21372 vertical bar escaped by % as dialect delimiters.
21373 (output_asm_insn): Print curly braces and vertical bar if escaped
21374 by % and ASSEMBLER_DIALECT defined.
21375 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21376 * doc/tm.texi: Regenerated.
21378 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21380 * config/mips/mips.c: Include tree-pass.h.
21381 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21382 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21383 (pass_mips_machine_reorg2): New machine specific pass.
21384 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21385 (mips_option_override): Register the new pass.
21386 * rtl.h (cleanup_barriers): Remove prototype.
21387 (dbr_schedule): Likewise.
21388 * jump.c (cleanup_barriers): Make static.
21389 * reorg.c (dbr_schedule): Likewise.
21391 2013-05-06 Richard Biener <rguenther@suse.de>
21393 PR tree-optimization/57185
21394 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21395 SSA names properly.
21397 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21400 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21401 between "first_arg" and "insn", not "last" and "insn".
21403 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21405 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21406 (find_candidates_in_block): Re-enable slsr_process_phi.
21407 (create_phi_basis): Fix double counting of candidate adjustment.
21409 2013-05-06 Richard Biener <rguenther@suse.de>
21411 PR middle-end/57147
21412 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21413 the edge is also fallthru, preserve it and just clear the
21415 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21416 also complex, preserve that and just clear the fallthru flag.
21417 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21418 update virtual operands.
21420 2013-05-06 Alan Modra <amodra@gmail.com>
21422 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21423 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21424 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21425 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21426 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21427 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21429 2013-05-06 Alan Modra <amodra@gmail.com>
21431 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21432 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21433 (DEFAULT_ASM_ENDIAN): Define.
21434 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21435 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21436 Update -K PIC clause from sysv4.h.
21437 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21438 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21440 2013-05-06 Alan Modra <amodra@gmail.com>
21442 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21443 twice for little-endian.
21444 (ashrdi3_no_power, ashrdi3): Support little-endian.
21446 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21449 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21450 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21452 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21453 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21454 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21457 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21458 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21460 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21461 * config/pa/pa.opt: Make mbig-switch a no-op.
21462 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21463 (CASE_VECTOR_MODE): Always return SImode.
21464 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21465 for the !TARGET_BIG_SWITCH case.
21466 * config/pa/pa-linux.h: Likewise.
21467 * config/pa/pa-openbsd.h: Likewise.
21468 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21469 * config/pa/pa.md (short_jump): Remove define_insn.
21470 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21471 (casesi0): Remove define_insn.
21472 (type): Remove btable_branch.
21473 (pa_combine_type): Likewise.
21474 (in_nullified_branch_delay): Likewise.
21475 (in_call_delay): Likewise.
21476 (define_delay): Likewise.
21477 (define_insn_reservation "Z3"): Likewise.
21478 (define_insn_reservation "Z4"): Likewise.
21479 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21480 (pa_adjust_insn_length): Remove adjustment for btable branches.
21481 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21484 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21486 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21487 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21488 Add m->r,x alternatives.
21489 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21490 splitters using SWI48x mode iterator.
21491 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21492 TARGET_64BIT. Add m->x alternative.
21493 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21494 Add o->x alternative. Enable for TARGET_SSE.
21495 (sse_storeq): Remove expander.
21496 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21497 with memory input operand.
21498 (*vec_extractv2di_1 splitter): New.
21499 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21500 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21502 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21504 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21505 (INT_LOWPART): Delete.
21506 (extract_MB): Adjust.
21507 (extract_ME): Adjust.
21508 (print_operand): Adjust.
21510 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21512 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21513 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21514 (reg_or_logical_cint_operand, easy_fp_constant,
21515 logical_const_operand): Delete "CONST_DOUBLE" case.
21516 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21517 "HOST_BITS_PER_WIDE_INT == 64" test.
21518 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21519 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21520 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21521 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21523 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21524 CONST_DOUBLE DImode/VOIDmode case.
21525 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21526 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21527 CONST_DOUBLE VOIDmode case.
21528 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21529 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21530 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21531 Delete CONST_DOUBLE case.
21532 (splitters for mov FMOVE64 const_double): Delete
21533 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21534 "HOST_BITS_PER_WIDE_INT >= 64" test.
21535 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21537 (mov DI const_double): Delete.
21539 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21541 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21542 on op shows all bits zero in mode of a lowpart subreg, return zero.
21544 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21547 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21548 to save TFmode registers and DImode to save TImode registers for
21549 caller save operations.
21550 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21551 mark being partially clobbered since they only use the first
21554 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21555 and TDmode only use the upper 64-bits of each VSX register.
21557 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21559 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21560 (find_candidates_in_block): Disable slsr_process_phi.
21562 2013-05-03 Guozhi Wei <carrot@google.com>
21564 * coverage.c (coverage_obj_init): Move the construction of gcov
21566 (build_init_ctor): ... here.
21568 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21570 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21571 (slsr_cand_d): Redefine def_phi.
21572 (stride_status, phi_adjust_status, count_phis_status): New enums.
21573 (find_phi_def): New.
21574 (find_basis_for_base_expr): New.
21575 (find_basis_for_candidate): Handle hidden bases.
21576 (alloc_cand_and_find_basis): Handle phi candidates.
21577 (slsr_process_phi): New.
21578 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21579 (create_mul_imm_cand): Likewise.
21580 (create_add_ssa_cand): Exclude phi base candidates.
21581 (create_add_imm_cand): Likewise.
21582 (slsr_process_cast): Likewise.
21583 (slsr_process_copy): Likewise.
21584 (find_candidates_in_block): Handle phi candidates.
21585 (dump_candidate): Likewise.
21586 (unconditional_cands): Delete.
21587 (unconditional_cands_with_known_stride_p): Delete.
21588 (phi_dependent_cand_p): New.
21589 (cand_increment): Handle phi-dependent candidates.
21590 (replace_dependent): Delete.
21591 (replace_mult_candidate): New.
21592 (replace_unconditional_candidate): New.
21593 (incr_vec_index): Move to avoid forward reference.
21594 (create_add_on_incoming_edge): New.
21595 (create_phi_basis): New.
21596 (replace_dependents): Delete.
21597 (replace_conditional_candidate): New.
21598 (phi_add_costs): New.
21599 (replace_uncond_cands_and_profitable_phis): New.
21600 (record_increment): Handle phi adjustments.
21601 (record_phi_increments): New.
21602 (record_increments): Handle phi adjustments.
21603 (phi_incr_cost): New.
21604 (lowest_cost_path): Handle phis.
21605 (total_savings): Likewise.
21606 (analyze_increments): Likewise.
21607 (ncd_with_phi): New.
21608 (ncd_of_cand_and_phis): New.
21609 (nearest_common_dominator_for_cands): Handle phi increments.
21610 (all_phi_incrs_profitable): New.
21611 (replace_profitable_candidates): Handle phi-dependent candidates.
21612 (analyze_candidates_and_replace): Likewise.
21614 2013-05-03 Teresa Johnson <tejohnson@google.com>
21617 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21618 do not exceed REG_BR_PROB_BASE.
21620 2013-05-03 Jeff Law <law@redhat.com>
21622 PR tree-optimization/57144
21623 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21624 operand of the condition will bit into the new type when eliminating
21625 a cast feeding a condition.
21627 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21629 PR rtl-optimization/57130
21630 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21631 of COMPARE as in_code to the recursive call if needed.
21633 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21635 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21636 (enabled): Handle new members.
21637 * config/i386/sse.md (*vec_concatv2si): Merge from
21638 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21639 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21641 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21643 PR tree-optimization/57027
21644 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21645 for fnms opportunity, check we got the prerequisite kind
21646 of tree / gimple before using accessor functions.
21648 2013-05-03 Richard Biener <rguenther@suse.de>
21650 * double-int.h (lshift): New overload without precision
21651 and arith argument.
21652 (operator *=, operator +=, operator -=): Move ...
21653 * double-int.c (operator *=, operator +=, operator -=): ... here
21654 and implement more efficiently.
21655 (mul_double_with_sign): Remove.
21656 (lshift_double): Adjust to take unsinged shift argument, push
21657 dispatching code to callers.
21658 (mul_double_wide_with_sign): Add early out for callers that
21659 are not interested in high parts or overflow.
21660 (lshift): New function.
21661 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21662 dispatch code here.
21663 (lrotate, rrotate): Use logical shifts.
21664 * expr.c (get_inner_reference): Use lshift.
21665 * fixed-value.c (do_fixed_divide): Likewise.
21666 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21667 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21668 (indirect_refs_may_alias_p): Likewise.
21669 (stmt_kills_ref_p_1): Likewise.
21671 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21673 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21675 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21677 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21678 scalar form of FABD instruction.
21680 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21682 * lra-constraints.c (process_alt_operands): Add checking alt
21683 number to choose the best alternative.
21685 2013-05-02 Richard Biener <rguenther@suse.de>
21687 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21688 bitmap and its handling.
21689 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21691 2013-05-02 Richard Biener <rguenther@suse.de>
21693 PR middle-end/57140
21694 * tree-inline.c (copy_loops): Properly handle removed loops.
21695 (copy_cfg_body): Mark destination loops for fixup if source
21696 loops needed fixup.
21698 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21701 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21702 generating simple_return for naked functions.
21704 2013-05-02 Martin Jambor <mjambor@suse.cz>
21706 PR middle-end/56988
21707 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21708 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21710 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21711 ipa_agg_replacement_value structures.
21712 (known_aggs_to_agg_replacement_list): Likewise.
21713 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21714 (read_agg_replacement_chain): Likewise.
21715 (ipcp_transform_function): Also check that by_ref flags match.
21717 2013-05-02 Richard Biener <rguenther@suse.de>
21719 * graphds.h (struct graph): Add obstack member.
21720 * graphds.c (new_graph): Initialize obstack and allocate
21722 (add_edge): Allocate edge from the obstack.
21723 (free_graph): Free the obstack instead of all edges and vertices.
21725 2013-05-02 Teresa Johnson <tejohnson@google.com>
21727 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21729 * cfg.c (update_bb_profile_for_threading): Ditto.
21730 * tree-inline.c (copy_bb): Ditto.
21731 (copy_edges_for_bb): Ditto.
21732 (initialize_cfun): Ditto.
21733 (copy_cfg_body): Ditto.
21734 (expand_call_inline): Ditto.
21735 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21736 (estimate_node_size_and_time): Ditto.
21737 (inline_merge_summary): Ditto.
21738 * cgraphclones.c (cgraph_clone_edge): Ditto.
21739 (cgraph_clone_node): Ditto.
21740 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21741 (compute_trg_info): Ditto.
21743 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21745 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21746 S reg when fp attribute set.
21747 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21749 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21751 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21753 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21754 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21755 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21757 2013-05-02 Richard Biener <rguenther@suse.de>
21759 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21760 (struct instantiate_cache_entry): New type.
21761 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21762 (struct instantiate_cache_type): New type.
21763 (set_instantiated_value, get_instantiated_value): Remove.
21764 (get_instantiated_value_entry): New function.
21765 (instantiate_scev_name): Use the new cache and adjust.
21766 (instantiate_scev_poly): Adjust.
21767 (instantiate_scev_binary): Likewise.
21768 (instantiate_array_ref): Likewise.
21769 (instantiate_scev_convert): Likewise.
21770 (instantiate_scev_not): Likewise.
21771 (instantiate_scev_3): Likewise.
21772 (instantiate_scev_2): Likewise.
21773 (instantiate_scev_r): Likewise.
21774 (instantiate_scev): Likewise.
21775 (resolve_mixers): Likewise.
21777 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21780 * lra-constraints.c (best_small_class_operands_num): Remove.
21781 (process_alt_operands): Remove small_class_operands_num. Take
21782 small classes operands into losers and only if the operand is not
21783 matched. Modify debugging output.
21784 (curr_insn_transform): Remove best_small_class_operands_num.
21787 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21789 * config/aarch64/aarch64-builtins.c
21790 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21791 * config/aarch64/aarch64-simd-builtins.def
21792 (reduc_splus_): Add new modes.
21793 (reduc_uplus_): New.
21794 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21795 (reduc_uplus_v4sf): Likewise.
21796 (reduc_splus_v4sf): Likewise.
21797 (aarch64_addv<mode>): Likewise.
21798 (reduc_uplus_<mode>): Likewise.
21799 (reduc_splus_<mode>): Likewise.
21800 (aarch64_addvv2di): Likewise.
21801 (reduc_uplus_v2di): Likewise.
21802 (reduc_splus_v2di): Likewise.
21803 (aarch64_addvv2si): Likewise.
21804 (reduc_uplus_v2si): Likewise.
21805 (reduc_splus_v2si): Likewise.
21806 (reduc_<sur>plus_<mode>): New.
21807 (reduc_<sur>plus_v2di): Likewise.
21808 (reduc_<sur>plus_v2si): Likewise.
21809 (reduc_<sur>plus_v4sf): Likewise.
21810 (aarch64_addpv4sf): Likewise.
21811 * config/aarch64/arm_neon.h
21812 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21813 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21814 add UNSPEC_SADDV, UNSPEC_UADDV.
21816 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21818 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21820 * config/aarch64/arm_neon.h
21821 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21823 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21825 * config/aarch64/aarch64-builtins
21826 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21828 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21830 * config/aarch64/aarch64-simd-builtins.def
21831 (reduc_smax_): New.
21832 (reduc_smin_): Likewise.
21833 (reduc_umax_): Likewise.
21834 (reduc_umin_): Likewise.
21835 (reduc_smax_nan_): Likewise.
21836 (reduc_smin_nan_): Likewise.
21839 (smax): Update for V2SF, V4SF and V2DF modes.
21842 (smin_nan): Likewise.
21843 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21844 (<su><maxmin><mode>3): ...This, refactor.
21845 (s<maxmin><mode>3): New.
21846 (<maxmin_uns><mode>3): Likewise.
21847 (reduc_<maxmin_uns>_<mode>): Refactor.
21848 (reduc_<maxmin_uns>_v4sf): Likewise.
21849 (reduc_<maxmin_uns>_v2si): Likewise.
21850 (aarch64_<fmaxmin><mode>: Remove.
21851 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21853 (vmin<q>_f<32,64>): Likewise.
21854 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21856 (su): Add mappings for smax, smin, umax, umin.
21858 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21859 (FMAXMIN): Rename as...
21860 (FMAXMIN_UNS): ...This.
21862 (fmaxminv): Likewise.
21863 (fmaxmin): Likewise.
21865 (maxmin_uns_op): Likewise.
21867 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21869 * config/aarch64/arm_neon.h
21870 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21871 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21872 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21874 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21876 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21877 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21879 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21881 * config/aarch64/aarch64-simd.md
21882 (vcond<mode>_internal): Handle special cases for constant masks.
21883 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21884 (vcondu<mode><mode>): Likewise.
21885 (vcond<v_cmp_result><mode>): New.
21887 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21889 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21890 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21891 * config/aarch64/aarch64-simd-builtins.def
21892 (cmeq): Update to BUILTIN_VALLDI.
21897 * config/aarch64/arm_neon.h
21898 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21899 to builtins or C as appropriate.
21901 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21903 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21905 (cmhi): Rename to...
21907 * config/aarch64/aarch64-simd.md
21908 (simd_mode): Add SF.
21909 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21910 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21911 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21912 (cstore<mode>_neg): ...This.
21913 * config/aarch64/iterators.md
21915 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21916 (COMPARISONS): New.
21917 (UCOMPARISONS): Likewise.
21918 (optab): Add missing comparisons.
21924 (VCMP_S): Likewise.
21925 (VCMP_U): Likewise.
21926 (V_cmp_result): Add DF, SF modes.
21927 (v_cmp_result): Likewise.
21929 (vmtype): Likewise.
21930 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21932 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21934 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21935 define_insn to define_insn_and_split.
21936 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21937 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21938 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21939 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21940 (thumb2_negscc): Likewise.
21942 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21944 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21946 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21948 * config/arm/thumb2.md: Remove trailing whitespaces.
21950 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21952 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21953 Use gen_int_mode rather than GEN_INT.
21955 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21957 * value-prof.c (stream_in_histogram_value): Remove the strayed
21960 2013-04-30 Richard Biener <rguenther@suse.de>
21962 PR middle-end/57122
21963 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21965 2013-04-30 Richard Biener <rguenther@suse.de>
21967 PR middle-end/57107
21968 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21970 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21972 PR rtl-optimization/56957
21973 PR rtl-optimization/57105
21974 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21975 variable. Use just INSN_UID for determining whether an insn
21976 should be only disconnected from the insn stream.
21977 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21979 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21981 PR tree-optimization/57104
21982 * tsan.c (instrument_expr): Don't instrument accesses to
21983 DECL_HARD_REGISTER VAR_DECLs.
21985 2013-04-30 Richard Biener <rguenther@suse.de>
21987 * function.h (loops_for_fn): New inline function.
21988 (set_loops_for_fn): Likewise.
21989 * cfgloop.h (place_new_loop): Add struct function parameter.
21990 (get_loop): Likewise.
21991 (get_loops): Likewise.
21992 (number_of_loops): Likewise.
21993 (fel_next): Adjust.
21994 (fel_init): Likewise.
21995 * cfg.c (get_loop_copy): Adjust.
21996 * cfgloop.c (flow_loops_dump): Likewise.
21997 (record_loop_exits): Likewise.
21998 (verify_loop_structure): Likewise.
21999 * cfgloopanal.c (mark_irreducible_loops): Likewise.
22000 (estimate_reg_pressure_cost): Likewise.
22001 (mark_loop_exit_edges): Likewise.
22002 * cfgloopmanip.c (place_new_loop): Likewise.
22003 (add_loop): Likewise.
22004 (duplicate_loop): Likewise.
22005 * graph.c (draw_cfg_nodes): Likewise.
22006 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
22007 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
22008 (extract_affine_chrec): Likewise.
22009 (build_scop_iteration_domain): Likewise.
22010 * graphite.c (graphite_initialize): Likewise.
22011 * ira-build.c (create_loop_tree_nodes): Likewise.
22012 (more_one_region_p): Likewise.
22013 (rebuild_regno_allocno_maps): Likewise.
22014 (mark_loops_for_removal): Likewise.
22015 (mark_all_loops_for_removal): Likewise.
22016 (remove_unnecessary_regions): Likewise.
22017 (ira_build): Likewise.
22018 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22019 * loop-init.c (fix_loop_structure): Likewise.
22020 (gate_rtl_move_loop_invariants): Likewise.
22021 (gate_rtl_unswitch): Likewise.
22022 (gate_rtl_unroll_and_peel_loops): Likewise.
22023 (rtl_doloop): Likewise.
22024 * lto-streamer-in.c (input_cfg): Likewise.
22025 * lto-streamer-out.c (output_cfg): Likewise.
22026 * modulo-sched.c (sms_schedule): Likewise.
22027 * predict.c (tree_estimate_probability): Likewise.
22028 (tree_estimate_probability_driver): Likewise.
22029 (estimate_loops): Likewise.
22030 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
22031 (move_sese_region_to_fn): Likewise.
22032 (debug_loop_num): Likewise.
22033 * tree-chrec.c (chrec_evaluate): Likewise.
22034 (hide_evolution_in_other_loops_than_loop): Likewise.
22035 (chrec_component_in_loop_num): Likewise.
22036 (reset_evolution_in_loop): Likewise.
22037 (evolution_function_is_invariant_rec_p): Likewise.
22038 * tree-if-conv.c (main_tree_if_conversion): Likewise.
22039 * tree-inline.c (copy_loops): Likewise.
22040 (copy_cfg_body): Likewise.
22041 (tree_function_versioning): Likewise.
22042 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
22043 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
22045 (add_to_evolution_1): Likewise.
22046 (scev_const_prop): Likewise.
22047 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
22048 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
22049 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
22050 (tree_ssa_lim_initialize): Likewise.
22051 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
22052 (verify_loop_closed_ssa): Likewise.
22053 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
22054 (tree_ssa_loop_im): Likewise.
22055 (tree_ssa_loop_unswitch): Likewise.
22056 (tree_vectorize): Likewise.
22057 (check_data_deps): Likewise.
22058 (tree_ssa_loop_ivcanon): Likewise.
22059 (tree_ssa_loop_bounds): Likewise.
22060 (tree_complete_unroll): Likewise.
22061 (tree_complete_unroll_inner): Likewise.
22062 (tree_parallelize_loops): Likewise.
22063 (tree_ssa_loop_prefetch): Likewise.
22064 (tree_ssa_loop_ivopts): Likewise.
22065 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
22066 * tree-vectorizer.c (vectorize_loops): Likewise.
22068 2013-04-29 Mike Frysinger <vapier@gentoo.org>
22070 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
22071 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
22072 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
22073 with EABI_LINK_SPEC.
22075 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22078 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
22081 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
22084 * lra-constraints.c (process_alt_operands): Discourage a bit more
22085 using memory for pseudos. Print cost dump for alternatives.
22086 Modify cost values for conflicts with early clobbers.
22087 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
22089 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22092 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
22094 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22096 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
22097 from/to S register.
22098 (movdi_aarch64): Support LDR/STR from/to D register.
22100 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22102 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
22103 or higher by default.
22105 2013-04-29 Richard Biener <rguenther@suse.de>
22107 PR middle-end/57075
22108 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
22109 even if not adding abnormal edges for calls that can make
22112 2013-04-29 Richard Biener <rguenther@suse.de>
22114 PR middle-end/57103
22115 * tree-cfg.c (move_stmt_op): Fix condition under which to update
22117 (move_stmt_r): Remove redundant checking.
22119 2013-04-29 Teresa Johnson <tejohnson@google.com>
22122 * basic-block.h (apply_scale): New function.
22123 (apply_probability): Use apply_scale.
22124 * gimple-streamer-in.c (input_bb): Ditto.
22125 * lto-streamer-in.c (input_cfg): Ditto.
22126 * lto-cgraph.c (merge_profile_summaries): Ditto.
22127 * tree-optimize.c (execute_fixup_cfg): Ditto.
22128 * tree-inline.c (copy_bb): Update comment to use apply_scale.
22129 (copy_edges_for_bb): Ditto.
22130 (copy_cfg_body): Ditto.
22132 2013-04-29 Tom de Vries <tom@codesourcery.com>
22134 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
22135 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
22136 (tail_merge_optimize): Handle current_loops == NULL.
22138 2013-04-26 Jeff Law <law@redhat.com>
22140 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
22141 (simplify_cond_using_ranges): Generalize code to simplify
22142 COND_EXPRs where one argument is a constant and the other
22143 is an SSA_NAME created by an integral type conversion.
22145 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22147 * config/arm/arm.md (store_minmaxsi): Use only when
22148 optimize_insn_for_size_p.
22150 2013-04-29 Christian Bruel <christian.bruel@st.com>
22153 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
22155 2013-04-29 Richard Biener <rguenther@suse.de>
22157 PR middle-end/57089
22158 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
22159 loop tree make sure to schedule a fixup for the child as well.
22160 (expand_omp_for_generic): Properly add loops.
22161 (expand_omp_for_static_nochunk): Likewise.
22162 (expand_omp_for_static_chunk): Likewise.
22163 (expand_omp_for): For the degenerate case fixup loops.
22164 (expand_omp_sections): Fix default bb placement in loops.
22165 (expand_omp_atomic_pipeline): Properly add loops.
22167 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22169 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
22171 2013-04-29 Tom de Vries <tom@codesourcery.com>
22173 * tree-ssa-tail-merge.c: Update header comment.
22175 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22177 * config/aarch64/arm_neon.h
22178 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
22179 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
22180 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
22181 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
22182 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
22183 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
22184 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
22185 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
22187 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22189 * config/aarch64/aarch64-simd.md
22190 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
22191 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
22192 fix_trunc, fixuns_trunc.
22193 (ftrunc<VDQF:mode>2): New.
22194 * config/aarch64/iterators.md (optab): Add fix, fixuns.
22195 (fix_trunc_optab): New.
22197 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22199 * config/aarch64/aarch64-builtins.c
22200 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
22201 iceilf, lround, iroundf.
22203 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22206 * config/i386/i386.h (enum ix86_tune_indices)
22207 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
22208 New, split from X86_TUNE_INTER_UNIT_MOVES.
22209 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
22210 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
22211 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
22212 (TARGET_INTER_UNIT_MOVES): Remove.
22213 * config/i386/i386.c (initial_ix86_tune_features): Update.
22214 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
22215 (ix86_expand_convert_uns_didf_sse): Use
22216 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22217 (ix86_expand_vector_init_one_nonzero): Ditto.
22218 (ix86_expand_vector_init_interleave): Ditto.
22219 (inline_secondary_memory_needed): Return true for moves from SSE class
22220 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
22221 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
22222 * config/i386/constraints.md (Yi, Ym): Depend on
22223 TARGET_INTER_UNIT_MOVES_TO_VEC.
22224 (Yj, Yn): New constraints.
22225 * config/i386/i386.md (*movdi_internal): Change constraints of
22226 operand 1 from Yi to Yj and from Ym to Yn.
22227 (*movsi_internal): Ditto.
22228 (*movdf_internal): Ditto.
22229 (*movsf_internal): Ditto.
22230 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
22231 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22232 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
22233 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
22234 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
22235 * config/i386/sse.md (movdi_to_sse): Ditto.
22236 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
22237 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
22238 TARGET_INTER_UNIT_MOVES.
22239 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
22240 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
22241 instead of TARGET_INTER_UNIT_MOVES.
22242 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
22243 operand 1 from Yi to Yj and from Ym to Yn.
22245 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22247 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
22248 (float_truncate_hi_): Likewise.
22249 (float_extend_lo_): Likewise.
22250 (float_truncate_lo_): Likewise.
22251 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
22252 (aarch64_float_extend_lo_v2df): Likewise.
22253 (vec_unpacks_hi_v4sf): Likewise.
22254 (aarch64_float_truncate_lo_v2sf): Likewise.
22255 (aarch64_float_truncate_hi_v4sf): Likewise.
22256 (vec_pack_trunc_v2df): Likewise.
22257 (vec_pack_trunc_df): Likewise.
22259 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22261 * config/aarch64/aarch64-builtins.c
22262 (aarch64_fold_builtin): Fold float conversions.
22263 * config/aarch64/aarch64-simd-builtins.def
22264 (floatv2si, floatv4si, floatv2di): New.
22265 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
22266 * config/aarch64/aarch64-simd.md
22267 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
22268 * config/aarch64/iterators.md (FLOATUORS): New.
22269 (optab): Add float, floatuns.
22270 (su_optab): Likewise.
22272 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22274 * config/aarch64/aarch64-builtins.c
22275 (aarch64_builtin_vectorized_function): Use new names for
22277 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
22278 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
22279 (fcvtzu): Split as...
22280 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
22281 (fcvtas): Split as...
22282 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
22283 (fcvtau): Split as...
22284 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
22285 (fcvtps): Split as...
22286 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
22287 (fcvtpu): Split as...
22288 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
22289 (fcvtms): Split as...
22290 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
22291 (fcvtmu): Split as...
22292 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
22293 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
22294 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
22295 (lfrintnusf, lfrintnudf): Likewise.
22296 * config/aarch64/aarch64-simd.md
22297 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
22299 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22300 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22301 (fcvt_pattern): Likewise.
22303 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22305 * config/aarch64/aarch64-simd.md
22306 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22307 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22309 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22311 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22312 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22313 (vrnd<a,m,n,p>_f32): Implement using builtins.
22314 (vrnd<i,x><q>_f<32, 64>): New.
22316 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22318 * config/aarch64/aarch64-builtins.c
22319 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22320 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22321 (frintz): Rename to...
22323 (frintp): Rename to...
22325 (frintm): Rename to...
22327 (frinti): Rename to...
22328 (nearbyint): ...this.
22329 (frintx): Rename to...
22331 (frinta): Rename to...
22333 * config/aarch64/aarch64-simd.md
22334 (aarch64_frint<frint_suffix><mode>): Delete.
22335 (<frint_pattern><mode>2): Convert to insn.
22336 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22337 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22338 (frint_pattern): Likewise.
22339 (frint_suffix): Likewise.
22341 2013-04-29 Richard Biener <rguenther@suse.de>
22343 PR tree-optimization/57081
22344 * loop-init.c: Include tree-flow.h.
22345 (loop_optimizer_finalize): Free number of iteration estimates.
22346 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22348 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22350 PR tree-optimization/57083
22351 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22352 non-singleton shift count range, zero extend low_bound for uns case.
22354 * config/i386/predicates.md (general_vector_operand): New predicate.
22355 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22356 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22357 if they aren't nonimmediate operands. If their original values
22358 satisfy const_vector_equal_evenodd_p, don't shift them.
22359 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22360 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22361 if not nonimmediate_operand.
22362 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22363 instead of register_operand.
22364 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22366 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22368 * stor-layout.c (finalize_size_functions): Allocate a structure and
22369 reset cfun before dumping the functions.
22371 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22373 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22376 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22377 use xop_pmacsdqh if uns_p.
22378 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22379 the immediate rotate count.
22381 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22383 * rtl.h (struct rtx_def): Add comment for field jump.
22384 (LRA_SUBREG_P): New macro.
22385 * recog.c (register_operand): Check LRA_SUBREG_P.
22386 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22388 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22389 (lra_final_code_change): Skip subreg change for operators.
22390 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22391 if there are no operand changes.
22392 * lra-constraints.c (curr_insn_set): New.
22393 (match_reload): Set LRA_SUBREG_P.
22394 (emit_spill_move): Ditto.
22395 (check_and_process_move): Use curr_insn_set. Process only single
22396 set insns. Don't initialize sec_mem_p and change_p.
22397 (simplify_operand_subreg): Use LRA_SUBREG_P.
22398 (reg_in_class_p): New function.
22399 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22400 of #ifdef. Add code to remove cycling.
22401 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22402 non-null disp. Reload inner instead of disp when base and index
22403 are null. Try to put lo_sum into register.
22404 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22405 (check_and_process_move): Move code for move cost check to
22406 simple_move_p. Remove equiv_substitution.
22407 (simple_move_p): New function.
22408 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22409 curr_insn_set. Call check_and_process_move only for single set
22410 insns. Use the new function. Move call of check_and_process_move
22411 after operand equiv substitution and address process.
22413 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22416 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22417 with nonlocal goto receivers or returns twice calls, ignore
22418 unininitialized values from abnormal edges to nl goto receiver
22419 or returns twice call.
22421 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22423 PR tree-optimization/57051
22424 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22425 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22428 2013-04-26 Richard Biener <rguenther@suse.de>
22430 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22431 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22432 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22433 (fixup_loop_arrays_after_move): New function.
22434 (move_sese_region_to_fn): Properly outline the loop tree parts
22435 of the SESE region.
22437 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22439 * config/i386/i386.md (type, unit): Fix long lines.
22441 2013-04-26 Richard Biener <rguenther@suse.de>
22443 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22444 (lto-streamer-out.o): Likewise.
22445 * cfgloop.c (init_loops_structure): Export, add struct function
22446 argument and adjust.
22447 (flow_loops_find): Adjust.
22448 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22449 (init_loops_structure): Declare.
22450 * lto-streamer-in.c: Include cfgloop.h.
22451 (input_cfg): Input the loop tree.
22452 * lto-streamer-out.c: Include cfgloop.h.
22453 (output_cfg): Output the loop tree.
22454 (output_struct_function_base): Do not drop PROP_loops.
22456 2013-03-26 Richard Biener <rguenther@suse.de>
22458 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22459 (pass_build_cfg): Provide PROP_loops.
22460 (move_sese_region_to_fn): Remove loops that are outlined into fn
22462 * tree-inline.c: Include cfgloop.h.
22463 (initialize_cfun): Do not drop PROP_loops.
22464 (copy_loops): New function.
22465 (copy_cfg_body): Copy loop structure.
22466 (tree_function_versioning): Initialize destination loop tree.
22467 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22468 (pass_parallelize_loops): Do IL verification.
22469 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22470 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22471 the CFG make sure we fixup loops as well.
22472 * tree-ssa-tail-merge.c: Include cfgloop.h.
22473 (replace_block_by): When merging loop latches mark loops for fixup.
22474 * lto-streamer-out.c (output_struct_function_base): Drop
22475 PROP_loops for now.
22476 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22477 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22478 * ipa-split.c: Include cfgloop.h.
22479 (split_function): Add the new return block to the loop tree root.
22480 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22481 whether we have removed the forwarder block.
22482 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22483 * cfgloop.h (place_new_loop): Declare.
22484 * cfgloopmanip.c (place_new_loop): Export.
22485 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22486 (tree-switch-conversion.o): Likewise.
22487 (tree-complex.o): Likewise.
22488 (tree-inline.o): Likewise.
22489 (tree-ssa-tailmerge.o): Likewise.
22490 (ipa-split.o): Likewise.
22491 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22492 (tree-ssa-copy.o): Likewise.
22493 * tree-switch-conversion.c: Include cfgloop.h
22494 (process_switch): If we emit a bit-test cascade, schedule loops
22496 * tree-complex.c: Include cfgloop.h.
22497 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22498 * asan.c: Include cfgloop.h.
22499 (create_cond_insert_point): Properly add new basic-blocks to
22500 loops, schedule loop fixup.
22501 * cfgloop.c (verify_loop_structure): Check that looks are not
22503 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22505 (expand_omp_for_generic): Likewise.
22506 (expand_omp_sections): Likewise.
22507 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22508 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22509 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22510 is initialized, not when loops are present.
22511 * tree-parloops.c (parallelize_loops): Remove checking here.
22512 * passes.c (init_optimization_passes): Schedule a copy-propagation
22513 pass before complete unrolling of inner loops.
22515 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22517 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22518 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22519 (colorize_init): Add argument to _WIN32 version.
22520 * toplev.c: Include diagnostic-color.h.
22521 (process_options): Default to -fdiagnostics-color=auto if
22522 GCC_COLORS env var is in the environment.
22523 * common.opt (fdiagnostics-color=): Add Var and Init.
22524 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22525 env var is in the environment, the default is auto rather than never.
22527 * diagnostic.h (file_name_as_prefix): Add context argument.
22528 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22529 the string as locus.
22530 * langhooks.c (lhd_print_error_function): Adjust caller.
22532 2013-04-25 Lawrence Crowl <crowl@google.com>
22534 * var-tracking.c (shared_hash_def::htab):
22535 Change type to hash_table. Update dependent calls and types.
22537 2013-04-25 Lawrence Crowl <crowl@google.com>
22539 * Makefile.in: Update as needed below.
22541 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22542 Move declaration to after the type's method definitons.
22544 * attribs.c (htab_t scoped_attributes::attribute_hash):
22545 Change type to hash_table. Update dependent calls and types.
22547 * bitmap.c (htab_t bitmap_desc_hash):
22548 Change type to hash_table. Update dependent calls and types.
22550 * cselib.c (htab_t cselib_hash_table):
22551 Change type to hash_table. Update dependent calls and types.
22553 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22554 (hash_string_slot_node): Move implementation into lto-streamer.h
22555 struct string_slot_hasher.
22556 (eq_string_slot_node): Likewise.
22558 * data-streamer-out.c: Update output_block::string_hash_table
22559 dependent calls and types.
22561 * dwarf2cfi.c (htab_t trace_index):
22562 Change type to hash_table. Update dependent calls and types.
22564 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22565 Change type to hash_table. Update dependent calls and types.
22566 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22567 (htab_t optimize_external_refs::map): Likewise.
22568 (htab_t output_comp_unit::extern_map): Likewise.
22569 (htab_t output_comdat_type_unit::extern_map): Likewise.
22570 (htab_t output_macinfo::macinfo_htab): Likewise.
22571 (htab_t optimize_location_lists::htab): Likewise.
22572 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22574 * except.c (htab_t ehspec_hash_type):
22575 Change type to hash_table. Update dependent calls and types.
22576 (assign_filter_values::ttypes): Likewise.
22577 (assign_filter_values::ehspec): Likewise.
22578 (sjlj_assign_call_site_values::ar_hash): Likewise.
22579 (convert_to_eh_region_ranges::ar_hash): Likewise.
22581 * gcse.c (htab_t pre_ldst_table):
22582 Change type to hash_table. Update dependent calls and types.
22584 * ggc-common.c (htab_t saving_htab):
22585 Change type to hash_table. Update dependent calls and types.
22586 (htab_t loc_hash): Likewise.
22587 (htab_t ptr_hash): Likewise.
22588 (call_count): Rename ggc_call_count.
22589 (call_alloc): Rename ggc_call_alloc.
22590 (loc_descriptor): Rename make_loc_descriptor.
22591 (add_statistics): Rename ggc_add_statistics.
22593 * ggc-common.c (saving_htab):
22594 Change type to hash_table. Update dependent calls and types.
22596 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22597 (push_gimplify_context): Likewise.
22598 (pop_gimplify_context): Likewise.
22599 (struct gimple_temp_hash_elt): Added.
22600 (struct gimplify_hasher): Likewise.
22601 (struct gimplify_ctx.temp_htab):
22602 Change type to hash_table. Update dependent calls and types.
22604 * gimple-fold.c: Include gimplify-ctx.h.
22606 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22607 Change type to hash_table. Update dependent calls and types.
22608 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22609 avoid potential global name collision.
22611 * gimplify.c: Include gimplify-ctx.h.
22612 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22613 (htab_t gimplify_ctx::temp_htab):
22614 Update dependent calls and types for new type hash_table.
22615 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22616 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22618 * gimplify-ctx.h: New.
22619 (struct gimple_temp_hash_elt): Move from gimplify.c.
22620 (class gimplify_hasher): New.
22621 (struct gimplify_ctx): Move from gimple.h.
22622 (htab_t gimplify_ctx::temp_htab):
22623 Change type to hash_table. Update dependent calls and types.
22625 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22626 (htab_t ivs_params::newivs_index):
22627 Change type to hash_table. Update dependent calls and types.
22628 (htab_t ivs_params::params_index): Likewise.
22629 (htab_t print_generated_program::params_index): Likewise.
22630 (htab_t gloog::newivs_index): Likewise.
22631 (htab_t gloog::params_index): Likewise.
22633 * graphite.c: Include graphite-htab.h.
22634 4htab_t graphite_transform_loops::bb_pbb_mapping):
22635 Change type to hash_table. Update dependent calls and types.
22637 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22638 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22639 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22641 * graphite-dependences.c: Include graphite-htab.h.
22642 (loop_is_parallel_p): Change hash table type of parameter.
22644 * graphite-htab.h: New.
22645 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22646 (extern find_pbb_via_hash): Move from graphite-poly.h.
22647 (extern loop_is_parallel_p): Move from graphite-poly.h.
22648 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22650 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22651 (extern loop_is_parallel_p): Move to graphite-htab.h.
22652 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22654 * haifa-sched.c (htab_t delay_htab):
22655 Change type to hash_table. Update dependent calls and types.
22656 (htab_t delay_htab_i2): Likewise.
22658 * ira-color.c (htab_t allocno_hard_regs_htab):
22659 Change type to hash_table. Update dependent calls and types.
22661 * ira-costs.c (htab_t cost_classes_htab):
22662 Change type to hash_table. Update dependent calls and types.
22664 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22665 Change type to hash_table. Update dependent calls and types.
22667 * loop-iv.c (htab_t bivs):
22668 Change type to hash_table. Update dependent calls and types.
22670 * loop-unroll.c (htab_t opt_info::insns_to_split):
22671 Change type to hash_table. Update dependent calls and types.
22672 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22674 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22675 (struct string_slot_hasher): New.
22676 (htab_t output_block::string_hash_table):
22677 Change type to hash_table. Update dependent calls and types.
22679 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22680 (htab_t file_name_hash_table):
22681 Change type to hash_table. Update dependent calls and types.
22683 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22686 * lto-streamer.c (htab_t tree_htab):
22687 Change type to hash_table. Update dependent calls and types.
22689 * omp-low.c: Include gimplify-ctx.h.
22691 * passes.c (htab_t name_to_pass_map):
22692 Change type to hash_table. Update dependent calls and types.
22693 (pass_traverse): Rename to passes_pass_traverse.
22695 * plugin.c (htab_t event_tab):
22696 Change type to hash_table. Update dependent calls and types.
22698 * postreload-gcse.c (htab_t expr_table):
22699 Change type to hash_table. Update dependent calls and types.
22700 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22702 * sese.c (debug_rename_map_1): Make extern.
22703 (htab_t copy_bb_and_scalar_dependences::rename_map):
22704 Change type to hash_table. Update dependent calls and types.
22706 * sese.h (extern debug_rename_map): Move to .c file.
22708 * store-motion.c (htab_t store_motion_mems_table):
22709 Change type to hash_table. Update dependent calls and types.
22711 * trans-mem.c (htab_t tm_new_mem_hash):
22712 Change type to hash_table. Update dependent calls and types.
22714 * tree-browser.c (htab_t TB_up_ht):
22715 Change type to hash_table. Update dependent calls and types.
22717 * tree-cfg.c (htab_t discriminator_per_locus):
22718 Change type to hash_table. Update dependent calls and types.
22720 * tree-complex.c: Include tree-hasher.h
22721 (htab_t complex_variable_components):
22722 Change type to hash_table. Update dependent calls and types.
22724 * tree-eh.c (htab_t finally_tree):
22725 Change type to hash_table. Update dependent calls and types.
22727 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22728 struct int_tree_hasher.
22729 (extern int_tree_map_eq): Likewise.
22730 (uid_decl_map_hash): Removed.
22731 (extern decl_tree_map_eq): Likewise.
22733 * tree-hasher.h: New.
22734 (struct int_tree_hasher): New.
22735 (typedef int_tree_htab_type): New.
22737 * tree-inline.c: Include gimplify-ctx.h.
22739 * tree-mudflap.c: Include gimplify-ctx.h.
22741 * tree-parloops.c: Include tree-hasher.h.
22742 (htab_t eliminate_local_variables_stmt::decl_address):
22743 Change type to hash_table. Update dependent calls and types.
22744 (htab_t separate_decls_in_region::decl_copies): Likewise.
22746 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22747 Change type to hash_table. Update dependent calls and types.
22749 * tree-sra.c (candidates):
22750 Change type to hash_table. Update dependent calls and types.
22752 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22754 (int_tree_map_hash): Likewise.
22756 * tree-ssa-dom.c (htab_t avail_exprs):
22757 Change type to hash_table. Update dependent calls and types.
22759 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22760 Change type to hash_table. Update dependent calls and types.
22762 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22763 Change type to hash_table. Update dependent calls and types.
22765 * tree-ssa-phiopt.c (seen_ssa_names):
22766 Change type to hash_table. Update dependent calls and types.
22768 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22769 Change type to hash_table. Update dependent calls and types.
22771 * tree-ssa-uncprop.c (equiv):
22772 Change type to hash_table. Update dependent calls and types.
22774 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22776 PR rtl-optimization/57003
22777 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22778 call note_stores with kill_clobbered_value callback again after
22779 killing regs_invalidated_by_call.
22781 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22783 * config/aarch64/aarch64-simd.md
22784 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22785 (aarch64_simd_bsl<mode>): Likewise.
22786 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22788 2013-04-25 Marek Polacek <polacek@redhat.com>
22790 PR tree-optimization/57066
22791 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22793 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22795 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22797 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22799 * config/aarch64/aarch64-builtins.c
22800 (aarch64_fold_builtin): New.
22801 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22802 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22803 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22804 * config/aarch64/arm_neon.h
22805 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22807 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22808 Tejas Belagod <tejas.belagod@arm.com>
22810 * config/aarch64/aarch64-builtins.c
22811 (aarch64_gimple_fold_builtin): New.
22812 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22813 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22814 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22815 (addvv4sf): Update.
22816 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22818 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22820 * config/aarch64/aarch64.md
22821 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22823 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22825 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22826 (*ngcsi_uxtw): New pattern.
22828 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22829 Julian Brown <julian@codesourcery.com>
22831 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22832 (TB_DREG): Add T_V4HF.
22833 (v4hf_UP): New macro.
22834 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22835 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22836 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22838 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22839 (arm_vector_mode_supported_p): Handle V4HF.
22840 (arm_mangle_map): Handle V4HFmode.
22841 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22842 * config/arm/arm_neon_builtins.def: Add entries for
22843 vcvtv4hfv4sf, vcvtv4sfv4hf.
22844 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22845 (neon_vcvtv4hfv4sf): Likewise.
22846 * config/arm/neon-gen.ml: Handle half-precision floating point
22848 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22849 * config/arm/arm_neon.h: Regenerate.
22850 * config/arm/neon.ml (type elts): Add F16.
22851 (type vectype): Add T_float16x4, T_floatHF.
22852 (type vecmode): Add V4HF.
22853 (type features): Add Requires_FP_bit feature.
22854 (elt_width): Handle F16.
22855 (elt_class): Likewise.
22856 (elt_of_class_width): Likewise.
22857 (mode_of_elt): Refactor.
22858 (type_for_elt): Handle F16, fix error messages.
22859 (vectype_size): Handle T_float16x4.
22860 (vcvt_sh): New function.
22861 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22862 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22863 (string_of_mode): Handle V4HF.
22864 * doc/arm-neon-intrinsics.texi: Regenerate.
22866 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22868 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22869 format specifier in 'X' case.
22871 2013-04-25 Alan Modra <amodra@gmail.com>
22874 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22875 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22876 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22877 Repeat for many other rotate/shift and mask patterns using subregs.
22878 Name lshiftrt insns.
22879 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22880 on WORDS_BIG_ENDIAN.
22882 2013-04-25 Alan Modra <amodra@gmail.com>
22884 * config.gcc: Support little-endian powerpc-linux targets.
22885 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22886 (LINK_OS_LINUX_SPEC): Define.
22887 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22888 Preserve MASK_LITTLE_ENDIAN.
22889 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22890 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22891 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22892 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22893 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22894 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22895 Correct fp word order for little-endian. Don't shift toc entries
22896 smaller than a word for little-endian.
22897 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22898 (bswapdi2 splits): Correct low-part subreg for little-endian.
22899 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22900 low/high where such is correct only for be.
22901 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22902 little-endian for -mcall-aixdesc.
22904 2013-04-25 Alan Modra <amodra@gmail.com>
22906 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22907 replace_equiv_address_nv.
22909 2013-04-25 Alan Modra <amodra@gmail.com>
22911 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22913 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22916 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22917 * rtl.h (struct rtx_def): ...
22919 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22921 PR rtl-optimizations/57046
22922 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22923 for multi-reg splits.
22925 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22927 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22929 2013-04-24 Sterling Augustine <saugustine@google.com>
22931 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22932 (comp_dir_string, debug_str_dwo_section): New.
22933 (DEBUG_STR_DWO_SECTION): Rename to ...
22934 (DEBUG_DWO_STR_SECTION): ... this.
22935 (DEBUG_NORM_STR_SECTION): Delete.
22936 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22937 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22938 (find_AT_string): Move most logic to ...
22939 (find_AT_string_in_table): ... here. New.
22940 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22941 add_skeleton_AT_string. Delete logic.
22942 (output_skeleton_debug_sections): Remove call to
22943 add_top_level_skeleton_die_attrs.
22944 (add_comp_dir_attribute): Move logic to comp_dir_string.
22945 (dwarf2out_init): Initialize debug_str_dwo_section.
22946 (output_indirect_string): Call find_string_form.
22947 (output_indirect_strings): Rewrite.
22948 (prune_unused_types): Empty skeleton_debug_str_hash.
22949 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22950 (dwarf2out_finish): Call output_indirect_strings.
22952 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22954 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22956 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22958 * rtl.h (struct rtx_def): Add comment for field jump.
22959 (LRA_SUBREG_P): New macro.
22960 * recog.c (register_operand): Check LRA_SUBREG_P.
22961 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22963 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22964 (lra_final_code_change): Skip subreg change for operators.
22965 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22966 if there are no operand changes.
22967 * lra-constraints.c (curr_insn_set): New.
22968 (match_reload): Set LRA_SUBREG_P.
22969 (emit_spill_move): Ditto.
22970 (check_and_process_move): Use curr_insn_set. Process only single
22971 set insns. Don't initialize sec_mem_p and change_p.
22972 (simplify_operand_subreg): Use LRA_SUBREG_P.
22973 (reg_in_class_p): New function.
22974 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22975 of #ifdef. Add code to remove cycling.
22976 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22977 non-null disp. Reload inner instead of disp when base and index
22978 are null. Try to put lo_sum into register.
22979 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22980 (check_and_process_move): Move code for move cost check to
22981 simple_move_p. Remove equiv_substitution.
22982 (simple_move_p): New function.
22983 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22984 curr_insn_set. Call check_and_process_move only for single set
22985 insns. Use the new function. Move call of check_and_process_move
22986 after operand equiv substitution and address process.
22988 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22990 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22991 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22992 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22994 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22996 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22998 2013-04-24 Marek Polacek <polacek@redhat.com>
23000 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
23001 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
23002 (select_loops_exit_conditions): Likewise.
23003 (number_of_iterations_for_all_loops): Likewise.
23004 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
23005 (scev_analysis): Likewise.
23007 2013-04-02 Catherine Moore <clm@codesourcery.com>
23008 Chao-ying Fu <fu@mips.com>
23010 * config/mips/micromips.md (jraddiusp): New pattern.
23011 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
23012 instruction if possible.
23014 2013-04-24 Alan Modra <amodra@gmail.com>
23016 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
23018 2013-04-24 Julian Brown <julian@codesourcery.com>
23019 Chung-Lin Tang <cltang@codesourcery.com>
23021 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
23022 dependency behavior in enumeration type DIE generation. Add TODO note
23023 to comments about future DW_FORM_sdata/udata re-work of related code.
23025 2013-04-23 Lawrence Crowl <crowl@google.com>
23027 * Makefile.in: Update as needed below.
23029 * hash-table.h (class hash_table):
23030 Correct many methods with parameter types compare_type to the correct
23031 value_type. (Correct code was unlikely to notice the change.)
23032 (hash_table::elements_with_deleted) New.
23033 (class hashtable::iterator): New.
23034 (hashtable::begin()): New.
23035 (hashtable::end()): New.
23036 (FOR_EACH_HASH_TABLE_ELEMENT): New.
23038 * statistics.c (statistics_hashes):
23039 Change type to hash_table. Update dependent calls and types.
23041 * tree-into-ssa.c (var_infos):
23042 Change type to hash_table. Update dependent calls and types.
23044 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
23045 Change type to hash_table. Update dependent calls and types.
23047 * tree-ssa-loop-im.c (struct mem_ref.refs):
23048 Change type to hash_table. Update dependent calls and types.
23050 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
23051 Change type to hash_table. Update dependent calls and types.
23053 * tree-ssa-sccvn.c (vn_tables_s::nary):
23054 Change type to hash_table. Update dependent calls and types.
23055 (vn_tables_s::phis): Likewise.
23056 (vn_tables_s::references): Likewise.
23058 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
23059 (vn_reference_eq): Update parameter and return types.
23061 * tree-ssa-structalias.c (pointer_equiv_class_table):
23062 Change type to hash_table. Update dependent calls and types.
23063 (location_equiv_class_table): Likewise.
23065 * tree-vect-data-refs.c: Consequential changes for making
23066 peeling a hash_table.
23068 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
23069 (destroy_loop_vec_info): Dependent hash_table update.
23071 * tree-vectorizer.h (peeling_htab):
23072 Change type to hash_table. Update dependent calls and types.
23074 2013-04-23 Shiva Chen <shiva0217@gmail.com>
23076 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
23077 to check the register content is equal or not.
23078 * lra-constraints.c (match_reload): Use lra_assign_reg_val
23079 to assign register content record.
23080 * lra-eliminations.c (update_reg_eliminate): Use
23081 lra_update_reg_val_offset to update register content offset.
23082 * lra-int.h (struct lra_reg): Add offset member.
23083 (lra_reg_val_equal_p): New static inline function.
23084 (lra_update_reg_val_offset): New static inline function.
23085 (lra_assign_reg_val): New static inline function.
23086 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
23087 to assign register content record.
23088 (initialize_lra_reg_info_element): Initial offset to zero.
23090 2013-04-23 Catherine Moore <clm@codesourcery.com>
23092 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
23093 operands. Record compression.
23095 2013-04-23 Xinliang David Li <davidxl@google.com>
23097 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
23099 2013-04-23 Richard Biener <rguenther@suse.de>
23101 PR middle-end/57036
23102 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
23103 parameter, only add abnormal goto edges from the copied body
23104 if the call could perform abnormal gotos.
23105 (copy_cfg_body): Adjust.
23107 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
23109 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
23111 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
23113 * coretypes.h (gimple_stmt_iterator): Add struct to make
23116 2013-04-23 Richard Biener <rguenther@suse.de>
23118 PR tree-optimization/57026
23119 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
23120 from SSA names occuring in abnormal PHI nodes.
23122 2013-04-22 Andi Kleen <ak@linux.intel.com>
23124 * lto/lto.c (print_lto_report_1): Fix LTO report names.
23126 2013-04-22 Andi Kleen <ak@linux.intel.com>
23128 * lto/lto.c (print_lto_report_1): Declare early.
23129 (read_cgraph_and_symbols): Call print_lto_report_1 early.
23131 2013-04-22 Andi Kleen <ak@linux.intel.com>
23133 * common.opt (-flto-report-wpa): Add.
23134 * doc/invoke.texi (-flto-report-wpa): Add.
23135 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
23138 2013-04-22 Xinliang David Li <davidxl@google.com>
23140 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
23141 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
23142 * Makefile.in: New dependency
23144 David Daney <ddaney.cavm@gmail.com>
23146 * configure.ac (gcc_cv_as_micromips_support): Use the
23147 --fatal-warnings option.
23148 * configure: Regenerate.
23150 2013-04-22 Marek Polacek <polacek@redhat.com>
23153 * tsan.c (instrument_expr): Don't instrument expression
23154 in case its size is zero.
23156 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
23160 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23162 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23164 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23166 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
23167 (gimple_stmt_iterator): New typedef.
23168 * gimple.h (gimple_stmt_iterator): Rename to...
23169 (gimple_stmt_iterator_d): ... This.
23170 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
23171 trees be valid for GIMPLE and GENERIC.
23172 (TARGET_GIMPLE_FOLD_BUILTIN): New.
23173 * gimple-fold.c (gimple_fold_call): Call target hook
23174 gimple_fold_builtin.
23175 * hooks.c (hook_bool_gsiptr_false): New.
23176 * hooks.h (hook_bool_gsiptr_false): New.
23177 * target.def (fold_stmt): New.
23178 * doc/tm.texi: Regenerate.
23180 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
23183 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
23184 a set sp if no stack realignment.
23186 2013-04-22 Nick Clifton <nickc@redhat.com>
23188 * config.gcc (tilegx-linux): Extend extra_objs rather than
23190 (tilepro-linux): Likewise.
23192 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23194 * config/aarch64/aarch64-builtins.c
23196 (CF0, CF1, CF2, CF3, CF4, CF10): New.
23197 (VAR<1-12>): Add MAP parameter.
23198 (BUILTIN_*): Likewise.
23199 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
23200 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
23201 (aarch64_ushl_n<mode>): Likewise.
23202 (aarch64_sshr_n<mode>): Likewise.
23203 (aarch64_ushr_n<mode>): Likewise.
23204 (aarch64_<maxmin><mode>): Likewise.
23205 (aarch64_sqrt<mode>): Likewise.
23206 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
23207 (vshr<q>_n_*): Likewise.
23209 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23211 * config/aarch64/aarch64-builtins.c
23212 (aarch64_simd_builtin_type_mode): Handle SF types.
23214 (BUILTIN_GPF): Define.
23215 (aarch64_init_simd_builtins): Handle SF types.
23216 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
23217 (frecps): Likewise.
23218 (frecpx): Likewise.
23219 * config/aarch64/aarch64-simd.md
23220 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
23221 (aarch64_frecpe<mode>): New.
23222 (aarch64_frecps<mode>): Likewise.
23223 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
23224 (v8type): Add frecp<esx>.
23225 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
23226 (aarch64_frecps<mode>): Likewise.
23227 * config/aarch64/iterators.md (FRECP): New.
23228 (frecp_suffix): Likewise.
23229 * config/aarch64/arm_neon.h
23230 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
23232 2013-04-22 Christian Bruel <christian.bruel@st.com>
23235 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
23236 (REG_CLASS_NAMES): Idem.
23237 (REG_CLASS_CONTENTS): Idem.
23238 (REGCLASS_HAS_FP_REG): Idem.
23239 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
23240 (sh_conditional_register_usage): Idem.
23242 2013-04-21 Jeff Law <law@redhat.com>
23244 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
23245 (ssa_forward_propagate_and_combine): Use it.
23247 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23249 * lra.c: Update the flow chart diagram.
23251 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23253 PR rtl-optimization/56847
23254 * lra-constraints.c (process_alt_operands): Discourage alternative
23255 with non-matche doffsettable memory constraint fro memory with
23258 2013-04-19 Richard Biener <rguenther@suse.de>
23260 PR tree-optimization/56982
23261 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
23263 * gimplify.c (gimplify_call_expr): Notice special calls.
23264 (gimplify_modify_expr): Likewise.
23265 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
23266 abnormal control flow receivers.
23267 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
23268 in the same way as cfun->has_nonlocal_labels.
23269 (gimple_purge_dead_abnormal_call_edges): Likewise.
23270 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
23271 receivers start a basic-block.
23273 2013-04-19 Richard Biener <rguenther@suse.de>
23275 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
23277 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
23278 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
23279 (SLP_TREE_LOAD_PERMUTATION): Add.
23280 (vect_transform_slp_perm_load): Adjust prototype.
23281 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
23282 (vect_free_slp_instance): Likewise.
23283 (vect_create_new_slp_node): Likewise.
23284 (vect_supported_slp_permutation_p): Remove.
23285 (vect_slp_rearrange_stmts): Adjust.
23286 (vect_supported_load_permutation_p): Likewise. Inline
23287 vect_supported_slp_permutation_p here.
23288 (vect_analyze_slp_instance): Compute load permutations per
23289 slp node instead of per instance.
23290 (vect_get_slp_defs): Adjust.
23291 (vect_transform_slp_perm_load): Likewise.
23292 (vect_schedule_slp_instance): Remove redundant code.
23293 (vect_schedule_slp): Remove hack for PR56270, add it ...
23294 * tree-vect-stmts.c (vectorizable_load): ... here, do not
23295 CSE loads for SLP. Adjust.
23297 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23299 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23300 spelling in two comments.
23302 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23305 * config/arm/arm.c (load_multiple_sequence): Require SP
23306 as base register for loads if SP is in the register list.
23308 2013-04-19 Martin Jambor <mjambor@suse.cz>
23310 PR tree-optimization/56718
23311 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23312 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23313 and made public. Adjusted all callers.
23314 (ipa_intraprocedural_devirtualization): New function.
23315 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23316 (ipa_intraprocedural_devirtualization): Likewise.
23317 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23319 2013-04-19 Richard Biener <rguenther@suse.de>
23321 PR tree-optimization/57000
23322 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23324 2013-04-19 Terry Guo <terry.guo@arm.com>
23326 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23328 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23329 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23330 (cortex_m4_fmacs): Use new reservations.
23331 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23333 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23335 PR rtl-optimization/56999
23336 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23338 (lra_coalesce): Remove split_origin_bitmap and related code.
23339 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23340 ranges if necessary.
23342 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23344 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23346 (ix86_expand_call): Remove clobbered_registers array and use
23347 x86_64_ms_sysv_extra_clobbered_registers instead.
23348 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23350 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23352 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23353 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23354 (*call_value_rex64_ms_sysv): Ditto.
23356 2013-04-18 Cary Coutant <ccoutant@google.com>
23358 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23359 parent instead of die_mark.
23361 2013-04-18 Diego Novillo <dnovillo@google.com>
23363 * gimple.c (create_gimple_tmp): New.
23364 (get_expr_type): New.
23365 (build_assign): New.
23366 (build_type_cast): New.
23367 * gimple.h (enum ssa_mode): Define.
23368 (gimple_seq_set_location): New.
23369 * asan.c (build_check_stmt): Change some gimple_build_* calls
23370 to use build_assign and build_type_cast.
23372 2013-04-18 Richard Biener <rguenther@suse.de>
23374 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23375 handle negative step. Remove redundant checks.
23376 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23377 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23378 for negative step and grouped loads fail to vectorize.
23380 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23382 * emit-rtl.c (reset_insn_used_flags): New function.
23383 (reset_all_used_flags): Use it.
23384 (verify_insn_sharing): New function.
23385 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23387 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23389 PR tree-optimization/56984
23390 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23391 and (x >> M) >= N don't register any assertion if N << M is the
23394 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23396 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23397 do not delete_insn insns that have not yet been emitted, only
23398 unlink them with remove_insn.
23399 * df-scan.c (df_insn_delete): Revert r197492.
23401 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23403 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23404 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23406 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23408 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23409 define_insn_and_split.
23410 (and_scc,ior_scc,negscc): Likewise.
23411 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23413 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23415 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23416 can benefit from using a sequence of LDRD instructions in epilogue
23417 instead of a single LDM instruction.
23419 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23422 * doc/extend.texi: Fix typo.
23424 2013-04-17 Richard Biener <rguenther@suse.de>
23426 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23427 (vect_build_slp_tree): ... here.
23428 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23429 match. Remove special-casing of mismatched complex loads.
23430 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23431 re-try the match with swapped commutative operands.
23432 (vect_supported_load_permutation_p): Remove special-casing of
23433 mismatched complex loads.
23434 (vect_analyze_slp_instance): Adjust.
23436 2013-04-17 Richard Biener <rguenther@suse.de>
23438 PR rtl-optimization/56921
23439 * cfgloop.h (struct loop): Add simple_loop_desc member.
23440 (struct niter_desc): Mark with GTY(()).
23441 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23442 * loop-iv.c (get_simple_loop_desc): Likewise.
23443 (free_simple_loop_desc): Likewise.
23446 2013-04-16 Richard Biener <rguenther@suse.de>
23448 PR rtl-optimization/56921
23449 * loop-init.c (pass_rtl_move_loop_invariants): Add
23450 TODO_do_not_ggc_collect to todo_flags_finish.
23451 (pass_rtl_unswitch): Same.
23452 (pass_rtl_unroll_and_peel_loops): Same.
23453 (pass_rtl_doloop): Same.
23455 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23457 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23458 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23459 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23461 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23462 * tree-streamer.c (record_common_node): Adjust reference in comment.
23464 2013-04-17 Terry Guo <terry.guo@arm.com>
23466 * config/arm/cortex-m4.md: Add a new bypass.
23468 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23470 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23472 (*subs_<optab><mode>_multp2): New pattern.
23473 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23474 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23476 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23478 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23479 (*subs_mul_imm_<mode>): New pattern.
23481 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23484 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23485 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23486 (vsx_movti_32bit): Change j->wa to O->wa.
23488 2013-04-16 Richard Biener <rguenther@suse.de>
23490 PR rtl-optimization/56921
23491 * loop-init.c (pass_rtl_move_loop_invariants): Add
23492 TODO_do_not_ggc_collect to todo_flags_finish.
23493 (pass_rtl_unswitch): Same.
23494 (pass_rtl_unroll_and_peel_loops): Same.
23495 (pass_rtl_doloop): Same.
23497 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23499 * config/arm/arm.c (emit_multi_reg_push): New declaration
23500 for an existing function.
23501 (arm_emit_strd_push): New function.
23502 (arm_expand_prologue): Used here.
23503 (arm_emit_ldrd_pop): New function.
23504 (arm_expand_epilogue): Used here.
23505 (arm_get_frame_offsets): Update condition.
23506 (arm_emit_multi_reg_pop): Add a special case for load of a single
23507 register with writeback.
23509 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23511 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23514 2013-04-16 Richard Biener <rguenther@suse.de>
23516 PR tree-optimization/56756
23517 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23518 (first_mem_ref_loc): New.
23519 (execute_sm): Place the load temporarily before a previous
23520 access instead of in the latch edge to ensure its SSA dependencies
23521 are defined at points dominating the load.
23523 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23525 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23526 correct fix by moving header and footer insn to the footer of
23527 the merged basic block. Clear BB_END of the merged-away block.
23529 PR middle-end/43631
23530 * emit-rtl.c (make_note_raw): New function.
23531 (link_insn_into_chain): New static inline function.
23532 (add_insn): Use it.
23533 (add_insn_before, add_insn_after): Factor insn chain linking code...
23534 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23535 using link_insn_into_chain.
23536 (note_outside_basic_block_p): New helper function for emit_note_after
23537 and emit_note_before.
23538 (emit_note_after): Use nobb variant of add_insn_after if the note
23539 should not be contained in a basic block.
23540 (emit_note_before): Use nobb variant of add_insn_before if the note
23541 should not be contained in a basic block.
23542 (emit_note_copy): Use make_note_raw.
23543 (emit_note): Likewise.
23544 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23545 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23546 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23547 the moved barrier the tail of the basic block it follows.
23548 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23550 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23552 PR tree-optimization/56962
23553 * gimple-ssa-strength-reduction.c (record_increment): Only set
23554 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23555 either rhs1 or rhs2 is equal to c->base_expr.
23557 2013-04-15 Richard Biener <rguenther@suse.de>
23559 PR tree-optimization/56933
23560 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23562 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23563 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23564 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23565 dependence check ...
23566 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23568 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23569 GROUP_READ_WRITE_DEPENDENCE.
23571 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23573 * emit-rtl.c (reset_all_used_flags): New function.
23574 (verify_rtl_sharing): Call reset_all_used_flags before and after
23575 performing the checks.
23577 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23579 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23580 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23581 * config/arm/constraints.md (De): New constraint.
23582 * config/arm/neon.md (anddi3_neon): Delete.
23583 (neon_vand<mode>): Expand to standard anddi3 pattern.
23584 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23585 Move earlier in the file.
23586 (neon_inv_logic_op2): Likewise.
23587 (arm_anddi_operand_neon): New predicate.
23589 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23591 * configure.ac (gcc_cv_ld_as_needed): Set
23592 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23593 Use -z ignore, -z record on *-*-solaris2*.
23594 (HAVE_LD_AS_NEEDED): Update comment.
23595 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23596 * configure: Regenerate.
23597 * config.in: Regenerate.
23598 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23599 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23600 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23601 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23602 equivalents. Fix markup.
23603 * doc/tm.texi: Regenerate.
23605 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23607 * config/i386/i386.opt: New option mstack-protector-guard=.
23608 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23609 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23610 TARGET_SSP_TLS_GUARD.
23611 * config/i386/i386.c (ix86_option_override_internal): Set
23612 ix86_stack_protector_guard.
23613 * config/i386/i386.md (stack_protect_set): Enable for
23614 TARGET_SSP_TLS_GUARD only.
23615 (stack_protect_set_<mode>): Ditto.
23616 (stack_protect_test): Ditto.
23617 (stack_protect_test_<mode>): Ditto.
23618 * doc/invoke.texi (i386 Option): Document.
23620 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23623 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23624 (S_MODES): Set H_MODE bit.
23625 (SF_MODES): Set only S_MODE and SF_MODE bits.
23626 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23627 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23628 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23629 <MODE_FLOAT>: Likewise.
23631 2013-04-15 Joey Ye <joey.ye@arm.com>
23633 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23635 2013-04-15 Joey Ye <joey.ye@arm.com>
23637 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23639 (thumb_far_jump_used_p): Count instruction size and set
23642 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23644 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23645 * resource.c (mark_target_live_regs): Reformat conditional block.
23647 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23649 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23650 notes, they are emitted only just before final.
23651 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23653 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23655 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23656 * cfgrtl.c (delete_insn): Call it here instead.
23657 * lra-spills.c (lra_final_code_change): Use delete_insn.
23658 * haifa-sched.c (sched_remove_insn): Likewise.
23659 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23660 returning to the nop pool.
23661 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23662 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23664 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23666 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23667 * doc/tm.texi: Regenerated.
23669 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23671 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23674 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23676 * df-core.c (df_find_def): Compare register numbers.
23677 (df_find_use): Likewise.
23679 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23682 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23683 lra_in_progress for return.
23685 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23687 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23688 define_insn into define_insn_and_split and emit movsicc patterns.
23690 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23692 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23694 2013-04-12 Richard Biener <rguenther@suse.de>
23696 * tree-pass.h (TODO_do_not_ggc_collect): New.
23697 * passes.c (execute_one_ipa_transform_pass): Honor
23698 TODO_do_not_ggc_collect.
23699 (execute_one_pass): Likewise.
23702 2013-04-10 Richard Biener <rguenther@suse.de>
23704 * passes.c (init_optimization_passes): Remove reload pass.
23705 * ira.c (do_reload): Merge into ...
23707 (rest_of_handle_reload): Remove.
23708 (pass_reload): Likewise.
23709 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23710 of reload for vzeroupper pass placement.
23712 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23714 PR tree-optimization/56918
23715 PR tree-optimization/56920
23716 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23717 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23718 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23719 use wide_mul_with_sign method.
23721 2013-04-12 Richard Biener <rguenther@suse.de>
23723 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23724 not be considered a gimple constant.
23726 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23728 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23729 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23732 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23733 Jakub Jelinek <jakub@redhat.com>
23735 * opts.c: Include diagnostic-color.h.
23736 (common_handle_option): Handle OPT_fdiagnostics_color_.
23737 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23738 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23739 (diagnostic-color.o): New.
23740 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23741 (diagnostic_color_rule): New enum.
23742 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23743 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23744 the location string.
23745 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23746 either NULL, or color kind.
23747 * diagnostic-color.c: New file.
23748 * diagnostic-color.h: New file.
23749 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23751 * doc/invoke.texi (-fdiagnostics-color): Document.
23752 * pretty-print.h (pp_show_color): Define.
23753 (struct pretty_print_info): Add show_color field.
23754 * diagnostic.c: Include diagnostic-color.h.
23755 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23756 macros. Colorize error:, warning: etc. strings and also the location
23758 (diagnostic_show_locus): Colorize the caret line.
23759 * pretty-print.c: Include diagnostic-color.h.
23760 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23761 inside of %< %> quotes or quoted through q format modifier.
23763 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23765 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23767 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23769 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23770 code in CC_NZ mode.
23771 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23774 2013-04-11 Marek Polacek <polacek@redhat.com>
23776 PR tree-optimization/48184
23777 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23779 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23781 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23782 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23783 (skip_simple_arithmetic): Tidy up.
23784 * tree.h (skip_simple_constant_arithmetic): Declare.
23786 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23788 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23790 2013-04-11 Richard Biener <rguenther@suse.de>
23792 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23793 generate vector constants.
23795 2013-04-11 Richard Biener <rguenther@suse.de>
23797 PR tree-optimization/56878
23798 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23799 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23801 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23802 Prefer to align the DR with the most invariant base address.
23804 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23806 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23809 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23811 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23812 floating-point vector comparisons against 0.
23814 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23816 PR tree-optimization/56899
23817 * fold-const.c (extract_muldiv_1): Apply distributive law
23818 only if TYPE_OVERFLOW_WRAPS (ctype).
23820 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23823 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23824 loading has side effect.
23826 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23828 * config/sparc/sparc.c: Include tree-pass.h.
23829 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23830 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23831 head of file. Change return type. Split off gate function.
23832 (sparc_gate_work_around_errata): New function.
23833 (pass_work_around_errata): New pass definition.
23834 (insert_pass_work_around_errata) New pass insert definition to
23835 insert pass_work_around_errata just after delayed-branch scheduling.
23836 (sparc_option_override): Insert the pass.
23837 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23839 2013-04-10 David S. Miller <davem@davemloft.net>
23841 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23842 or -mcpu=hypersparc.
23844 * target.def (cstore_mode): New hook.
23845 * target.h: Include insn-codes.h
23846 * targhooks.c: Likewise.
23847 (default_cstore_mode): New function.
23848 * targhooks.h: Declare it.
23849 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23850 * doc/tm.texi: Rebuild.
23851 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23852 target hook, rather than inspecting the insn_data.
23853 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23854 (TARGET_CSTORE_MODE): Redefine.
23855 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23857 * config/sparc/predicates.md (cstore_result_operand): New special
23859 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23860 Use it for operand 0.
23861 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23862 (*snesi_special): Likewise.
23863 (*snesi_zero): Likewise.
23864 (*seqsi_zero): Likewise.
23865 (*sltu_insn): Likewise.
23866 (*sgeu_insn): Likewise.
23867 (*seqdi_special): Make operand 0 and comparison operation be of
23869 (*snedi_special): Likewise.
23870 (*snedi_special_vis3): Likewise.
23871 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23872 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23873 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23874 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23875 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23876 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23877 (*sltu_extend_sp64): Likewise.
23878 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23879 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23880 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23881 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23882 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23884 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23886 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23887 (aarch64_start_file): Use the new function.
23889 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23890 Jason Merrill <jason@redhat.com>
23892 * common.opt: Add -gdwarf.
23893 * opts.c (common_handle_option): Handle it.
23894 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23896 2013-04-10 Richard Biener <rguenther@suse.de>
23898 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23899 (execute_one_ipa_transform_pass): But unconditionally here.
23900 (execute_one_pass): And here.
23901 (init_optimization_passes): Remove reload pass.
23902 * tree-pass.h (TODO_ggc_collect): Remove.
23903 (pass_reload): Likewise.
23904 * ira.c (do_reload): Merge into ...
23906 (rest_of_handle_reload): Remove.
23907 (pass_reload): Likewise.
23908 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23909 of reload for vzeroupper pass placement.
23910 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23911 and todo_flags_finish of all passes.
23913 2013-04-10 Richard Biener <rguenther@suse.de>
23915 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23916 first_const_oprnd field, rename first_def_type to first_op_type.
23917 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23918 (vect_get_and_check_slp_defs): Always use the type of the
23919 operand. Allow mixed vect_external_def, vect_constant_def types.
23920 (vect_get_constant_vectors): Handle mixed vect_external_def,
23921 vect_constant_def types.
23923 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23925 PR tree-optimization/55524
23926 * tree-ssa-math-opts.c
23927 (convert_mult_to_fma): Don't use an fms construct
23928 when we don't have an fms operation, but fnma, and it looks
23929 likely that we'll be able to use the latter.
23931 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23933 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23935 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23936 inline fail caused by overwritable functions.
23938 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23940 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23941 unnecessary bits in the constant power of two case.
23943 2013-04-10 Richard Biener <rguenther@suse.de>
23945 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23946 broken code swapping operands.
23947 (vect_build_slp_tree): Do not compute load permutations here.
23948 (vect_analyze_slp_instance): Compute load permutations here,
23949 after building the SLP tree.
23951 2013-04-09 Christian Bruel <christian.bruel@st.com>
23953 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23954 of next/prev_real_insn.
23956 2013-04-09 Jan Hubicka <jh@suse.cz>
23958 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23959 Drop aliased parameter.
23960 (function_and_variable_visibility): Do not handle alias pairs.
23961 * cgraph.c (varpool_externally_visible_p): Update prototype.
23962 * varpool.c (varpool_add_new_variable): Update.
23964 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23966 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23968 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23970 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23972 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23974 2013-04-09 Marek Polacek <polacek@redhat.com>
23976 PR tree-optimization/48762
23977 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23979 2013-04-09 Richard Biener <rguenther@suse.de>
23981 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23983 (vect_build_slp_tree): Likewise.
23984 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23985 calculating the cost of a SLP instance.
23986 (vect_analyze_slp_instance): Use it from here, after building
23989 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23991 PR middle-end/56883
23992 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23993 expand_omp_for_static_chunk): Use simple_p = true in
23994 force_gimple_operand_gsi calls when assigning to addressable decls.
23996 2013-04-09 Jeff Law <law@redhat.com>
23998 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23999 when the boolean was created by converting a wider object which
24000 had a boolean range.
24002 2013-04-09 Richard Biener <rguenther@suse.de>
24004 * tree-vectorizer.h (slp_void_p): Remove.
24005 (slp_tree): Typedef before _slp_tree declaration.
24006 (struct _slp_tree): Use a vector of slp_tree as children.
24007 (vect_get_place_in_interleaving_chain): Remove.
24008 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
24010 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
24012 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
24013 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
24014 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
24015 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
24016 Use slp_node instead of slp_void_p and adjust.
24018 2013-04-09 Richard Biener <rguenther@suse.de>
24020 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
24021 work that is not necessary.
24023 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24025 PR tree-optimization/56854
24026 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
24027 forward into clobber stmts if it would change MEM_REF lhs into
24030 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
24032 * tree.c (type_hash_lookup, type_hash_add): Make static.
24033 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
24035 2013-04-09 Richard Biener <rguenther@suse.de>
24037 * tree.h (unsave_expr_now): Remove.
24038 * tree-inline.c (mark_local_for_remap_r): Remove.
24039 (unsave_expr_1): Likewise.
24040 (unsave_r): Likewise.
24041 (unsave_expr_now): Likewise.
24042 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
24043 (propagate_tree_value): Likewise.
24045 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
24047 * doc/rtl.texi (sequence): Rewrite documentation to match the
24048 current use of SEQUENCE rtl objects.
24049 * rtl.def (SEQUENCE): Likewise.
24051 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
24052 Update documentation.
24053 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
24054 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
24056 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
24058 2013-04-08 Teresa Johnson <tejohnson@google.com>
24060 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
24061 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
24063 (estimate_edge_size_and_time): Add comment to suggest using rounding
24065 (estimate_node_size_and_time): Ditto.
24066 (remap_edge_change_prob): Use helper rounding divide methods.
24067 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
24068 (gimple_mod_pow2_value_transform): Ditto.
24069 (gimple_mod_subtract_transform): Ditto.
24070 (gimple_ic_transform): Ditto.
24071 (gimple_stringops_transform): Ditto.
24072 * stmt.c (conditional_probability): Ditto.
24073 (emit_case_dispatch_table): Ditto.
24074 * lto-cgraph.c (merge_profile_summaries): Ditto.
24075 * tree-optimize.c (execute_fixup_cfg): Ditto.
24076 * cfgcleanup.c (try_forward_edges): Ditto.
24077 * cfgloopmanip.c (scale_loop_profile): Ditto.
24079 (duplicate_loop_to_header_edge): Ditto.
24080 (lv_adjust_loop_entry_edge): Ditto.
24081 * tree-vect-loop.c (vect_transform_loop): Ditto.
24082 * profile.c (compute_branch_probabilities): Ditto.
24083 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
24084 * lto-streamer-in.c (input_cfg): Ditto.
24085 * gimple-streamer-in.c (input_bb): Ditto.
24086 * ipa-cp.c (update_profiling_info): Ditto.
24087 (update_specialized_profile): Ditto.
24088 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
24089 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
24091 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24092 (compute_trg_info): Ditto.
24093 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
24094 (purge_dead_edges): Ditto.
24095 * loop-unswitch.c (unswitch_loop): Ditto.
24096 * cgraphclones.c (cgraph_clone_edge): Ditto.
24097 (cgraph_clone_node): Ditto.
24098 * tree-inline.c (copy_bb): Ditto.
24099 (copy_edges_for_bb): Ditto.
24100 (initialize_cfun): Ditto.
24101 (copy_cfg_body): Ditto.
24102 (expand_call_inline): Ditto.
24104 2013-04-08 Kai Tietz <ktietz@redhat.com>
24106 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
24107 TARGET_CYGWIN64 by TARGET_64BIT.
24109 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24111 * config/epiphany/epiphany.md (GPR_1): New constant.
24112 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
24113 * config/epiphany/epiphany.c (gen_compare_reg):
24114 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
24115 is already in place.
24116 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
24117 Don't require being called during rtl expansion; If y operlaps r0,
24119 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
24120 (epiphany_expand_epilogue): Likewise.
24122 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
24123 Don't use CC_FPmode for ORDERED / UNORDERED.
24124 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
24126 * config/epiphany/constraints.md (CnL): New constraint.
24127 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
24128 * config/epiphany/predicates.md (add_operand): Allow 1024.
24130 * config/epiphany/epiphany.md (logical_op): New code iterator.
24131 (op_mnc): New code attribute.
24132 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
24133 (mov_f+1, mov_f+2): New peephole2 patterns.
24135 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
24136 (cstoresi4): Also allow re-use of zero result when doing a NE
24137 comparison to a non-zero operand.
24138 Use (clobber (scratch)) for first insn if the gpr output is not needed.
24140 * config/epiphany/epiphany.md (<insn_opname>v2si3):
24141 Use gen_addsi3_i / gen_subsi3_i.
24143 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24147 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
24148 contain anything but clobbers, at most one __builtin_stack_restore,
24149 optionally debug stmts and final resx, and if it has at least one
24150 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
24151 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
24152 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
24153 which isn't defaut definition, remove them.
24154 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
24155 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
24156 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
24157 with MEM_REF LHS with SSA_NAME address.
24159 2013-04-08 Jeff Law <law@redhat.com>
24161 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
24163 2013-04-08 Richard Biener <rguenther@suse.de>
24165 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
24167 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
24168 determined vector type.
24169 (vect_analyze_data_refs): Likewise.
24170 (vect_get_new_vect_var): Adjust.
24171 (vect_create_destination_var): Preserve SSA name versions.
24172 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
24173 not dump anything here.
24175 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24177 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
24178 Add member lr_slot_known.
24179 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
24181 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
24182 Remove code that sets lr_slot_offset according to what a previous
24183 version of epiphany_emit_save_restore used to do.
24184 (epiphany_emit_save_restore): When doing an lr save or restore,
24185 set/verify lr_slot_known and lr_slot_offset.
24187 2013-04-08 Xinyu Qi <xyqi@marvell.com>
24190 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
24193 2013-04-08 Richard Biener <rguenther@suse.de>
24195 * alias.c (find_base_term): Fix thinko in previous change.
24197 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24199 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
24200 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
24201 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
24202 if possible to compute val.
24203 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
24204 For QImode integers don't require anything about precision. Use
24205 const_with_all_bytes_same to find out if the constant doesn't have
24206 repeated bytes in it.
24208 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24210 * config/s390/s390.c (s390_expand_insv): Only accept insertions
24213 2013-04-08 Marek Polacek <polacek@redhat.com>
24215 PR rtl-optimization/48182
24216 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
24219 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24222 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
24223 nuses, make sure we have a label.
24225 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24228 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
24229 (rs6000_emit_swdiv_low_precision): Remove.
24230 (rs6000_emit_swdiv): Rewrite to handle between one and four
24231 iterations of Newton-Raphson generally; modify required number of
24232 iterations for some cases.
24233 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
24235 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
24237 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
24238 set-but-unused variable.
24240 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
24241 basic blocks of released function bodies garbage-collectable.
24243 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
24244 (struct rtl_opt_pass): Add TODO_df_finish.
24246 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
24248 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24250 * config/arm/constraints.md (q): New constraint.
24251 * config/arm/ldrdstrd.md: New file.
24252 * config/arm/arm.md (ldrdstrd.md) New include.
24253 (arm_movdi): Use "q" instead of "r" constraint
24254 for double-word memory access.
24255 (movdf_soft_insn): Likewise.
24256 * config/arm/vfp.md (movdi_vfp): Likewise.
24257 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
24258 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
24259 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
24260 (mem_ok_for_ldrd_strd): Likewise.
24261 (output_move_double): Update assertion.
24263 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24265 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
24267 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24269 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
24270 define_insn_and_split.
24271 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
24273 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24275 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
24276 define_insn_and_split.
24277 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
24278 (shiftsi3_compare): New pattern.
24279 (rrx): New pattern.
24280 * config/arm/unspecs.md (UNSPEC_RRX): New.
24282 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24284 * config/arm/arm.md (negdi_extendsidi): New pattern.
24285 (negdi_zero_extendsidi): Likewise.
24287 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24289 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
24290 define_insn_and_split.
24291 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
24292 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
24294 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24296 * config/arm/arm.md (arm_subdi3): Convert define_insn into
24297 define_insn_and_split.
24298 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
24299 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24301 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24303 * config/arm/arm.md (subsi3_carryin): New pattern.
24304 (subsi3_carryin_const): Likewise.
24305 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24306 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24308 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24310 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24312 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24314 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24315 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24317 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24319 * config/arm/arm.c (arm_expand_builtin): Change fcode
24320 type to unsigned int.
24322 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24324 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24326 2013-04-04 Ian Lance Taylor <iant@google.com>
24328 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24331 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24333 PR middle-end/56729
24334 * df-scan.c (df_insn_delete): Disable failing assert.
24336 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24338 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24339 New function prototype.
24340 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24341 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24342 (arm_builtin_vectorized_function): New function.
24344 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24346 * config/arm/arm_neon_builtins.def: New file.
24347 * config/arm/arm.c (neon_builtin_data): Move contents to
24348 arm_neon_builtins.def.
24349 (enum arm_builtins): Include neon builtin definitions.
24350 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24351 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24353 2013-04-04 Marek Polacek <polacek@redhat.com>
24355 PR tree-optimization/48186
24356 * predict.c (maybe_hot_frequency_p): Return false if
24357 HOT_BB_FREQUENCY_FRACTION is 0.
24358 (cgraph_maybe_hot_edge_p): Likewise.
24360 2013-04-04 Richard Biener <rguenther@suse.de>
24362 PR tree-optimization/56826
24363 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24366 2013-04-04 Richard Biener <rguenther@suse.de>
24368 PR tree-optimization/56213
24369 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24370 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24372 2013-04-04 Richard Biener <rguenther@suse.de>
24374 PR tree-optimization/56837
24375 * tree-loop-distribution.c (classify_partition): For non-zero
24376 values require that the value has the same precision as its
24377 mode to be useful as memset value.
24379 2013-04-03 Nick Clifton <nickc@redhat.com>
24381 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24382 (fmssf4): Use fmsf.s on E3V5 architectures.
24383 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24384 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24386 2013-04-03 Jeff Law <law@redhat.com>
24388 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24389 (lra-eliminations.o): Likewise.
24391 2013-04-03 Teresa Johnson <tejohnson@google.com>
24393 * gcov-io.c (compute_working_sets): Moved most of body of old
24394 compute_working_sets here from profile.c.
24395 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24396 (gcov_working_set_t): Moved typedef here from basic-block.h
24397 (compute_working_set): Declare.
24398 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24399 (get_working_sets): Renamed from compute_working_set,
24400 replace most of body with call to new compute_working_sets.
24401 (get_exec_counts): Replace call to compute_working_sets
24402 to get_working_sets.
24403 * profile.h (get_working_sets): Renamed from compute_working_set.
24404 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24405 to get_working_sets.
24406 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24407 * gcov-dump.c (dump_working_sets): New function.
24409 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24411 * hwint.c (sext_hwi, zext_hwi): New functions.
24412 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24413 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24414 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24415 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24416 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24417 (sext_hwi, zext_hwi): New functions.
24419 2013-04-03 Jeff Law <law@redhat.com>
24421 PR tree-optimization/56799
24422 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24423 back test for widening conversion erroneously dropped in prior change.
24425 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24428 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24429 instead of next_real_insn.
24431 2013-04-03 Marek Polacek <polacek@redhat.com>
24434 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24436 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24439 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24441 (thumb1_output_casesi): Likewise.
24442 (thumb2_output_casesi): Likewise.
24444 2013-04-03 Richard Biener <rguenther@suse.de>
24446 PR tree-optimization/56817
24447 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24449 (tree_unroll_loops_completely_1): ... new function to manually
24450 walk the loop tree, properly defering outer loops of unrolled
24451 loops to later iterations.
24453 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24455 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24456 (vectorizable_load): Likewise.
24457 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24458 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24460 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24462 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24465 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24467 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24469 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24471 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24473 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24475 PR tree-optimization/56790
24476 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24479 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24481 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24483 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24484 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24487 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24490 * tree.h (force_folding_builtin_constant_p): New decl.
24491 * builtins.c (force_folding_builtin_constant_p): New variable.
24492 (fold_builtin_constant_p): Fold immediately also if
24493 force_folding_builtin_constant_p.
24495 2013-04-03 Richard Biener <rguenther@suse.de>
24497 PR tree-optimization/56812
24498 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24499 DRs of the same interleaving chain are independent.
24501 2013-04-02 Jason Merrill <jason@redhat.com>
24503 * gdbinit.in (pbb): Use debug fn.
24505 2013-04-02 Lawrence Crowl <crowl@google.com>
24507 * sese.h (struct ivtype_map_elt_s): Remove unused.
24508 (extern debug_ivtype_map): Remove unused.
24509 (extern eq_ivtype_map_elts): Remove unused.
24510 * sese.c (debug_ivtype_map): Removed unused.
24511 (debug_ivtype_map_1): Removed unused.
24512 (debug_ivtype_elt): Remove unused.
24513 (eq_ivtype_map_elts): Remove unused.
24515 2013-04-02 Kai Tietz <ktietz@redhat.com>
24518 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24519 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24520 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24522 (legitimize_pe_coff_symbol): Likewise.
24523 (is_imported_p): New helper-function.
24524 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24525 for Windows x64 targets.
24526 (ix86_expand_prologue): Optimize for pe-coff targets.
24527 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24528 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24529 medium/large code-model.
24530 (legitimize_pic_address): Likewise.
24531 (legitimize_tls_address): Likewise.
24532 (ix86_expand_call): Likewise.
24533 (x86_output_mi_thunk): Likewise.
24534 (get_dllimport_decl): Add new beimport argument.
24535 (construct_plt_address): Don't assert for x64 pe-coff targets.
24536 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24538 (SYMBOL_FLAG_STUBVAR): New macro.
24539 (SYMBOL_REF_STUBVAR_P): Likewise.
24540 * config/i386/winnt.c (stub_list): New structure.
24541 (stub_head): New local variable.
24542 (i386_pe_record_stub): New function.
24543 (i386_pe_file_end): Emit refptr-stubs.
24545 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24547 PR rtl-optimization/56745
24548 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24549 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24552 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24553 and both of them are MEM_REFs, just compare first argument for
24554 equality and attempt to deal even with differing offsets.
24557 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24558 of gimple_clobber_p to be MEM_REF.
24559 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24560 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24561 after gimplification.
24562 * asan.c (get_mem_ref_of_assignment): Don't instrument
24563 gimple_clobber_p stmts.
24564 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24565 gimple_clobber_p stmt if they have MEM_REF lhs and
24566 are dead because of another gimple_clobber_p stmt.
24567 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24568 gimple_clobber_p stmts like debug stmts.
24569 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24570 that refer to unused VAR_DECLs or uninitialized values.
24571 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24572 gimple_clobber_p stmts if they refer to removed parameters.
24573 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24576 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24578 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24579 using SWI48 mode attribute.
24581 2013-04-02 Wei Mi <wmi@google.com>
24583 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24584 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24585 *<rotate_insn><mode>3_mask in i386.md.
24587 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24589 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24591 2013-04-02 Richard Biener <rguenther@suse.de>
24593 PR tree-optimization/56778
24594 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24595 Runtime alias tests are not supported for gather loads.
24596 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24597 stmts referenced from SSA operands before updating SSA form.
24599 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24600 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24602 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24603 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24604 * config/arm/cortex-a53.md: New file.
24605 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24606 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24607 * config/arm/arm.c (arm_issue_rate): Likewise.
24608 * config/arm/arm-tune.md: Regenerate
24609 * config/arm/arm-tables.opt: Regenerate.
24610 * config/arm/arm-cores.def: Add cortex-a53.
24612 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24614 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24617 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24619 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24620 scalar load/store operations using B/H registers.
24621 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24623 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24625 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24627 * config/aarch64/aarch64.c
24628 (aarch64_simd_scalar_immediate_valid_for_move): New.
24629 * config/aarch64/aarch64-protos.h
24630 (aarch64_simd_scalar_immediate_valid_for_move): New.
24631 * config/aarch64/constraints.md (Dh, Dq): New.
24632 * config/aarch64/iterators.md (hq): New.
24634 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24636 * reorg.c (get_branch_condition): Deal with conditional returns.
24637 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24639 2013-04-01 Wei Mi <wmi@google.com>
24641 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24642 Truncate operand 2 using %b asm operand modifier.
24643 (*<shift_insn><mode>3_mask): Ditto.
24644 (*<rotate_insn><mode>3_mask): Ditto.
24646 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24648 PR middle-end/56798
24649 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24651 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24653 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24655 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24657 2013-03-30 Lawrence Crowl <crowl@google.com>
24659 * dse.c (clear_alias_sets): Remove never set.
24660 (disqualified_clear_alias_sets): Remove never set.
24661 (clear_alias_mode_pool): Remove never set.
24662 (dse_step0): Remove condition that is never true.
24663 (canon_address): Remove condition that is never true.
24664 (dse_step7): Remove condition that is never true.
24665 (rest_of_handle_dse): Remove condition that is never true.
24666 (rest_of_handle_dse::did_global): Remove never read from above.
24667 (dse_step2_spill): Remove never called from above.
24668 (dse_step5_spill): Remove never called from above.
24670 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24672 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24673 JUMP_TABLE_DATA changes.
24674 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24675 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24676 (Insns) <jump_table_data>: New entry.
24677 * doc/tm.texi: Regenerate.
24679 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24681 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24682 for table jump at the end of a basic block using tablejump_p.
24683 * targhooks.c (default_invalid_within_doloop): Likewise.
24684 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24685 target hook implementation that is identical to the default hook.
24686 (rs6000_invalid_within_doloop): Remove.
24688 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24689 unused variable from tablejump_p call.
24691 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24692 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24693 (INSN_DELETED_P): Likewise.
24694 (emit_jump_table_data): New prototype.
24695 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24696 after 4th as unused.
24697 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24698 * sched-vis.c (print_insn): Likewise.
24699 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24700 insn for compatibility with back ends that use next_active_insn to
24701 identify jump table data.
24702 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24703 (remove_insn): Likewise.
24704 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24706 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24707 (emit_jump_table_data): New function.
24709 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24710 basic block, a JUMP_TABLE_DATA never is.
24711 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24712 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24713 off from code handling real insns.
24714 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24715 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24716 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24717 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24718 is not a NONDEBUG_INSN_P.
24719 * ira-costs.c (scan_one_insn): Likewise.
24720 * jump.c (mark_all_labels): Likewise.
24721 (mark_jump_label_1): Likewise.
24722 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24723 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24725 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24726 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24727 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24728 JUMP_TABLE_DATA_P insns.
24729 (calculate_elim_costs_all_insns): Likewise.
24730 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24731 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24732 (delete_output_reload): Code style fixups.
24733 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24734 insn flags on this non-insn.
24735 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24736 as scheduling barriers, for pre-change compatibility.
24737 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24738 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24740 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24741 redundant JUMP_TABLE_DATA_P test.
24742 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24743 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24744 (frv_for_each_packet): Likewise.
24745 * config/i386/i386.c (min_insn_size): Likewise.
24746 (ix86_avoid_jump_mispredicts): Likewise.
24747 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24748 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24749 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24750 (mips16_insn_length): Robustify.
24751 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24752 (mips16_split_long_branches): Likewise.
24753 * config/pa/pa.c (pa_combine_instructions): Likewise.
24754 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24755 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24756 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24757 as contributing to pool range lengths.
24758 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24759 Remove redundant JUMP_TABLE_DATA_P test.
24760 (sh_loop_align): Likewise.
24761 (split_branches): Likewise.
24762 (sh_insn_length_adjustment): Likewise.
24763 * config/spu/spu.c (get_branch_target): Likewise.
24765 2013-03-29 Jan Hubicka <jh@suse.cz>
24767 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24768 gcov streaming; stream hot bb threshold to ltrans.
24769 * predict.c (get_hot_bb_threshold): Break out from ....
24770 (maybe_hot_count_p): ... here.
24771 (set_hot_bb_threshold): New function.
24772 * lto-section-in.c (lto_section_name): Add profile.
24773 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24774 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24775 and data-streamer.h
24776 (histogram_entry): New structure.
24777 (histogram, histogram_pool): New global vars.
24778 (histogram_hash): New structure.
24779 (histogram_hash::hash): New method.
24780 (histogram_hash::equal): Likewise.
24781 (account_time_size): New function.
24782 (cmp_counts): New function.
24783 (dump_histogram): New function.
24784 (ipa_profile_generate_summary): New function.
24785 (ipa_profile_write_summary): New function.
24786 (ipa_profile_read_summary): New function.
24787 (ipa_profile): Decide on threshold.
24788 (pass_ipa_profile): Add ipa_profile_write_summary and
24789 ipa_profile_read_summary.
24790 * Makefile.in (ipa.o): Update dependencies.
24791 * lto-streamer.h (LTO_section_ipa_profile): New section.
24793 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24795 * tree.h (VAR_P): New.
24797 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24800 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24802 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24804 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24805 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24806 (control_flow_insn_p): Likewise.
24807 * cfgrtl.c (duplicate_insn_chain): Likewise.
24808 * final.c (get_attr_length_1): Likewise.
24809 (shorten_branches): Likewise.
24810 (final_scan_insn): Likewise.
24811 * function.c (instantiate_virtual_regs): Likewise.
24812 * gcse.c (insert_insn_end_basic_block): Likewise.
24813 * ira-costs.c (scan_one_insn): Likewise.
24814 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24815 * lra.c (check_rtl): Likewise.
24816 * reload1.c (elimination_costs_in_insn): Likewise.
24817 * reorg.c (follow_jumps): Likewise.
24819 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24820 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24821 (thumb_far_jump_used_p): Likewise.
24822 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24823 (workaround_speculation): Likewise.
24824 (add_sched_insns_for_speculation): Likewise.
24825 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24826 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24827 (frv_for_each_packet): Likewise.
24828 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24829 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24830 (final_emit_insn_group_barriers): Likewise.
24831 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24832 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24833 (mips16_insn_length): Likewise.
24834 * config/pa/pa.c (pa_reorg): Likewise.
24835 (pa_combine_instructions): Likewise.
24836 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24837 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24838 (sh_reorg): Likewise.
24839 (split_branches): Likewise.
24840 * config/spu/spu.c (get_branch_target): Likewise.
24842 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24845 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24847 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24848 Fix declaration name.
24850 2013-03-28 Lawrence Crowl <crowl@google.com>
24852 * graphds.h (struct graph.indicies): Remove unused.
24853 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24854 (SCOP_ORIGINAL_PDDRS): Remove unused.
24855 * sese.h (extern insert_loop_close_phis): Removed unused.
24856 (extern insert_guard_phis): Removed unused.
24857 (extern ivtype_map_elt_info): Removed unused.
24858 (new_ivtype_map_elt): Removed unused.
24859 * sese.c (ivtype_map_elt_info): Removed unused.
24861 2013-03-28 Lawrence Crowl <crowl@google.com>
24863 * Makefile.in: Add several missing include dependences.
24865 (test-dump.o): New. This object is not added to any executable,
24866 but is present for ad-hoc testing.
24868 (debug (const bitmap_head_def &)): New.
24869 (debug (const bitmap_head_def *)): New.
24871 (extern debug (const bitmap_head_def &)): New.
24872 (extern debug (const bitmap_head_def *)): New.
24874 (debug (edge_def &)): New.
24875 (debug (edge_def *)): New.
24877 (debug (basic_block_def &)): New.
24878 (debug (basic_block_def *)): New.
24880 (dump_node (const_tree, int, FILE *)): Correct source file.
24882 (debug (die_struct &)): New.
24883 (debug (die_struct *)): New.
24885 (extern debug (die_struct &)): New.
24886 (extern debug (die_struct *)): New.
24887 * gimple-pretty-print.c
24888 (debug (gimple_statement_d &)): New.
24889 (debug (gimple_statement_d *)): New.
24890 * gimple-pretty-print.h
24891 (extern debug (gimple_statement_d &)): New.
24892 (extern debug (gimple_statement_d *)): New.
24894 (debug (ira_allocno_copy &)): New.
24895 (debug (ira_allocno_copy *)): New.
24896 (debug (ira_allocno &)): New.
24897 (debug (ira_allocno *)): New.
24899 (extern debug (ira_allocno_copy &)): New.
24900 (extern debug (ira_allocno_copy *)): New.
24901 (extern debug (ira_allocno &)): New.
24902 (extern debug (ira_allocno *)): New.
24904 (debug (live_range &)): New.
24905 (debug (live_range *)): New.
24907 (debug (lra_live_range &)): New.
24908 (debug (lra_live_range *)): New.
24910 (debug (lra_live_range &)): New.
24911 (debug (lra_live_range *)): New.
24913 (debug (omega_pb_d &)): New.
24914 (debug (omega_pb_d *)): New.
24916 (extern debug (omega_pb_d &)): New.
24917 (extern debug (omega_pb_d *)): New.
24919 (debug (const rtx_def &)): New.
24920 (debug (const rtx_def *)): New.
24922 (debug_tree (tree): Move within file.
24923 (debug_raw (const tree_node &)): New.
24924 (debug_raw (const tree_node *)): New.
24925 (dump_tree_via_hooks (const tree_node *, int)): New.
24926 (debug (const tree_node &)): New.
24927 (debug (const tree_node *)): New.
24928 (debug_verbose (const tree_node &)): New.
24929 (debug_verbose (const tree_node *)): New.
24930 (debug_head (const tree_node &)): New.
24931 (debug_head (const tree_node *)): New.
24932 (debug_body (const tree_node &)): New.
24933 (debug_body (const tree_node *)): New.
24934 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24935 (debug (vec<tree, va_gc> &)): New.
24936 (debug (vec<tree, va_gc> *)): New.
24938 (extern debug (const rtx_def &)): New.
24939 (extern debug (const rtx_def *)): New.
24941 (debug_raw (simple_bitmap_def &)): New.
24942 (debug_raw (simple_bitmap_def *)): New.
24943 (debug (simple_bitmap_def &)): New.
24944 (debug (simple_bitmap_def *)): New.
24946 (extern debug (simple_bitmap_def &)): New.
24947 (extern debug (simple_bitmap_def *)): New.
24948 (extern debug_raw (simple_bitmap_def &)): New.
24949 (extern debug_raw (simple_bitmap_def *)): New.
24951 (debug (vinsn_def &)): New.
24952 (debug (vinsn_def *)): New.
24953 (debug_verbose (vinsn_def &)): New.
24954 (debug_verbose (vinsn_def *)): New.
24955 (debug (expr_def &)): New.
24956 (debug (expr_def *)): New.
24957 (debug_verbose (expr_def &)): New.
24958 (debug_verbose (expr_def *)): New.
24959 (debug (vec<rtx> &)): New.
24960 (debug (vec<rtx> *)): New.
24962 (extern debug (vinsn_def &)): New.
24963 (extern debug (vinsn_def *)): New.
24964 (extern debug_verbose (vinsn_def &)): New.
24965 (extern debug_verbose (vinsn_def *)): New.
24966 (extern debug (expr_def &)): New.
24967 (extern debug (expr_def *)): New.
24968 (extern debug_verbose (expr_def &)): New.
24969 (extern debug_verbose (expr_def *)): New.
24970 (extern debug (vec<rtx> &)): New.
24971 (extern debug (vec<rtx> *)): New.
24973 (_list_iter_cond_expr): Make inline instead of static.
24975 (debug (sreal &)): New.
24976 (debug (sreal *)): New.
24978 (extern debug (sreal &)): New.
24979 (extern debug (sreal *)): New.
24981 (extern debug_raw (const tree_node &)): New.
24982 (extern debug_raw (const tree_node *)): New.
24983 (extern debug (const tree_node &)): New.
24984 (extern debug (const tree_node *)): New.
24985 (extern debug_verbose (const tree_node &)): New.
24986 (extern debug_verbose (const tree_node *)): New.
24987 (extern debug_head (const tree_node &)): New.
24988 (extern debug_head (const tree_node *)): New.
24989 (extern debug_body (const tree_node &)): New.
24990 (extern debug_body (const tree_node *)): New.
24991 (extern debug (vec<tree, va_gc> &)): New.
24992 (extern debug (vec<tree, va_gc> *)): New.
24994 (debug (struct loop &)): New.
24995 (debug (struct loop *)): New.
24996 (debug_verbose (struct loop &)): New.
24997 (debug_verbose (struct loop *)): New.
24998 * tree-dump.c: Add header dependence.
25000 (extern debug (struct loop &)): New.
25001 (extern debug (struct loop *)): New.
25002 (extern debug_verbose (struct loop &)): New.
25003 (extern debug_verbose (struct loop *)): New.
25005 (debug (data_reference &)): New.
25006 (debug (data_reference *)): New.
25007 (debug (vec<data_reference_p> &)): New.
25008 (debug (vec<data_reference_p> *)): New.
25009 (debug (vec<ddr_p> &)): New.
25010 (debug (vec<ddr_p> *)): New.
25012 (extern debug (data_reference &)): New.
25013 (extern debug (data_reference *)): New.
25014 (extern debug (vec<data_reference_p> &)): New.
25015 (extern debug (vec<data_reference_p> *)): New.
25016 (extern debug (vec<ddr_p> &)): New.
25017 (extern debug (vec<ddr_p> *)): New.
25019 (debug (pt_solution &)): New.
25020 (debug (pt_solution *)): New.
25022 (extern debug (pt_solution &)): New.
25023 (extern debug (pt_solution *)): New.
25025 (debug (_var_map &)): New.
25026 (debug (_var_map *)): New.
25027 (debug (tree_live_info_d &)): New.
25028 (debug (tree_live_info_d *)): New.
25030 (extern debug (_var_map &)): New.
25031 (extern debug (_var_map *)): New.
25032 (extern debug (tree_live_info_d &)): New.
25033 (extern debug (tree_live_info_d *)): New.
25035 2013-03-28 Jan Hubicka <jh@suse.cz>
25037 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
25039 2013-03-28 Ian Bolton <ian.bolton@arm.com>
25041 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
25042 record only when desired or required.
25044 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
25046 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
25047 *vec_extractv2di_1_rex64. Use x64 isa attribute.
25049 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25051 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
25052 (*andsi3_compare0_uxtw): New pattern.
25053 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
25054 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
25056 2013-03-28 Jan Hubicka <jh@suse.cz>
25058 * data-streamer-in.c (streamer_read_gcov_count): New function.
25059 * gimple-streamer-out.c: Include value-prof.h.
25060 (output_gimple_stmt): Output histogram.
25061 (output_bb): Use streamer_write_gcov_count.
25062 * value-prof.c: Include data-streamer.h
25063 (dump_histogram_value): Add HIST_TYPE_MAX.
25064 (stream_out_histogram_value): New function.
25065 (stream_in_histogram_value): New function.
25066 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
25067 (stream_out_histogram_value, stream_in_histogram_value): Declare.
25068 * data-streamer-out.c (streamer_write_gcov_count): New function.
25069 (streamer_write_gcov_count_stream): New function.
25070 * lto-cgraph.c (lto_output_edge): Update counter streaming.
25071 (lto_output_node): Likewise.
25072 (input_node, input_edge): Likewise.
25073 * lto-streamer-out.c (output_cfg): Update streaming.
25074 * lto-streamer-in.c (input_cfg): Likewise.
25075 * data-streamer.h (streamer_write_gcov_count,
25076 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
25077 * gimple-streamer-in.c: Include value-prof.h
25078 (input_gimple_stmt): Input histograms.
25079 (input_bb): Update profile streaming.
25081 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25083 * genmodes.c (emit_max_int): New function.
25084 (emit_insn_modes_h): Added call to emit_max_function.
25085 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
25087 * machmode.def: Fixed comment.
25089 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25091 * combine.c (try_combine): Removed useless assert.
25092 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
25094 2013-03-28 Marek Polacek <polacek@redhat.com>
25095 Richard Biener <rguenther@suse.de>
25097 PR tree-optimization/56695
25098 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
25099 build signed result of a vector comparison.
25100 * tree-cfg.c (verify_gimple_comparison): Check that a result
25101 of a vector comparison has signed type.
25103 2013-03-28 Richard Biener <rguenther@suse.de>
25105 PR tree-optimization/37021
25106 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
25107 do not restrict gaps between groups.
25108 * tree-vect-stmts.c (vectorizable_load): Properly account for
25109 a gap between groups.
25111 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
25113 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
25114 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
25117 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
25119 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
25120 * doc/extend.texi (Named Address Spaces): Ditto.
25121 (Variable Attributes): Ditto.
25123 2013-03-27 Kai Tietz <ktietz@redhat.com>
25125 * config.build: Add support for cygwin x64 target.
25126 * config.gcc: Likewise.
25127 * config.host: Likewise.
25128 * configure.ac: Likewise
25129 * configure: Regenerated.
25131 2013-03-27 Kai Tietz <ktietz@redhat.com>
25133 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
25134 * config/i386/t-cygwin-w64: New file.
25135 * config/i386/cygwin-w64.h: New file.
25136 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
25137 and add support for x64-cygwin target.
25138 (CPP_SPEC): Likewise.
25139 (CXX_WRAP_SPEC_LIST): Undefine before define.
25140 (LIBGCJ_SONAME): Use 15 as version.
25142 2013-03-27 Richard Biener <rguenther@suse.de>
25144 PR tree-optimization/56716
25145 * tree-ssa-structalias.c (perform_var_substitution): Adjust
25146 dumping for ref nodes.
25148 2013-03-27 Martin Jambor <mjambor@suse.cz>
25150 PR tree-optimization/55334
25151 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
25152 restricted pointers to arrays.
25154 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
25156 * Makefile.in (.SUFFIXES): Add .cc.
25157 (.c.o): Apply same recipe for implicit rule .cc.o.
25159 2013-03-27 Richard Biener <rguenther@suse.de>
25161 PR tree-optimization/37021
25162 * tree-vect-data-refs.c (vect_check_strided_load): Allow
25163 REALPART/IMAGPART_EXPRs around the supported refs.
25164 * tree-ssa-structalias.c (find_func_aliases): Assume that
25165 floating-point values are not used to transfer pointers.
25167 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
25169 * target.def (TARGET_HAS_IFUNC_P): New target hook.
25170 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
25171 * doc/tm.texi: Regenerate.
25172 * targhooks.h (default_has_ifunc_p): New.
25173 * targhooks.c (default_has_ifunc_p): Ditto.
25174 * config/linux-protos.h: New file.
25175 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
25176 hook for linux which disables support of indirect functions in android.
25177 * config/linux-android.c: New file.
25178 * config/t-linux-android.c: Ditto.
25179 * config.gcc: Added new object file linux-android.o.
25180 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
25181 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
25182 * varasm.c (do_assemble_alias): Likewise.
25183 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
25184 doesn't support indirect functions.
25185 * configure: Regenerate.
25187 2013-03-27 Bin Cheng <bin.cheng@arm.com>
25190 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
25191 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
25194 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25196 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
25198 2013-03-27 Terry Guo <terry.guo@arm.com>
25200 * config/arm/arm-cores.def: Added core cortex-r7.
25201 * config/arm/arm-tune.md: Regenerated.
25202 * config/arm/arm-tables.opt: Regenerated.
25203 * doc/invoke.texi: Added entry for core cortex-r7.
25205 2013-03-27 Walter Lee <walt@tilera.com>
25207 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
25208 double-decrement of next_scratch_regno.
25210 2013-03-27 Walter Lee <walt@tilera.com>
25212 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
25214 (insn_v1mulus): Ditto.
25215 (insn_v2muls): Ditto.
25217 2013-03-27 Walter Lee <walt@tilera.com>
25219 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
25220 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
25222 2013-03-27 Walter Lee <walt@tilera.com>
25224 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
25225 (*sibcall_value): Ditto.
25227 2013-03-27 Walter Lee <walt@tilera.com>
25229 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
25230 (insn_mnz_v8qi): ... this ...
25231 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
25233 (insn_v<n>mnz): Replaced by ...
25234 (insn_v1mnz): ... this ...
25235 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
25237 (insn_mz_<mode>): Replaced by ...
25238 (insn_mz_v8qi): ... this ...
25239 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
25241 (insn_v<n>mz): Replaced by ...
25242 (insn_v1mz): ... this ...
25243 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
25246 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
25248 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
25250 2013-03-26 Roland McGrath <mcgrathr@google.com>
25252 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
25253 than fprintf with a non-constant, non-format string.
25255 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25257 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
25258 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
25259 operand 0 predicate.
25260 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
25261 attribute. Use general_x64nomem_operand as operand 1 predicate.
25262 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
25263 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25264 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
25265 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25266 (mov<mode>_insv_1): Remove expander. Merge insn with
25267 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
25268 Use general_x64nomem_operand as operand 1 predicate.
25269 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
25270 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
25271 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25272 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
25273 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
25274 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
25275 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25276 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
25277 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
25278 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
25279 (general_x64nomem_operand): Ditto.
25281 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
25283 * config/rtems.opt: Add -pthread option.
25285 2013-03-26 Richard Biener <rguenther@suse.de>
25287 * alias.c (find_base_term): Avoid redundant and not used recursion.
25288 (base_alias_check): Get the initial base term from the caller.
25289 (true_dependence_1): Compute and pass base terms to base_alias_check.
25290 (write_dependence_p): Likewise.
25291 (may_alias_p): Likewise.
25293 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
25295 * config/aarch64/aarch64.c (aarch64_classify_address): Support
25296 PC-relative load in SI modes and above only.
25298 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25300 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25301 * config/arm/iwmmxt.md (WCGR0): Update.
25302 (WCGR1, WCGR2, WCGR3): Likewise.
25304 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25306 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25307 Use x64 and nox64 isa attributes.
25309 2013-03-26 Richard Biener <rguenther@suse.de>
25311 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25312 alignment computations and rely on get_object_alignment_1
25313 for the !TYPE_P case.
25314 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25316 2013-03-26 Walter Lee <walt@tilera.com>
25318 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25319 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25321 2013-03-25 Jeff Law <law@redhat.com>
25323 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25324 check for INTEGRAL_TYPE_P that was missing due to checking in
25325 wrong version of prior patch.
25327 2013-03-25 Walter Lee <walt@tilera.com>
25329 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25330 TILEGX_INSN_SHUFFLEBYTES1.
25331 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25333 (tilegx_builtins): Ditto.
25334 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25336 2013-03-25 Walter Lee <walt@tilera.com>
25338 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25339 (floatunssisf2): New pattern.
25340 (floatsidf2): New pattern.
25341 (floatunssidf2): New pattern.
25343 2013-03-25 Walter Lee <walt@tilera.com>
25345 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25346 tests for constraint J, K, N, P.
25348 2013-03-25 Walter Lee <walt@tilera.com>
25350 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25351 Use indirect/pcrel encoding.
25352 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25355 2013-03-25 Steve Ellcey <sellcey@mips.com>
25357 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25358 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25359 * config/mips/mips.c (mips_option_override): Set IMADD default.
25360 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25361 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25362 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25363 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25364 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25366 2013-03-25 Jeff Law <law@redhat.com>
25368 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25369 slightly to avoid creating and folding useless trees. Simplify
25370 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25372 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25374 * config/i386/i386.md (*zero_extendsidi2): Merge with
25375 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25376 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25377 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25378 targets. Clarify comment.
25380 2013-03-25 Martin Jambor <mjambor@suse.cz>
25382 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25383 pass-through jump functions differently.
25384 (ipa_read_jump_function): Likewise. Also use setter functions to set
25387 2013-03-25 Martin Jambor <mjambor@suse.cz>
25389 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25390 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25392 (ipa_get_indirect_edge_target): New function.
25393 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25394 ipa_get_indirect_edge_target. Update all callers.
25395 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25396 ipa_get_indirect_edge_target_1 instead of calling
25397 ipa_get_indirect_edge_target.
25398 (create_specialized_node): Pass aggvlas to
25399 ipcp_discover_new_direct_edges.
25401 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25403 * config/arm/arm.md (f_sels, f_seld): New types.
25404 (*cmov<mode>): New pattern.
25405 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25408 2013-03-25 Kai Tietz <ktietz@redhat.com>
25410 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25411 POSIX-printf for mingw-hosted builds.
25413 2013-03-25 Richard Biener <rguenther@suse.de>
25415 PR middle-end/56694
25416 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25417 must-not-throw stmt location.
25419 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25421 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25422 Emit load-acquire versions when acq is true.
25423 (arm_emit_store_exclusive): Add rel parameter.
25424 Emit store-release versions when rel is true.
25425 (arm_split_compare_and_swap): Use acquire-release instructions
25427 of barriers when appropriate.
25428 (arm_split_atomic_op): Likewise.
25429 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25430 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25431 (VUNSPEC_SLX): Likewise.
25432 (VUNSPEC_LDA): Likewise.
25433 (VUNSPEC_STL): Likewise.
25434 * config/arm/sync.md (atomic_load<mode>): New pattern.
25435 (atomic_store<mode>): Likewise.
25436 (arm_load_acquire_exclusive<mode>): Likewise.
25437 (arm_load_acquire_exclusivesi): Likewise.
25438 (arm_load_acquire_exclusivedi): Likewise.
25439 (arm_store_release_exclusive<mode>): Likewise.
25441 2013-03-25 Catherine Moore <clm@codesourcery.com>
25443 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25444 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25445 * config/mip/predicates.md (lwsp_swsp_operand,
25446 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25447 sb16_operand, db4_operand, db7_operand, ib3_operand,
25448 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25449 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25450 andi16_operand): New predicates.
25451 * config/mips/mips.md (compression): New attribute.
25452 (enabled): New attribute.
25453 (length): Consider compression in computing length.
25454 (shift_compression): New code attribute.
25455 (*add<mode>3): New operands. Record compression.
25456 (sub<mode>3): Likewise.
25457 (one_cmpl<mode>2): Likewise.
25458 (*and<mode>3): Likewise.
25459 (*ior<mode>3): Likewise.
25460 (unnamed pattern for xor): Likewise.
25461 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25462 (*<optab><mode>3): Likewise.
25463 (*mov<mode>_internal: Likewise.
25464 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25465 (mips_unsigned_immediate_p): New.
25466 (umips_lwsp_swsp_address_p): New.
25467 (m16_based_address_p): New.
25468 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25469 (mips_unsigned_immediate_p): New prototype.
25470 (lwsp_swsp_address_p): New prototype.
25471 (m16_based_address_p): New prototype.
25472 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25473 (mips_signed_immediate_p): New function.
25474 (m16_based_address_p): New function.
25475 (lwsp_swsp_address_p): New function.
25476 (mips_print_operand_punctuation): Recognize short delay slot insns
25477 for microMIPS.add<mode>3"
25479 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25482 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25483 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25485 2013-03-25 Richard Biener <rguenther@suse.de>
25487 PR tree-optimization/56689
25488 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25491 2013-03-25 Richard Biener <rguenther@suse.de>
25493 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25495 (memory_references): Likewise.
25496 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25497 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25498 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25499 (gather_mem_refs_in_loops): Fold into ...
25500 (analyze_memory_references): ... this. Move initialization
25501 to tree_ssa_lim_initialize.
25502 (fill_always_executed_in): Rename to ...
25503 (fill_always_executed_in_1): ... this.
25504 (fill_always_executed_in): Move contains_call computation to
25505 this new function from ...
25506 (tree_ssa_lim_initialize): ... here.
25507 (tree_ssa_lim): Call fill_always_executed_in.
25509 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25511 * postreload.c (reload_combine): Fix code detecting returns.
25513 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25515 * function.c (emit_use_return_register_into_block): On cc0 targets,
25516 do not emit the sequence between cc0 setter and user.
25518 2013-03-25 Kai Tietz <ktietz@redhat.com>
25520 * config/i386/predicates.md (local_symbolic_operand): Interpret
25521 dll-imported symbols as none-local.
25523 2013-03-25 Richard Biener <rguenther@suse.de>
25525 * tree-ssa-loop-im.c (struct depend): Remove.
25526 (struct lim_aux_data): Make depends a vec of gimples.
25527 (free_lim_aux_data): Adjust.
25528 (add_dependency): Likewise.
25529 (set_level): Likewise.
25531 2013-03-25 Richard Biener <rguenther@suse.de>
25533 PR middle-end/56434
25534 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25535 the pointer returned by calls with ECF_MALLOC set.
25537 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25539 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25541 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25543 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25544 using MMXMODE mode iterator.
25545 (*move<mode>_internal): Merge with *movv2sf_internal and
25546 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25548 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25550 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25551 (record_last_mem_set_info): Likewise.
25553 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25554 of XNEWVEC followed by memset.
25555 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25557 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25559 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25560 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25561 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25562 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25563 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25564 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25565 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25566 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25567 BARRIER_P instead of GET_CODE.
25569 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25571 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25572 inaccuracy in the probing code.
25574 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25575 (ctrapdi4): Likewise.
25577 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25579 * calls.c (expand_call): Add missing guard to code handling return
25580 of non-BLKmode structures in MSB.
25581 * function.c (expand_function_end): Likewise.
25583 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25585 * combine.c (try_combine): Adjust comment. Do not add the set of
25586 insn #0 if the destination indirectly is set or dies in insn #2.
25587 Tidy up code to distribute a new note.
25589 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25591 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25592 also for alternatives 16 and 17.
25594 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25596 * config/i386/sse.md (*mov<mode>_internal): Merge with
25597 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25598 Emit insn template depending on type attribute. Use
25599 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25600 movd instead of movq mnemonic for interunit moves. Rewrite mode
25601 attribute calculation. Remove unit attribute calculation.
25602 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25603 Set prefix_data16 attribute for DImode ssemov types.
25604 Use Ym instead of y for SSE-MMX conversion alternatives.
25605 Reorder operand constraints.
25607 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25609 * df.h (df_insn_delete): Adjust prototype.
25610 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25611 and let it decide whether mark the basic block dirty.
25612 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25613 * df-scan.c (df_insn_info_delete): New helper function, split
25614 off from df_insn_delete.
25615 (df_scan_free_bb_info): Use it.
25616 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25618 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25619 that the insn is actually an insn and it has a non-NULL basic block.
25620 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25622 2013-03-22 Richard Biener <rguenther@suse.de>
25624 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25626 (mem_ref_alloc): Do not allocate them.
25627 (refs_independent_p): Do not query or maintain a cache.
25629 2013-03-22 Richard Biener <rguenther@suse.de>
25631 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25632 (gather_mem_refs_in_loops): Do not compute it.
25633 (analyze_memory_references): Do not allocate it.
25634 (tree_ssa_lim_finalize): Do not free it.
25635 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25637 2013-03-22 Richard Biener <rguenther@suse.de>
25639 * is-a.h (as_a): Use gcc_checking_assert.
25641 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25643 * config/aarch64/aarch64.c (aarch64_print_operand): New
25644 format specifier for printing a constant in hex.
25645 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25646 format specifier for printing second operand.
25648 2013-03-22 Richard Biener <rguenther@suse.de>
25650 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25652 (gather_mem_refs_in_loops): Perform store accumulation here.
25653 (create_vop_ref_mapping_loop): Remove.
25654 (create_vop_ref_mapping): Likewise.
25655 (analyze_memory_references): Initialize refs_stored_in_loop.
25656 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25657 (record_indep_loop): Remove.
25658 (record_dep_loop): New function.
25659 (ref_indep_loop_p_1): Adjust to only walk over references
25660 in the loop, not its subloops.
25661 (ref_indep_loop_p): Rename to ...
25662 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25663 maintaining a more fine-grained cache.
25664 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25665 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25667 2013-03-22 Richard Biener <rguenther@suse.de>
25669 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25670 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25671 aggregate mem_ref_loc.
25672 (free_mem_ref_locs): Inline into ...
25673 (memref_free): ... this and adjust.
25674 (mem_ref_alloc): Adjust.
25675 (mem_ref_locs_alloc): Remove.
25676 (record_mem_ref_loc): Adjust.
25677 (get_all_locs_in_loop): Rewrite into ...
25678 (for_all_locs_in_loop): ... this iterator.
25679 (rewrite_mem_ref_loc): New functor.
25680 (rewrite_mem_refs): Use for_all_locs_in_loop.
25681 (sm_set_flag_if_changed): New functor.
25682 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25683 (ref_always_accessed): New functor.
25684 (ref_always_accessed_p): Use for_all_locs_in_loop.
25686 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25688 * tree-pass.h (PROP_gimple_lvec): New.
25689 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25690 (init_optimization_passes): Move pass_lower_vector.
25691 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25693 (pass_lower_vector): Provide PROP_gimple_lvec.
25694 (pass_lower_vector_ssa): Likewise.
25695 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25697 2013-03-21 Mark Wielaard <mjw@redhat.com>
25699 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25701 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25703 * config/i386/i386.md (*movdi_internal): Disparage slightly
25704 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25705 conversion alternatives.
25707 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25709 PR middle-end/48087
25710 * diagnostic.def (DK_WERROR): New kind.
25711 * diagnostic.h (werrorcount): Define.
25712 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25713 promoted to DK_ERROR, increment DK_WERROR counter instead of
25715 * toplev.c (toplev_main): Call print_ignored_options even if
25716 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25717 even if just werrorcount is non-zero.
25720 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25722 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25723 (string_cst_pool_decl): New function.
25724 (optimize_one_addr_into_implicit_ptr): New function.
25725 (resolve_addr_in_expr): Optimize DWARF location expression
25726 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25727 which doesn't live in memory, but has DW_AT_location or
25728 DW_AT_const_value, or refers to a string literal, into
25729 DW_OP_GNU_implicit_pointer.
25730 (optimize_location_into_implicit_ptr): New function.
25731 (resolve_addr): If removing DW_AT_location of a variable because
25732 it was DW_OP_addr of address of the variable, but the variable doesn't
25733 live in memory, try to emit const value attribute for the initializer.
25735 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25737 * tree.h (VECTOR_TYPE_P): New macro.
25738 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25739 TYPE_MODE): Use it.
25740 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25741 VEC_COND_EXPR cannot be lvalues.
25742 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25744 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25746 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25747 Restrict the transformation to equal modes.
25749 2013-03-21 Richard Biener <rguenther@suse.de>
25751 PR tree-optimization/39326
25752 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25753 (MEM_ANALYZABLE): Adjust.
25754 (record_mem_ref_loc): Move bitmap ops ...
25755 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25756 unanalyzable refs, do not record locations for it.
25757 (analyze_memory_references): Allocate ref zero as shared
25759 (refs_independent_p): Do not test for unanalyzed mems here.
25760 (ref_indep_loop_p_1): Special-case disambiguation against
25761 the unanalyzed ref.
25762 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25764 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25766 * config/arm/arm-protos.h (tune_params): Add
25767 prefer_neon_for_64bits field.
25768 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25769 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25770 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25771 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25772 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25773 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25774 (arm_option_override): Handle -mneon-for-64bits new option.
25775 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25776 (prefer_neon_for_64bits): Declare new variable.
25777 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25778 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25779 (arch_enabled): Handle new arch types. Remove support for onlya8
25781 (one_cmpldi2): Use new arch names.
25782 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25783 * config/arm/arm.opt (mneon-for-64bits): Add option.
25784 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25785 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25786 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25788 * doc/invoke.texi (-mneon-for-64bits): Document.
25790 2013-03-21 Richard Biener <rguenther@suse.de>
25792 PR tree-optimization/39326
25793 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25794 (sort_bbs_in_loop_postorder_cmp): New function.
25795 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25797 2013-03-21 Richard Biener <rguenther@suse.de>
25799 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25800 (vect_insert_into_interleaving_chain): Likewise.
25801 (vect_drs_dependent_in_basic_block): Inline ...
25802 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25804 (vect_analyze_data_ref_dependence): ... here. Simplify.
25805 (vect_check_interleaving): Simplify.
25806 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25807 (vect_slp_analyze_data_ref_dependences): ... this new function.
25808 (dr_group_sort_cmp): New function.
25809 (vect_analyze_data_ref_accesses): Compute data-reference groups
25810 here instead of in vect_analyze_data_ref_dependence. Use
25811 a more efficient algorithm.
25812 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25813 vect_slp_analyze_data_ref_dependences. Call
25814 vect_analyze_data_ref_accesses earlier.
25815 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25816 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25817 (vect_slp_analyze_data_ref_dependences): New prototype.
25819 2013-03-21 Richard Biener <rguenther@suse.de>
25821 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25822 ref is stored in the loop.
25823 (find_refs_for_sm): Walk only over all stores.
25824 (store_motion_loop): Allocate from lim_bitmap_obstack.
25825 (store_motion): Likewise.
25827 2013-03-21 Richard Biener <rguenther@suse.de>
25829 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25830 Update virtual SSA form.
25832 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25834 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25835 * configure: Regenerate.
25836 * config.in: Regenerate.
25837 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25838 if !HAVE_LD_EH_FRAME_CIEV3.
25840 2013-03-21 Richard Biener <rguenther@suse.de>
25842 * tree-cfg.c (verify_expr_no_block): New function.
25843 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25844 nor DECL_VALUE_EXPR have locations with associated blocks.
25845 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25846 (clear_unused_block_pointer): Remove code dealing with
25847 blocks in DECL_DEBUG_EXPR locations.
25849 2013-03-21 Richard Biener <rguenther@suse.de>
25851 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25852 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25853 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25854 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25855 instead of DECL_DEBUG_EXPR_IS_FROM.
25856 * gimplify.c (gimplify_modify_expr): Likewise.
25857 * tree-cfg.c (verify_expr_location_1): Likewise.
25858 * tree-complex.c (create_one_component_var): Likewise.
25859 * tree-sra.c (create_access_replacement): Likewise.
25860 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25861 (clear_unused_block_pointer): Likewise.
25862 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25863 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25864 * var-tracking.c (var_debug_decl): Likewise.
25865 (track_expr_p): Likewise.
25866 * tree-inline.c (add_local_variables): Likewise. Set
25867 DECL_HAS_DEBUG_EXPR_P after copying it.
25868 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25869 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25871 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25874 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25875 * configure: Regenerate.
25876 * config.in: Regenerate.
25877 * config/i386/i386.md (*movdf_internal): Use
25878 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25879 movd instead of movq mnemonic for interunit moves.
25880 (*movdi_internal): Ditto.
25882 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25884 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25885 (abd<mode>_3): New pattern.
25886 (aba<mode>_3): New pattern.
25887 (fabd<mode>_3): New pattern.
25889 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25891 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25892 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25893 occurrence of REGISTER_PREFIX as its empty string.
25895 2013-03-20 Jeff Law <law@redhat.com>
25897 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25898 addititional equivalences for equality comparisons between an SSA_NAME
25899 and a constant where the SSA_NAME was set from a widening conversion.
25901 2013-03-20 Walter Lee <walt@tilera.com>
25903 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25905 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25907 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25908 depending on type attribute.
25909 (*movti_internal): Ditto.
25910 (*movtf_internal): Ditto.
25911 (*movxf_internal): Ditto.
25912 (*movdf_internal): Ditto.
25913 (*movsf_internal): Ditto.
25915 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25917 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25918 maybe_vex for sselog1 and ssemov types.
25919 (*movdi_internal): Reorder operand constraints.
25920 (*movsi_internal): Ditto. Set prefix attribute to
25921 maybe_vex for sselog1 and ssemov types.
25922 (*movtf_internal): Set prefix attribute to maybe_vex
25923 for sselog1 and ssemov types.
25924 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25925 DImode ssemov types. Reorder operand constraints.
25926 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25927 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25928 attribute for SImode ssemov types. Reorder operand constraints.
25930 2013-03-20 Martin Jambor <mjambor@suse.cz>
25932 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25933 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25935 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25937 * config/rs6000/predicates.md (indexed_address, update_address_mem
25938 update_indexed_address_mem): New predicates.
25939 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25940 attribute for load/store instructions.
25941 * config/rs6000/dfp.md (movsd_store): Likewise.
25942 (movsd_load): Likewise.
25943 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25944 (unnamed HI->DI extend define_insn): Likewise.
25945 (unnamed SI->DI extend define_insn): Likewise.
25946 (unnamed QI->SI extend define_insn): Likewise.
25947 (unnamed QI->HI extend define_insn): Likewise.
25948 (unnamed HI->SI extend define_insn): Likewise.
25949 (unnamed HI->SI extend define_insn): Likewise.
25950 (extendsfdf2_fpr): Likewise.
25951 (movsi_internal1): Likewise.
25952 (movsi_internal1_single): Likewise.
25953 (movhi_internal): Likewise.
25954 (movqi_internal): Likewise.
25955 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25956 attribute for load/store instructions.
25957 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25959 (mov<mode>_softfloat): Likewise.
25960 (mov<mode>_hardfloat32): Likewise.
25961 (mov<mode>_hardfloat64): Likewise.
25962 (mov<mode>_softfloat64): Likewise.
25963 (movdi_internal32): Likewise.
25964 (movdi_internal64): Likewise.
25965 (probe_stack_<mode>): Likewise.
25967 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25969 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25970 floating point, and decimal floating point to reload iterator.
25972 * config/rs6000/constraints.md (wl constraint): New constraints to
25973 return FLOAT_REGS if certain options are used to reduce the number
25974 of separate patterns that exist in the file.
25975 (wx constraint): Likewise.
25976 (wz constraint): Likewise.
25978 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25979 -mdebug=reg, print wg, wl, wx, and wz constraints.
25980 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25981 Initialize the reload functions for 64-bit binary/decimal floating
25983 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25984 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25985 create the buffer on the stack to overcome not having a 32-bit
25987 (rs6000_emit_move): Likewise.
25988 (rs6000_secondary_memory_needed_rtx): Likewise.
25989 (rs6000_alloc_sdmode_stack_slot): Likewise.
25990 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25991 via xxlxor, just like DFmode 0.0.
25993 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25994 define as 1 if we are running on a power7 or newer.
25995 (enum r6000_reg_class_enum): Add new constraints.
25997 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25998 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25999 with other moves by using conditional constraits (wg). Use LFIWZX
26000 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
26001 (movsd splitter): Likewise.
26002 (movsd_hardfloat): Likewise.
26003 (movsd_softfloat): Likewise.
26005 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
26006 binary and decimal floating point moves.
26007 (fmove_ok): New attributes to combine binary and decimal floating
26008 point moves, and to combine power6x (mfpgpr) moves along normal
26010 (real_value_to_target): Likewise.
26011 (f32_lr): Likewise.
26012 (f32_lm): Likewise.
26013 (f32_li): Likewise.
26014 (f32_sr): Likewise.
26015 (f32_sm): Likewise.
26016 (f32_si): Likewise.
26017 (movsf): Combine binary and decimal floating point moves. Combine
26018 power6x (mfpgpr) moves with other moves by using conditional
26019 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
26020 (mov<mode> for SFmode/SDmode); Likewise.
26021 (SFmode/SDmode splitters): Likewise.
26022 (movsf_hardfloat): Likewise.
26023 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
26024 (movsf_softfloat): Likewise.
26025 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
26027 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
26028 wx and wz constraints.
26030 * config/rs6000/constraints.md (wg constraint): New constraint to
26031 return FLOAT_REGS if -mmfpgpr (power6x) was used.
26033 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
26036 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26037 -mdebug=reg, print wg, wl, wx, and wz constraints.
26038 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26039 Initialize the reload functions for 64-bit binary/decimal floating
26041 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26042 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26043 create the buffer on the stack to overcome not having a 32-bit
26045 (rs6000_emit_move): Likewise.
26046 (rs6000_secondary_memory_needed_rtx): Likewise.
26047 (rs6000_alloc_sdmode_stack_slot): Likewise.
26048 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26049 via xxlxor, just like DFmode 0.0.
26051 * config/rs6000/dfp.md (movdd): Delete, combine with binary
26052 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26053 with other moves by using conditional constraits (wg). Use LFIWZX
26054 and STFIWX for loading SDmode on power7.
26055 (movdd splitters): Likewise.
26056 (movdd_hardfloat32): Likewise.
26057 (movdd_softfloat32): Likewise.
26058 (movdd_hardfloat64_mfpgpr): Likewise.
26059 (movdd_hardfloat64): Likewise.
26060 (movdd_softfloat64): Likewise.
26062 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
26063 64-bit binary and decimal floating point moves.
26064 (FMOVE64X): Likewise.
26065 (movdf): Combine 64-bit binary and decimal floating point moves.
26066 Combine power6x (mfpgpr) moves with other moves by using
26067 conditional constraits (wg).
26068 (mov<mode> for DFmode/DDmode): Likewise.
26069 (DFmode/DDmode splitters): Likewise.
26070 (movdf_hardfloat32): Likewise.
26071 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
26072 (movdf_softfloat32): Likewise.
26073 (movdf_hardfloat64_mfpgpr): Likewise.
26074 (movdf_hardfloat64): Likewise.
26075 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
26076 (movdf_softfloat64): Likewise.
26077 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
26078 (reload_<mode>_load): Move to later in the file so they aren't in
26079 the middle of the floating point move insns.
26080 (reload_<mode>_store): Likewise.
26082 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
26085 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
26086 constraint if -mdebug=reg.
26087 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
26088 Enable using dd reload support if needed.
26090 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
26091 binary and decimal floating point moves in rs6000.md.
26092 (movtd_internal): Likewise.
26094 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
26095 decimal floating point moves.
26097 (movtf_internal): Likewise.
26098 (mov<mode>_internal, TDmode/TFmode): Likewise.
26099 (movtf_softfloat): Likewise.
26100 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
26102 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
26103 movdi_internal64, using wg constraint for move direct operations.
26104 (movdi_internal64): Likewise.
26106 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
26107 MODES_TIEABLE_P for selected modes. Print the numerical value of
26108 the various virtual registers. Use GPR/FPR first/last values,
26109 instead of hard coding the register numbers. Print which modes
26110 have reload functions registered.
26111 (rs6000_option_override_internal): If -mdebug=reg, trace the options
26112 settings before/after setting cpu, target and subtarget settings.
26113 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
26114 and for secondary reload failures in rs6000_secondary_reload_inner.
26115 (rs6000_secondary_reload_fail): Likewise.
26116 (rs6000_secondary_reload_inner): Likewise.
26118 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
26119 macros for first/last GPR and FPR registers.
26120 (LAST_GPR_REGNO): Likewise.
26121 (FIRST_FPR_REGNO): Likewise.
26122 (LAST_FPR_REGNO): Likewise.
26124 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
26125 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
26126 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
26127 (vcond<mode><mode>): Likewise.
26128 (vcondu<mode><mode>): Likewise.
26129 (vector_gtu<mode>): Likewise.
26130 (vector_gte<mode>): Likewise.
26131 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
26132 to prevent the compiler from converting DImode operations to TImode.
26133 (ior<mode>3): Likewise.
26134 (and<mode>3): Likewise.
26135 (one_cmpl<mode>2): Likewise.
26136 (nor<mode>3): Likewise.
26137 (andc<mode>3): Likewise.
26139 * config/rs6000/constraints.md (wt constraint): New constraint
26140 that returns VSX_REGS if TImode is allowed in VSX registers.
26142 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
26143 constant under VSX.
26145 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
26146 similar to TImode, but it is restricted to being in the GPRs.
26148 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
26149 TImode to occupy a single VSX register.
26151 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
26152 -mvsx-timode for power7/power8.
26153 (power7 cpu): Likewise.
26154 (power8 cpu): Likewise.
26156 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
26157 sure that TFmode/TDmode take up two registers if they are ever
26158 allowed in the upper VSX registers.
26159 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
26161 (rs6000_init_hard_regno_mode_ok): Likewise.
26162 (rs6000_debug_reg_global): Add debugging for PTImode and wt
26163 constraint. Print if LRA is turned on.
26164 (rs6000_option_override_internal): Give an error if -mvsx-timode
26165 and VSX is not enabled.
26166 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
26167 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
26168 to reg+offset addressing. Use PTImode when checking offset
26169 addresses for validity.
26170 (reg_offset_addressing_ok_p): Likewise.
26171 (rs6000_legitimate_offset_address_p): Likewise.
26172 (rs6000_legitimize_address): Likewise.
26173 (rs6000_legitimize_reload_address): Likewise.
26174 (rs6000_legitimate_address_p): Likewise.
26175 (rs6000_eliminate_indexed_memrefs): Likewise.
26176 (rs6000_emit_move): Likewise.
26177 (rs6000_secondary_reload): Likewise.
26178 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
26179 reloads to fpr registers to continue to use reg+offset addressing,
26180 but 64-bit reloads to altivec registers need reg+reg addressing.
26181 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
26182 it. Treat LO_SUM like a PLUS operation.
26183 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
26184 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
26185 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
26186 registers to share a register with a smaller sized type, since VSX
26187 puts scalars in the upper 64-bits.
26188 (print_operand): Add support for PTImode.
26189 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
26190 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
26191 registers, but don't have arithmetic support.
26192 (rs6000_memory_move_cost): Add test for VSX.
26193 (rs6000_opt_masks): Add -mvsx-timode.
26195 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
26198 (VSr): Use wt constraint for TImode.
26199 (VSv): Drop TImode support.
26200 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
26201 (vsx_movti_64bit): Likewise.
26202 (vsx_movti_32bit): Likewise.
26203 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
26204 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
26205 one '?' on the appropriate output constraint. Do not allow TImode
26206 logical operations on 32-bit systems.
26207 (vsx_ior<mode>3): Likewise.
26208 (vsx_xor<mode>3): Likewise.
26209 (vsx_one_cmpl<mode>2): Likewise.
26210 (vsx_nor<mode>3): Likewise.
26211 (vsx_andc<mode>3): Likewise.
26212 (vsx_concat_<mode>): Likewise.
26213 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
26215 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
26216 OPTION_MASK_VSX_TIMODE.
26217 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
26218 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
26220 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
26221 (TI2 iterator): New iterator for TImode, PTImode.
26222 (wd mode attribute): Add values for vector types.
26223 (movti_string): Replace TI move operations with operations for TImode
26224 and PTImode. Add support for TImode being allowed in VSX registers.
26225 (mov<mode>_string, TImode/PTImode): Likewise.
26226 (movti_ppc64): Likewise.
26227 (mov<mode>_ppc64, TImode/PTImode): Likewise.
26228 (TI mode splitters): Likewise.
26230 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
26233 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
26235 PR tree-optimization/56355
26236 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
26237 Also handle integers with undefined overflow.
26239 2013-03-20 Catherine Moore <clm@codesourcery.com>
26240 Maciej W. Rozycki <macro@codesourcery.com>
26241 Tom de Vries <tom@codesourcery.com>
26242 Nathan Sidwell <nathan@codesourcery.com>
26243 Iain Sandoe <iain@codesourcery.com>
26244 Nathan Froyd <froydnj@codesourcery.com>
26245 Chao-ying Fu <fu@mips.com>
26247 * doc/extend.texi: (micromips, nomicromips, nocompression):
26248 Document new function attributes.
26249 * doc/invoke.texi (minterlink-compressed, mmicromips,
26250 m14k, m14ke, m14kec): Document new options.
26251 (minterlink-mips16): Update documentation.
26252 * doc/md.texi (ZC, ZD): Document new constraints.
26253 * configure.ac (gcc_cv_as_micromips): Check if linker
26254 supports the .set micromips directive.
26255 * configure: Regenerate.
26256 * config.in: Regenerate.
26257 * config/mips/mips-tables.opt: Regenerate.
26258 * config/mips/micromips.md: New file.
26259 * constraints.md (ZC, ZD): New constraints.
26260 * config/mips/predicates.md (movep_src_register): New predicate.
26261 (movep_src_operand): New predicate.
26262 (non_volatile_mem_operand): New predicate.
26263 * config/mips/mips.md (multimem): New type.
26264 (length): Differentiate between 17-bit and 18-bit branch offsets.
26265 (MOVEP1, MOVEP2): New mode iterator.
26266 (mov_<load>l): Use ZC constraint.
26267 (mov_<load>r): Likewise.
26268 (mov_<store>l): Likewise.
26269 (mov_<store>r): Likewise.
26270 (*branch_equality<mode>_inverted): Add microMIPS support.
26271 (*branch_equality<mode>): Likewise.
26272 (*jump_absolute): Likewise.
26273 (indirect_jump_<mode>): Likewise.
26274 (tablejump_<mode>): Likewise.
26275 (<optab>_internal): Likewise.
26276 (sibcall_internal): Likewise.
26277 (sibcall_value_internal): Likewise.
26278 (prefetch): Use constraint ZD.
26279 * config/mips/mips.opt (minterlink-compressed): New option.
26280 (minterlink-mips16): Now an alias for minterlink-compressed.
26281 (mmicromips): New option.
26282 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26283 (compare_and_swap_12): Likewise.
26284 (sync_add<mode>): Likewise.
26285 (sync_<optab>_12): Likewise.
26286 (sync_old_<optab>_12): Likewise.
26287 (sync_new_<optab>_12): Likewise.
26288 (sync_nand_12): Likewise.
26289 (sync_old_nand_12): Likewise.
26290 (sync_new_nand_12): Likewise.
26291 (sync_sub<mode>): Likewise.
26292 (sync_old_add<mode>): Likewise.
26293 (sync_old_sub<mode>): Likewise.
26294 (sync_new_add<mode>): Likewise.
26295 (sync_new_sub<mode>): Likewise.
26296 (sync_<optab><mode>): Likewise.
26297 (sync_old_<optab><mode>): Likewise.
26298 (sync_new_<optab><mode>): Likewise.
26299 (sync_nand<mode>): Likewise.
26300 (sync_old_nand<mode>): Likewise.
26301 (sync_new_nand<mode>): Likewise.
26302 (sync_lock_test_and_set<mode>): Likewise.
26303 (test_and_set_12): Likewise.
26304 (atomic_compare_and_swap<mode>): Likewise.
26305 (atomic_exchange<mode>_llsc): Likewise.
26306 (atomic_fetch_add<mode>_llsc): Likewise.
26307 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26308 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26309 (umips_save_restore_pattern_p): Likewise.
26310 (umips_load_store_pair_p): Likewise.
26311 (umips_output_load_store_pair): Likewise.
26312 (umips_movep_target_p): Likewise.
26313 (umips_12bit_offset_address_p): Likewise.
26314 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26315 (mips_base_mips16): Rename this...
26316 (mips_base_compression_flags): ...to this. Update all uses.
26317 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26318 (mips_mips16_decl_p): Delete.
26319 (mips_nomips16_decl_p): Delete.
26320 (mips_get_compress_on_flags): New function.
26321 (mips_get_compress_off_flags): New function.
26322 (mips_get_compress_mode): New function.
26323 (mips_get_compress_on_name): New function.
26324 (mips_get_compress_off_name): New function.
26325 (mips_insert_attributes): Support multiple compression types.
26326 (mips_merge_decl_attributes): Likewise.
26327 (umips_12bit_offset_address_p): New function.
26328 (mips_start_function_definition): Emit .set micromips directive.
26329 (mips_call_may_need_jalx_p): New function.
26330 (mips_function_ok_for_sibcall): Add microMIPS support.
26331 (mips_print_operand_punctuation): Support short delay slots and
26333 (umips_swm_mask, umips_swm_encoding): New.
26334 (umips_build_save_restore): New function.
26335 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26336 (was_mips16_p): Remove.
26337 (old_compression_mode): New.
26338 (mips_set_compression_mode): New function.
26339 (mips_set_current_function): Add microMIPS support.
26340 (mips_option_override): Likewise.
26341 (umips_save_restore_pattern_p): New function.
26342 (umips_output_save_restore): New function.
26343 (umips_load_store_pair_p_1): New function.
26344 (umips_load_store_pair_p): New function.
26345 (umips_output_load_store_pair_1): New function.
26346 (umips_output_load_store_pair): New function.
26347 (umips_movep_target_p) New function.
26348 (mips_prepare_pch_save): Add microMIPS support.
26349 * config/mips/mips.h (TARGET_COMPRESSION): New.
26350 (TARGET_CPU_CPP_BUILTINS): Update macro
26351 to use new compression flags and to support microMIPS.
26352 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26353 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26354 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26355 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26356 (ASM_SPEC): Support mmicromips and mno-micromips.
26357 (M16STORE_REG_P): New macro.
26358 (MIPS_CALL): Support TARGET_MICROMIPS.
26359 (MICROMIPS_J): New macro.
26360 (mips_base_mips16): Rename this...
26361 (mips_base_compression_flags): ...to this.
26362 (UMIPS_12BIT_OFFSET_P): New macro.
26363 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26364 (MULTILIB_DIRNAMES): Likewise.
26365 2013-03-20 Richard Biener <rguenther@suse.de>
26367 PR tree-optimization/56661
26368 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26369 the result does not have to be distinct.
26371 2013-03-20 Richard Biener <rguenther@suse.de>
26373 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26376 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26377 Steven Bosscher <steven@gcc.gnu.org>
26379 PR rtl-optimization/56605
26380 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26382 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26385 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26386 that require movd instead of movq.
26388 2013-03-20 Richard Biener <rguenther@suse.de>
26390 * tree-ssa-structalias.c (struct variable_info): Add pointer
26391 to the first field of an aggregate with sub-vars. Make
26392 this and the pointer to the next subfield its ID.
26393 (vi_next): New function.
26394 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26395 storedanything_id, integer_id): Increment by one.
26396 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26397 get_call_clobber_vi): Adjust.
26398 (solution_set_expand): Simplify and speedup.
26399 (solution_set_add): Inline into ...
26400 (set_union_with_increment): ... this. Adjust accordingly.
26401 (do_sd_constraint): Likewise.
26402 (do_ds_constraint): Likewise.
26403 (do_complex_constraint): Simplify.
26404 (build_pred_graph): Adjust.
26405 (solve_graph): Likewise. Simplify and speedup.
26406 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26407 get_constraint_for_component_ref, get_constraint_for_1,
26408 first_vi_for_offset, first_or_preceding_vi_for_offset,
26409 create_function_info_for, create_variable_info_for_1,
26410 create_variable_info_for, intra_create_variable_infos): Adjust.
26411 (init_base_vars): Push NULL for ID zero.
26412 (compute_points_to_sets): Adjust.
26414 2013-03-20 Richard Biener <rguenther@suse.de>
26416 * cfgloop.c (verify_loop_structure): Streamline and avoid
26417 ICEing on corrupt loop tree.
26418 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26421 2013-03-20 Richard Biener <rguenther@suse.de>
26423 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26424 check whether an SSA update is needed.
26426 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26428 * config/mips/constraints.md (T): Rename to...
26432 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26433 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26435 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26437 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26438 (*subsi3_carryin_uxtw): Likewise.
26440 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26442 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26443 (*rorsi3_insn_uxtw): Likewise.
26445 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26447 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26448 (*extrsi5_insn_uxtw): Likewise.
26450 2013-03-19 Richard Biener <rguenther@suse.de>
26452 PR tree-optimization/56273
26453 * passes.c (init_optimization_passes): Move second VRP after DOM.
26455 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26457 * config/i386/i386.md (*movti_internal): Merge from
26458 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26459 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26460 nox64 isa attributes.
26462 2013-03-18 Richard Biener <rguenther@suse.de>
26464 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26466 (merge_node_constraints): Likewise.
26467 (build_succ_graph): Likewise.
26468 (valid_graph_edge): Inline into single caller.
26469 (unify_nodes): Likewise. Use bitmap_set_bit return value
26471 (scc_visit): Fix formatting and variable use.
26472 (do_sd_constraint): Use gcc_checking_assert.
26473 (do_ds_constraint): Likewise.
26474 (do_complex_constraint): Likewise.
26475 (condense_visit): Likewise. Cleanup.
26476 (dump_pred_graph): New function.
26477 (perform_var_substitution): Dump the pred-graph before
26478 variable substitution.
26479 (find_equivalent_node): Use gcc_checking_assert.
26480 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26482 2013-03-18 Richard Biener <rguenther@suse.de>
26484 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26485 Remove cond_expr_stmt_list argument and do not gimplify the
26487 (vect_loop_versioning): Adjust.
26488 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26489 Cleanup to use less temporaries.
26490 (vect_create_data_ref_ptr): Cleanup.
26492 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26494 PR tree-optimization/56635
26495 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26496 require types_compatible_p types.
26498 2013-03-18 Nick Clifton <nickc@redhat.com>
26500 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26501 spurious backslash.
26503 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26504 Add missing line to comment describing function.
26506 2013-03-18 Richard Biener <rguenther@suse.de>
26508 PR tree-optimization/56210
26509 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26510 Handle string / character search functions.
26511 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26513 2013-03-18 Richard Biener <rguenther@suse.de>
26515 PR middle-end/56483
26516 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26517 and implement properly.
26518 * gimple.h (gimple_cond_single_var_p): Remove.
26520 2013-03-18 Richard Biener <rguenther@suse.de>
26522 * tree-data-ref.h (find_data_references_in_loop): Declare.
26523 * tree-data-ref.c (get_references_in_stmt): Use a stack
26524 vector pre-allocated in the callers.
26525 (find_data_references_in_stmt): Adjust.
26526 (graphite_find_data_references_in_stmt): Likewise.
26527 (create_rdg_vertices): Likewise.
26528 (find_data_references_in_loop): Export.
26529 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26530 Compute dependences here...
26531 (vect_analyze_data_refs): ...not here. When we encounter
26532 a non-vectorizable data reference in basic-block vectorization
26533 truncate the data reference vector. Do not bother to
26534 fixup data-dependence information for gather loads.
26535 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26536 of data references, as reported.
26538 2013-03-18 Richard Biener <rguenther@suse.de>
26540 PR tree-optimization/3713
26541 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26542 has_constants and expr.
26543 (stmt_has_constants): Properly valueize SSA names when deciding
26544 whether the stmt has constants.
26546 2013-03-18 Richard Biener <rguenther@suse.de>
26548 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26549 whole function when there is nothing to do.
26550 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26551 * tree-vectorizer.c (vectorize_loops): Update virtual and
26552 loop-closed SSA once.
26553 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26555 2013-03-18 Richard Biener <rguenther@suse.de>
26557 PR middle-end/56113
26558 * domwalk.c (bb_postorder): New global static.
26559 (cmp_bb_postorder): New function.
26560 (walk_dominator_tree): Replace scheme imposing an order for
26561 visiting dominator sons by one sorting them at the time they
26562 are pushed on the stack.
26564 2013-03-18 Richard Biener <rguenther@suse.de>
26566 PR tree-optimization/39326
26567 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26568 (struct mem_ref): Replace mem member with ao_ref typed member.
26569 (MEM_ANALYZABLE): Adjust.
26570 (memref_eq): Likewise.
26571 (mem_ref_alloc): Likewise.
26572 (gather_mem_refs_stmt): Likewise.
26573 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26574 (execute_sm_if_changed_flag_set): Adjust.
26575 (execute_sm): Likewise.
26576 (ref_always_accessed_p): Likewise.
26577 (refs_independent_p): Likewise.
26578 (can_sm_ref_p): Likewise.
26580 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26583 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26584 return 1 even for !unsignedp.
26586 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26588 * config/i386/i386.md (isa): Add x64 and nox64.
26589 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26590 (*pushtf): Enable *roF alternative for x64 isa only.
26591 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26592 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26593 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26594 (*movtf_internal): Merge from *movtf_internal_rex64 and
26595 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26596 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26597 nox64 isa attributes.
26598 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26599 nox64 isa attributes.
26600 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26602 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26604 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26606 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26609 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26610 class names. Remove trailing comma after "ALL_REGS".
26612 2013-03-16 Jan Hubicka <jh@suse.cz>
26614 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26615 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26616 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26617 of cgraph_get_create_node.
26618 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26620 2013-03-16 Jason Merrill <jason@redhat.com>
26623 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26624 with DW_AT_default_value.
26626 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26628 * BASE-VER: Set to 4.9.0.
26630 2013-03-14 Andi Kleen <ak@linux.intel.com>
26633 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26634 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26635 Document _x* TSX intrinsics.
26637 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26638 David Holsgrove <david.holsgrove@xilinx.com>
26640 * configure.ac: Add MicroBlaze TLS support detection.
26641 * configure: Regenerate.
26642 * config/microblaze/microblaze-protos.h
26643 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26644 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26645 * config/microblaze/microblaze.c (microblaze_address_type): Add
26646 ADDRESS_TLS and tls_reloc address types.
26647 (microblaze_address_info): Add tls_reloc.
26648 (TARGET_HAVE_TLS): Define.
26649 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26650 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26651 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26652 load_tls_operand, microblaze_call_tls_get_addr,
26653 microblaze_legitimize_tls_address): New functions.
26654 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26655 (get_base_reg): Use microblaze_tls_symbol_p.
26656 (microblaze_classify_address): Handle TLS.
26657 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26658 label_mentioned_p and microblaze_tls_referenced_p.
26659 (microblaze_legitimize_address): Handle TLS.
26660 (microblaze_address_insns): Handle ADDRESS_TLS.
26661 (pic_address_needs_scratch): Handle TLS.
26662 (print_operand_address): Handle TLS.
26663 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26664 (microblaze_expand_move): Handle TLS.
26665 (microblaze_legitimate_constant_p): Check
26666 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26667 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26668 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26669 (PIC_OFFSET_TABLE_REGNUM): Set.
26670 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26671 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26672 (addsi3, movsi_internal2, movdf_internal): Update constraints
26673 * config/microblaze/predicates.md (arith_plus_operand): Define
26674 (move_operand): Redefine as move_src_operand,
26675 check microblaze_tls_referenced_p.
26677 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26679 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26680 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26682 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26684 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26687 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26689 PR tree-optimization/53265
26690 * common.opt (Waggressive-loop-optimizations): New option.
26691 * tree-ssa-loop-niter.c: Include tree-pass.h.
26692 (do_warn_aggressive_loop_optimizations): New function.
26693 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26694 if number_of_latch_executions returned constant.
26695 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26696 early. If number_of_latch_executions returned constant, set
26697 nb_iterations_upper_bound back to it.
26698 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26700 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26701 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26703 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26704 (MULTILIB_OSDIRNAMES): Set.
26705 * genmultilib: If defaultosdirname doesn't start with :: , set
26706 defaultosdirname2 instead, clear it and emit two . multilib_raw
26707 entries instead of just one.
26709 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26711 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26712 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26713 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26714 (SUBTARGET_OVERRIDE_OPTIONS): New.
26716 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26719 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26720 (musermode): Convert to Var(TARGET_USERMODE).
26721 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26722 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26723 * config/sh/sh.c (sh_option_override): Use
26724 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26725 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26727 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26729 (udivsi3_i4_single, divsi3_i4_single): Use
26730 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26732 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26734 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26737 2013-03-13 Richard Biener <rguenther@suse.de>
26739 PR tree-optimization/56608
26740 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26741 calls when vectorizing basic-blocks.
26743 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26746 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26749 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26751 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26753 2013-03-11 Jan Hubicka <jh@suse.cz>
26756 * lto-streamer-out.c (output_symbol_p): Skip references from
26757 constructors of external variables.
26759 2013-03-11 Jan Hubicka <jh@suse.cz>
26761 PR middle-end/56571
26762 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26764 * emit-rtl.c (verify_rtx_sharing): Likewise.
26765 (copy_insn_1): Likewise.
26766 * rtl.c (copy_rtx): Likewise.
26768 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26771 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26772 output_operand_lossage message.
26774 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26777 * arm.c (shift_op): Validate RTL pattern on the fly.
26778 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26781 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26784 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26785 new call patterns that clobber %fr12.
26786 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26787 split and postreload patterns.
26788 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26789 registers %fr12 and %fr12R as call used.
26791 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26793 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26794 (canon_address, record_store, replace_read, check_mem_read_rtx,
26795 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26796 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26797 rest_of_handle_dse): Likewise.
26799 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26801 PR middle-end/56524
26802 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26804 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26805 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26806 (save_optabs_if_changed): Replace with...
26807 (init_tree_optimization_optabs): ...this.
26808 * optabs.c (save_optabs_if_changed): Rename to...
26809 (init_tree_optimization_optabs): ...this. Take the optimization node
26810 as argument. Do nothing if the base optabs are already correct.
26811 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26812 to recompute optabs.
26813 * function.h (function): Remove optabs field.
26814 * function.c (invoke_set_current_function_hook): Call
26815 init_tree_optimization_optabs. Use the result to initialize
26818 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26820 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26821 if GTMA_HAS_NO_INSTRUMENTATION.
26822 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26823 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26824 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26825 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26826 GTMA_HAS_NO_INSTRUMENTATION.
26828 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26830 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26833 2013-03-08 Marek Polacek <polacek@redhat.com>
26834 Jakub Jelinek <jakub@redhat.com>
26836 PR tree-optimization/56478
26837 * predict.c (is_comparison_with_loop_invariant_p): Change the
26838 type of loop_step to tree.
26839 (predict_loops): Adjust.
26840 (predict_iv_comparison): Perform the computations on double_ints.
26842 2013-03-08 Richard Biener <rguenther@suse.de>
26844 PR tree-optimization/56570
26845 * tree-cfg.c (verify_expr_location_1): Verify locations for
26847 * tree-sra.c (create_access_replacement): Strip locations
26848 from DECL_DEBUG_EXPRs.
26850 2013-03-08 Richard Biener <rguenther@suse.de>
26852 * tree-inline.c (expand_call_inline): Do not associate
26853 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26854 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26856 2013-03-08 Richard Biener <rguenther@suse.de>
26858 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26859 or block changes with -Og. Fix for location / block encoding
26860 changes and PHI arguments with locations.
26862 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26864 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26866 (struct output_info): Likewise.
26867 (register_overhead): Remove bad gcc_assert.
26868 (bitmap_find_bit): If there is only a single bitmap element, do not
26869 count a miss as a search.
26870 (print_statistics): Update for counter type changes.
26871 (dump_bitmap_statistics): Likewise. Print headers such that they
26872 are properly lined up with the printed counters.
26874 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26876 PR tree-optimization/56559
26877 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26878 check that it has only a single use.
26880 2013-03-07 Richard Biener <rguenther@suse.de>
26882 * doc/invoke.texi (fwhole-program): Discourage use in combination
26885 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26887 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26889 PR tree-optimization/56539
26890 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26891 instead of GSI_CONTINUE_LINKING as last argument to
26892 force_gimple_operand_gsi. Adjust function comment.
26894 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26897 PR middle-end/56548
26898 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26899 promoted mode, convert the result back to the original mode.
26901 2013-03-06 Richard Biener <rguenther@suse.de>
26903 PR middle-end/56294
26904 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26905 (insert_updated_phi_nodes_compare_uids): New function.
26906 (update_ssa): Sort symbols_to_rename after UID before
26907 traversing it to insert PHI nodes.
26909 2013-03-06 Richard Biener <rguenther@suse.de>
26911 PR middle-end/50494
26912 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26913 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26916 2013-02-13 Richard Biener <rguenther@suse.de>
26919 * varasm.c (output_constant_def_1): Get the decl representing
26920 the constant as argument.
26921 (output_constant_def): Wrap output_constant_def_1.
26922 (make_decl_rtl): Use output_constant_def_1 with the decl
26923 representing the constant.
26924 (build_constant_desc): Optionally re-use a decl already
26925 representing the constant.
26926 (tree_output_constant_def): Adjust.
26928 2013-03-06 Joey Ye <joey.ye@arm.com>
26931 * gcc.c (convert_white_space): New function.
26932 (main): Handles white space in function name.
26934 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26937 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26938 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26939 to SH_DIV_CALL_TABLE for TARGET_SH2.
26940 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26942 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26943 call-table options.
26945 2013-03-05 Sterling Augustine <saugustine@google.com>
26946 Cary Coutant <ccoutant@google.com>
26949 * dwarf2out.c (resolve_addr): Don't call
26950 remove_loc_list_addr_table_entries a second time for the same
26953 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26956 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26957 (avoid_complex_debug_insns): New function.
26958 (expand_debug_locations): Call it.
26960 PR rtl-optimization/56484
26961 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26962 lifetimes of hard registers on small register class machines.
26964 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26966 * config/microblaze/microblaze-protos.h: Rename
26967 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26968 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26970 (microblaze_fast_interrupt_function_p): New function.
26971 (microblaze_is_interrupt_handler): Rename to
26972 microblaze_is_interrupt_variant and add fast_interrupt check.
26973 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26974 (save_restore_insns): Likewise.
26975 (compute_frame_size): Likewise.
26976 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26977 (microblaze_globalize_label): Likewise.
26978 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26979 * config/microblaze/microblaze.md: Use wrapper
26980 microblaze_is_interrupt_variant.
26982 2013-03-05 Kai Tietz <ktietz@redhat.com>
26984 * sdbout.c (sdbout_one_type): Switch to current function's section
26985 supporting cold/hot.
26987 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26989 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26992 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26994 PR middle-end/56461
26995 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26996 if VALGRIND_GET_VBITS is defined, temporarily make object
26997 memory all defined, and restore previous valgrind addressability
26998 and definability afterwards. Free this_object at the end.
27000 PR middle-end/56461
27001 * lra.c (lra): Call lra_clear_live_ranges if live_p,
27002 right before calling lra_create_live_ranges, also call it
27003 when clearing live_p. Only call lra_clear_live_ranges
27004 at the end if live_p.
27006 PR middle-end/56461
27007 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
27009 2013-03-05 Richard Biener <rguenther@suse.de>
27011 PR tree-optimization/56521
27012 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
27015 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
27018 * except.h (remove_unreachable_eh_regions): New prototype.
27019 * except.c (remove_eh_handler_splicer): New function, split out
27020 of remove_eh_handler.
27021 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
27022 warning about running it on many EH regions one at a time.
27023 (remove_unreachable_eh_regions_worker): New function, walk the
27024 EH tree in depth-first order and remove non-marked regions.
27025 (remove_unreachable_eh_regions): New function.
27026 * tree-eh.c (mark_reachable_handlers): New function, split out
27027 from remove_unreachable_handlers.
27028 (remove_unreachable_handlers): Use mark_reachable_handlers and
27029 remove_unreachable_eh_regions.
27030 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
27031 and remove_unreachable_eh_regions.
27033 2013-03-05 Richard Biener <rguenther@suse.de>
27035 PR middle-end/56525
27036 * loop-init.c (fix_loop_structure): Remove loops in two stages,
27037 not freeing them until the end.
27039 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27041 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
27043 2013-03-05 Richard Biener <rguenther@suse.de>
27045 PR tree-optimization/56270
27046 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
27047 of loads after scheduling an SLP instance.
27049 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27051 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
27053 (check_gcc_parallelize): Run guality.exp as a separate job from
27054 vect.exp with unsorted.exp and $(dg_target_exps) separately from
27055 struct-layout-1.exp with stackalign.exp.
27057 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
27059 PR middle-end/56461
27060 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
27061 load_index sbitmap even if some bit in it isn't set.
27063 PR middle-end/56461
27064 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
27065 (discover_iteration_bound_by_body_walk): Change queues to
27066 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
27067 spelling in comment. Call safe_push on queues[bound_index] directly.
27068 Release queues[queue_index] in every iteration unconditionally.
27069 Release bounds vector.
27071 PR middle-end/56461
27072 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
27073 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
27074 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
27077 2013-03-05 Richard Biener <rguenther@suse.de>
27080 * tree-inline.c (remap_blocks_to_null): New function.
27081 (expand_call_inline): When expanding a call stmt without
27082 an associated block inline remap all callee blocks to NULL.
27084 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27086 PR rtl-optimization/56494
27087 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
27088 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
27089 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
27091 PR middle-end/56461
27092 * sel-sched-ir.c (free_sched_pools): Release
27093 succs_info_pool.stack[succs_info_pool.max_top] vectors too
27094 if succs_info_pool.max_top isn't -1.
27097 * opts.c (opts_obstack, opts_concat): Moved to...
27098 * opts-common.c (opts_obstack, opts_concat): ... here.
27100 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27102 PR middle-end/56461
27103 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
27105 2013-03-04 Martin Jambor <mjambor@suse.cz>
27107 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
27108 all appropriate places.
27110 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
27112 PR tree-optimization/56424
27113 * ipa-split.c (split_function): Do not set the RSO flag if result is
27114 not by reference and its type is a register type.
27116 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
27118 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
27119 (microblaze_legitimate_pic_operand): Likewise
27120 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
27121 new function microblaze_legitimate_pic_operand
27122 * config/microblaze/microblaze-protos.h
27123 (microblaze_legitimate_pic_operand): Declare.
27125 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27127 * config/microblaze/predicates.md (call_insn_simple_operand):
27128 New predicate for supported rtx code types.
27129 * config/microblaze/microblaze.md (call_internal1): Use
27130 call_insn_simple_operand predicate.
27132 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27134 PR middle-end/56461
27135 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
27136 partitions.ordered_remove.
27138 PR middle-end/56461
27139 * tree-vect-stmts.c (vectorizable_conversion): Don't call
27140 vec_oprnds0.create (1) for modifier == NONE.
27142 PR middle-end/56461
27143 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
27144 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
27145 vec_oprnds1 right before pushing anything to it for
27148 PR middle-end/56461
27149 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
27150 set nbbs to 0 instead of having separate code path.
27151 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
27152 instead of false as last argument if returning NULL.
27154 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
27156 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
27157 the attribute is now called "target" instead of "option".
27158 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
27159 * doc/tm.texi.in (Target Attributes): Likewise document the correct
27160 attribute/pragma name for TARGET_OPTION_VALID_P and
27161 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
27162 * doc/tm.texi: Regenerated.
27164 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
27166 * config/microblaze/microblaze.c:
27167 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
27168 * config/microblaze/microblaze.h: Add -mxl-reorder to
27170 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
27171 instructions emitted if TARGET_REORDER.
27172 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
27173 or 0 for -m/-mno case, but initialises as 2 to detect default use case
27176 2013-03-01 Xinliang David Li <davidxl@google.com>
27178 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
27181 2013-03-01 Jakub Jelinek <jakub@redhat.com>
27183 PR middle-end/56461
27184 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
27185 vector even when returning true. Fix up function comment formatting.
27187 PR middle-end/56461
27188 * ira-build.c (ira_loop_nodes_count): New variable.
27189 (create_loop_tree_nodes): Initialize it.
27190 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
27192 PR middle-end/56461
27193 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
27194 method on dr_chain and result_chain.
27195 * tree-vect-stmts.c (vectorizable_store): Only call
27196 result_chain.create if j == 0.
27198 PR middle-end/56461
27199 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
27200 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
27201 before overwriting it.
27203 2013-03-01 Tobias Burnus <burnus@net-b.de>
27205 * doc/extended.texi (C Extensions): Change order in @menu
27207 (Other MIPS Built-in Functions): Move last MIPS entry before
27208 "picoChip Built-in Functions".
27209 (SH Built-in Functions): Move after RX Built-in Functions.
27210 * doc/gcc.texi (Introduction): Change order in @menu
27212 * doc/md.texi (Constraints): Ditto.
27213 * gty.texi (Type Information): Ditto.
27214 (User-provided marking routines for template types): Make
27216 * doc/invoke.texi (AArch64 Options): Move before
27217 "Adapteva Epiphany Options".
27219 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
27220 Jakub Jelinek <jakub@redhat.com>
27223 * asan.c (gate_asan): Lookup no_sanitize_address instead of
27224 no_address_safety_analysis attribute.
27225 * doc/extend.texi (no_address_safety_attribute): Rename to
27226 no_sanitize_address attribute, mention no_address_safety_analysis
27227 attribute as deprecated alias.
27229 2013-02-28 Jakub Jelinek <jakub@redhat.com>
27231 PR middle-end/56461
27232 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
27233 type to vec<vec<tree> > *.
27234 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
27235 to be vec<tree> instead of vec<tree> *, set vec_defs
27236 to vNULL and call vec_defs.create (number_of_vects), adjust other
27238 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
27239 vectorizable_condition): Adjust vect_get_slp_defs callers.
27241 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27243 * config/aarch64/aarch64.c
27244 (aarch64_float_const_representable): Remove unused variable.
27246 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27248 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
27250 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27252 * config/aarch64/aarch64-builtins.c
27253 (aarch64_init_simd_builtins): Make static.
27255 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27257 * config/aarch64/aarch64.c
27258 (aarch64_simd_make_constant): Make static.
27260 2013-02-28 Martin Jambor <mjambor@suse.cz>
27262 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
27263 with no initialization to the RHS of debug statements.
27265 2013-02-28 Martin Jambor <mjambor@suse.cz>
27267 PR tree-optimization/56294
27268 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
27270 (get_access_replacement): Do not call create_access_replacement.
27271 Assert a replacement exists.
27272 (get_repl_default_def_ssa_name): Create the replacement declaration
27275 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27277 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
27278 final_end_function.
27280 2013-02-28 Marek Polacek <polacek@redhat.com>
27282 PR rtl-optimization/56466
27283 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
27284 if we're changing a loop.
27285 (peel_loops_completely): Likewise.
27287 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
27290 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
27292 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27295 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
27296 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
27297 INTX_FTYPE_FX, FX_FTYPE_INTX.
27298 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27300 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27302 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27303 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27304 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27305 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27306 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27307 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27308 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27309 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27310 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27311 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27312 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27313 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27314 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27315 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27316 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27317 (avrxmega6): Increase max flash segments from 5 to 6.
27318 * config/avr/t-multilib: Regenerate.
27319 * config/avr/avr-tables.opt: Regenerate.
27320 * doc/avr-mmcu.texi: Regenerate.
27322 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27324 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27325 (avr_device_to_arch): Rename to avr_device_to_ld.
27326 (avr_device_to_as): New prototype.
27327 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27328 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27329 * config/avr/driver-avr.c (avr_device_to_as): New.
27330 (avr_device_to_arch): Rename to avr_device_to_ld.
27332 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27334 PR middle-end/56461
27335 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27336 method on dr_chain and result_chain.
27338 PR middle-end/56461
27339 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27340 pointer_set_destroy on not_executed_last_iteration.
27342 PR middle-end/56461
27343 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27345 PR middle-end/56461
27346 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27347 FOR_EACH_DEFINED_FUNCTION when freeing state.
27349 PR middle-end/56461
27350 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27352 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27355 PR middle-end/56461
27356 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27357 known_aggs[i].items and release known_aggs vector.
27359 PR middle-end/56461
27360 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27362 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27364 * config/microblaze/microblaze.c (microblaze_emit_compare):
27365 Use xor for EQ/NE comparisions.
27366 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27367 (cbranchsf4): Adjust operator to comparison_operator.
27369 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27371 PR middle-end/56461
27372 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27374 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27375 vec_safe_push, always update *slot.
27376 (redirect_edge_var_map_clear): Use vec_free.
27377 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27378 (free_var_map_entry): Use vec_free.
27379 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27380 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27382 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27384 PR middle-end/45472
27385 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27386 when the may_trap_p bit of the exprs being merged differs.
27387 Reorder tests for speculativeness in the logical and operator.
27389 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27391 * incpath.c (add_standard_paths): Use reconcat instead of concat
27392 where appropriate and avoid leaking memory.
27394 * opts.h: Include obstack.h.
27395 (opts_concat): New prototype.
27396 (opts_obstack): New declaration.
27397 * opts.c (opts_concat): New function.
27398 (opts_obstack): New variable.
27399 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27400 (finish_options): Use opts_concat instead of concat
27401 and XOBNEWVEC instead of XNEWVEC.
27402 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27403 generate_option): Likewise.
27404 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27405 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27408 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27409 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27411 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27413 PR middle-end/56461
27414 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27416 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27418 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27419 (arm_block_move_unaligned_straight): Likewise.
27420 (arm_adjust_block_mem): Likewise.
27422 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27425 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27426 temp, cond and label.
27427 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27430 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27431 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27434 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27435 of prologue/epilogue functions.
27438 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27439 Remove unused variable cfa_offset.
27440 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27443 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27444 type promotion to unsigned.
27447 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27448 for HOST_WIDE_INT of 32 bit / same size as int.
27449 (arm_block_move_unaligned_straight): Likewise.
27450 (arm_adjust_block_mem): Likewise.
27453 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27456 2013-02-26 Marek Polacek <polacek@redhat.com>
27458 PR tree-optimization/56426
27459 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27461 2013-02-26 Richard Biener <rguenther@suse.de>
27464 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27465 unused variable loops.
27467 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27469 PR tree-optimization/56448
27470 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27471 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27472 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27473 later operands of the references, or even first operand for
27474 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27476 PR tree-optimization/56443
27477 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27478 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27479 to type_for_mode langhook.
27481 2013-02-25 Matt Turner <mattst88@gmail.com>
27483 * doc/invoke.texi: Document r4700.
27485 2013-02-25 Richard Biener <rguenther@suse.de>
27487 PR tree-optimization/56175
27488 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27490 (simplify_bitwise_binary): ... here. Also guard the conversion
27491 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27493 2013-02-25 Catherine Moore <clm@codesourcery.com>
27496 2013-02-24 Catherine Moore <clm@codesourcery.com>
27497 Maciej W. Rozycki <macro@codesourcery.com>
27498 Tom de Vries <tom@codesourcery.com>
27499 Nathan Sidwell <nathan@codesourcery.com>
27500 Iain Sandoe <iain@codesourcery.com>
27501 Nathan Froyd <froydnj@codesourcery.com>
27502 Chao-ying Fu <fu@mips.com>
27504 * doc/extend.texi: (micromips, nomicromips, nocompression):
27505 Document new function attributes.
27506 * doc/invoke.texi (minterlink-compressed, mmicromips,
27507 m14k, m14ke, m14kec): Document new options.
27508 (minterlink-mips16): Update documentation.
27509 * doc/md.texi (ZC, ZD): Document new constraints.
27510 * configure.ac (gcc_cv_as_micromips): Check if linker
27511 supports the .set micromips directive.
27512 * configure: Regenerate.
27513 * config.in: Regenerate.
27514 * config/mips/mips-tables.opt: Regenerate.
27515 * config/mips/micromips.md: New file.
27516 * constraints.md (ZC, AD): New constraints.
27517 * config/mips/predicates.md (movep_src_register): New predicate.
27518 (movep_src_operand): New predicate.
27519 (non_volatile_mem_operand): New predicate.
27520 * config/mips/mips.md (multimem): New type.
27521 (length): Differentiate between 17-bit and 18-bit branch offsets.
27522 (MOVEP1, MOVEP2): New mode iterator.
27523 (mov_<load>l): Use ZC constraint.
27524 (mov_<load>r): Likewise.
27525 (mov_<store>l): Likewise.
27526 (mov_<store>r): Likewise.
27527 (*branch_equality<mode>_inverted): Add microMIPS support.
27528 (*branch_equality<mode>): Likewise.
27529 (*jump_absolute): Likewise.
27530 (indirect_jump_<mode>): Likewise.
27531 (tablejump_<mode>): Likewise.
27532 (<optab>_internal): Likewise.
27533 (sibcall_internal): Likewise.
27534 (sibcall_value_internal): Likewise.
27535 (prefetch): Use constraint ZD.
27536 * config/mips/mips.opt (minterlink-compressed): New option.
27537 (minterlink-mips16): Now an alias for minterlink-compressed.
27538 (mmicromips): New option.
27539 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27540 (compare_and_swap_12): Likewise.
27541 (sync_add<mode>): Likewise.
27542 (sync_<optab>_12): Likewise.
27543 (sync_old_<optab>_12): Likewise.
27544 (sync_new_<optab>_12): Likewise.
27545 (sync_nand_12): Likewise.
27546 (sync_old_nand_12): Likewise.
27547 (sync_new_nand_12): Likewise.
27548 (sync_sub<mode>): Likewise.
27549 (sync_old_add<mode>): Likewise.
27550 (sync_old_sub<mode>): Likewise.
27551 (sync_new_add<mode>): Likewise.
27552 (sync_new_sub<mode>): Likewise.
27553 (sync_<optab><mode>): Likewise.
27554 (sync_old_<optab><mode>): Likewise.
27555 (sync_new_<optab><mode>): Likewise.
27556 (sync_nand<mode>): Likewise.
27557 (sync_old_nand<mode>): Likewise.
27558 (sync_new_nand<mode>): Likewise.
27559 (sync_lock_test_and_set<mode>): Likewise.
27560 (test_and_set_12): Likewise.
27561 (atomic_compare_and_swap<mode>): Likewise.
27562 (atomic_exchange<mode>_llsc): Likewise.
27563 (atomic_fetch_add<mode>_llsc): Likewise.
27564 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27565 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27566 (umips_save_restore_pattern_p): Likewise.
27567 (umips_load_store_pair_p): Likewise.
27568 (umips_output_load_store_pair): Likewise.
27569 (umips_movep_target_p): Likewise.
27570 (umips_12bit_offset_address_p): Likewise.
27571 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27572 (mips_base_mips16): Rename this...
27573 (mips_base_compression_flags): ...to this. Update all uses.
27574 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27575 (mips_mips16_decl_p): Delete.
27576 (mips_nomips16_decl_p): Delete.
27577 (mips_get_compress_on_flags): New function.
27578 (mips_get_compress_off_flags): New function.
27579 (mips_get_compress_mode): New function.
27580 (mips_get_compress_on_name): New function.
27581 (mips_get_compress_off_name): New function.
27582 (mips_insert_attributes): Support multiple compression types.
27583 (mips_merge_decl_attributes): Likewise.
27584 (umips_12bit_offset_address_p): New function.
27585 (mips_start_function_definition): Emit .set micromips directive.
27586 (mips_call_may_need_jalx_p): New function.
27587 (mips_function_ok_for_sibcall): Add microMIPS support.
27588 (mips_print_operand_punctuation): Support short delay slots and
27590 (umips_swm_mask, umips_swm_encoding): New.
27591 (umips_build_save_restore): New function.
27592 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27593 (was_mips16_p): Remove.
27594 (old_compression_mode): New.
27595 (mips_set_compression_mode): New function.
27596 (mips_set_current_function): Add microMIPS support.
27597 (mips_option_override): Likewise.
27598 (umips_save_restore_pattern_p): New function.
27599 (umips_output_save_restore): New function.
27600 (umips_load_store_pair_p_1): New function.
27601 (umips_load_store_pair_p): New function.
27602 (umips_output_load_store_pair_1): New function.
27603 (umips_output_load_store_pair): New function.
27604 (umips_movep_target_p) New function.
27605 (mips_prepare_pch_save): Add microMIPS support.
27606 * config/mips/mips.h (TARGET_COMPRESSION): New.
27607 (TARGET_CPU_CPP_BUILTINS): Update macro
27608 to use new compression flags and to support microMIPS.
27609 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27610 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27611 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27612 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27613 (ASM_SPEC): Support mmicromips and mno-micromips.
27614 (M16STORE_REG_P): New macro.
27615 (MIPS_CALL): Support TARGET_MICROMIPS.
27616 (MICROMIPS_J): New macro.
27617 (mips_base_mips16): Rename this...
27618 (mips_base_compression_flags): ...to this.
27619 (UMIPS_12BIT_OFFSET_P): New macro.
27620 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27621 (MULTILIB_DIRNAMES): Likewise.
27623 2013-02-25 Tom de Vries <tom@codesourcery.com>
27625 PR rtl-optimization/56131
27626 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27627 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27628 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27630 2013-02-25 Tobias Burnus <burnus@net-b.de>
27632 * doc/invoke.texi (-fsanitize=): Move from optimization
27633 to debugging options.
27635 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27637 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27639 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27640 Alexander Monakov <amonakov@ispras.ru>
27642 PR middle-end/56077
27643 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27644 flush pending lists also on non-jumps. Adjust comment.
27646 2013-02-24 Catherine Moore <clm@codesourcery.com>
27647 Maciej W. Rozycki <macro@codesourcery.com>
27648 Tom de Vries <tom@codesourcery.com>
27649 Nathan Sidwell <nathan@codesourcery.com>
27650 Iain Sandoe <iain@codesourcery.com>
27651 Nathan Froyd <froydnj@codesourcery.com>
27652 Chao-ying Fu <fu@mips.com>
27654 * doc/extend.texi: (micromips, nomicromips, nocompression):
27655 Document new function attributes.
27656 * doc/invoke.texi (minterlink-compressed, mmicromips,
27657 m14k, m14ke, m14kec): Document new options.
27658 (minterlink-mips16): Update documentation.
27659 * doc/md.texi (ZC, ZD): Document new constraints.
27660 * configure.ac (gcc_cv_as_micromips): Check if linker
27661 supports the .set micromips directive.
27662 * configure: Regenerate.
27663 * config.in: Regenerate.
27664 * config/mips/mips-tables.opt: Regenerate.
27665 * config/mips/micromips.md: New file.
27666 * constraints.md (ZC, AD): New constraints.
27667 * config/mips/predicates.md (movep_src_register): New predicate.
27668 (movep_src_operand): New predicate.
27669 (non_volatile_mem_operand): New predicate.
27670 * config/mips/mips.md (multimem): New type.
27671 (length): Differentiate between 17-bit and 18-bit branch offsets.
27672 (MOVEP1, MOVEP2): New mode iterator.
27673 (mov_<load>l): Use ZC constraint.
27674 (mov_<load>r): Likewise.
27675 (mov_<store>l): Likewise.
27676 (mov_<store>r): Likewise.
27677 (*branch_equality<mode>_inverted): Add microMIPS support.
27678 (*branch_equality<mode>): Likewise.
27679 (*jump_absolute): Likewise.
27680 (indirect_jump_<mode>): Likewise.
27681 (tablejump_<mode>): Likewise.
27682 (<optab>_internal): Likewise.
27683 (sibcall_internal): Likewise.
27684 (sibcall_value_internal): Likewise.
27685 (prefetch): Use constraint ZD.
27686 * config/mips/mips.opt (minterlink-compressed): New option.
27687 (minterlink-mips16): Now an alias for minterlink-compressed.
27688 (mmicromips): New option.
27689 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27690 (compare_and_swap_12): Likewise.
27691 (sync_add<mode>): Likewise.
27692 (sync_<optab>_12): Likewise.
27693 (sync_old_<optab>_12): Likewise.
27694 (sync_new_<optab>_12): Likewise.
27695 (sync_nand_12): Likewise.
27696 (sync_old_nand_12): Likewise.
27697 (sync_new_nand_12): Likewise.
27698 (sync_sub<mode>): Likewise.
27699 (sync_old_add<mode>): Likewise.
27700 (sync_old_sub<mode>): Likewise.
27701 (sync_new_add<mode>): Likewise.
27702 (sync_new_sub<mode>): Likewise.
27703 (sync_<optab><mode>): Likewise.
27704 (sync_old_<optab><mode>): Likewise.
27705 (sync_new_<optab><mode>): Likewise.
27706 (sync_nand<mode>): Likewise.
27707 (sync_old_nand<mode>): Likewise.
27708 (sync_new_nand<mode>): Likewise.
27709 (sync_lock_test_and_set<mode>): Likewise.
27710 (test_and_set_12): Likewise.
27711 (atomic_compare_and_swap<mode>): Likewise.
27712 (atomic_exchange<mode>_llsc): Likewise.
27713 (atomic_fetch_add<mode>_llsc): Likewise.
27714 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27715 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27716 (umips_save_restore_pattern_p): Likewise.
27717 (umips_load_store_pair_p): Likewise.
27718 (umips_output_load_store_pair): Likewise.
27719 (umips_movep_target_p): Likewise.
27720 (umips_12bit_offset_address_p): Likewise.
27721 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27722 (mips_base_mips16): Rename this...
27723 (mips_base_compression_flags): ...to this. Update all uses.
27724 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27725 (mips_mips16_decl_p): Delete.
27726 (mips_nomips16_decl_p): Delete.
27727 (mips_get_compress_on_flags): New function.
27728 (mips_get_compress_off_flags): New function.
27729 (mips_get_compress_mode): New function.
27730 (mips_get_compress_on_name): New function.
27731 (mips_get_compress_off_name): New function.
27732 (mips_insert_attributes): Support multiple compression types.
27733 (mips_merge_decl_attributes): Likewise.
27734 (umips_12bit_offset_address_p): New function.
27735 (mips_start_function_definition): Emit .set micromips directive.
27736 (mips_call_may_need_jalx_p): New function.
27737 (mips_function_ok_for_sibcall): Add microMIPS support.
27738 (mips_print_operand_punctuation): Support short delay slots and
27740 (umips_swm_mask, umips_swm_encoding): New.
27741 (umips_build_save_restore): New function.
27742 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27743 (was_mips16_p): Remove.
27744 (old_compression_mode): New.
27745 (mips_set_compression_mode): New function.
27746 (mips_set_current_function): Add microMIPS support.
27747 (mips_option_override): Likewise.
27748 (umips_save_restore_pattern_p): New function.
27749 (umips_output_save_restore): New function.
27750 (umips_load_store_pair_p_1): New function.
27751 (umips_load_store_pair_p): New function.
27752 (umips_output_load_store_pair_1): New function.
27753 (umips_output_load_store_pair): New function.
27754 (umips_movep_target_p) New function.
27755 (mips_prepare_pch_save): Add microMIPS support.
27756 * config/mips/mips.h (TARGET_COMPRESSION): New.
27757 (TARGET_CPU_CPP_BUILTINS): Update macro
27758 to use new compression flags and to support microMIPS.
27759 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27760 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27761 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27762 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27763 (ASM_SPEC): Support mmicromips and mno-micromips.
27764 (M16STORE_REG_P): New macro.
27765 (MIPS_CALL): Support TARGET_MICROMIPS.
27766 (MICROMIPS_J): New macro.
27767 (mips_base_mips16): Rename this...
27768 (mips_base_compression_flags): ...to this.
27769 (UMIPS_12BIT_OFFSET_P): New macro.
27770 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27771 (MULTILIB_DIRNAMES): Likewise.
27773 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27776 * target-globals.c (save_target_globals): For init_reg_sets and
27777 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27779 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27781 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27782 * config/aarch64/t-aarch64
27783 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27785 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27787 PR inline-asm/56148
27788 * lra-constraints.c (process_alt_operands): Reload operand
27789 conflicting with earlier clobber only if no more other conflicting
27792 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27795 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27796 if not linking a shared library.
27798 2013-02-22 Seth LaForge <sethml@google.com>
27800 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27802 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27804 * config/arm/arm.md (split for extendsidi): Update condition.
27805 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27806 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27807 (qhs_zextenddi_cstr): Likewise.
27809 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27811 PR middle-end/56420
27812 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27813 avoid signed wrapping.
27814 (expand_mult): Handle properly multiplication by
27815 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27816 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27817 in the compiler if coeff is HOST_WIDE_INT_MIN.
27818 (expand_divmod): Don't make ext_op1 static, change it's type to
27819 uhwi. Avoid undefined behavior in -INTVAL (op1).
27821 PR rtl-optimization/50339
27822 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27824 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27825 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27826 into splitting_ashiftrt field.
27827 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27829 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27832 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27834 PR middle-end/56108
27835 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27836 are sure to go irrevocable.
27838 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27840 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27841 scalars are valid operands.
27843 2013-02-21 Martin Jambor <mjambor@suse.cz>
27845 PR tree-optimization/56310
27846 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27847 only matching indices and non-negative final offsets.
27848 (intersect_aggregates_with_edge): Pass src_idx to
27849 agg_replacements_to_vector. Pass src_idx insstead of index to
27850 intersect_with_agg_replacements.
27852 2013-02-21 Martin Jambor <mjambor@suse.cz>
27854 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27855 instead of hard-wired defaults.
27857 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27859 * doc/invoke.texi (MIPS Options): Update documentation of the
27860 floating-point multiply-accumulate instruction restrictions.
27862 2013-02-21 Kostya Serebryany <kcc@google.com>
27864 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27865 asan_shadow_offset on x86_64 linux.
27867 2013-02-21 Richard Biener <rguenther@suse.de>
27869 PR tree-optimization/56415
27871 2013-02-11 Richard Biener <rguenther@suse.de>
27873 PR tree-optimization/56273
27874 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27877 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27880 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27883 PR inline-asm/56405
27884 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27885 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27887 2013-02-20 Jan Hubicka <jh@suse.cz>
27889 PR tree-optimization/56265
27890 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27891 when target is referenced for first time.
27893 2013-02-20 Richard Biener <rguenther@suse.de>
27895 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27896 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27897 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27898 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27899 not return anything.
27900 (rename_ssa_copies): Do not remove unused locals.
27901 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27902 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27903 * passes.c (execute_function_todo): Do not schedule unused locals
27904 removal if cleanup_tree_cfg did something.
27905 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27906 about the number of removed locals.
27908 2013-02-20 Richard Biener <rguenther@suse.de>
27910 PR tree-optimization/56398
27911 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27913 2013-02-20 Martin Jambor <mjambor@suse.cz>
27915 PR tree-optimization/55334
27916 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27917 restricted pointers to arrays.
27919 2013-02-20 Richard Biener <rguenther@suse.de>
27920 Jakub Jelinek <jakub@redhat.com>
27922 PR tree-optimization/56396
27923 * tree-ssa-ccp.c (n_const_val): New static variable.
27924 (get_value): Return NULL for SSA names we don't have a lattice
27926 (ccp_initialize): Initialize n_const_val.
27927 * tree-ssa-copy.c (n_copy_of): New static variable.
27928 (init_copy_prop): Initialize n_copy_of.
27929 (get_value): Return NULL_TREE for SSA names we don't have a
27932 2013-02-20 Martin Jambor <mjambor@suse.cz>
27934 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27936 2013-02-20 Richard Biener <rguenther@suse.de>
27938 * genpreds.c (write_lookup_constraint): Do not compare first
27939 letter of the constraint again.
27941 2013-02-20 Richard Biener <rguenther@suse.de>
27943 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27945 (get_use_iv_cost): Terminate hashtable walk when coming across
27948 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27950 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27951 reassociation for avx2 targets.
27953 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27955 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27956 Add version check for v8.10.a to enable microblaze_has_clz
27957 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27958 version and TARGET_PATTERN_COMPARE check
27959 * config/microblaze/microblaze.md: New clzsi2 instruction
27961 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27963 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27964 function before branching.
27966 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27968 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27970 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27972 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27974 PR middle-end/55889
27975 * sel-sched.c: Include ira.h.
27976 (implicit_clobber_conflict_p): New function.
27977 (moveup_expr): Use it.
27978 * Makefile.in (sel-sched.o): Depend on ira.h.
27980 2013-02-19 Richard Biener <rguenther@suse.de>
27982 PR tree-optimization/56384
27983 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27984 (vn_hash_type): Split out from ...
27985 (vn_hash_constant_with_type): ... here.
27986 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27987 (vn_phi_eq): Compare types from vn_phi_s structure.
27988 (vn_phi_lookup): Populate vn_phi_s type.
27989 (vn_phi_insert): Likewise.
27991 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27993 PR tree-optimization/56350
27994 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27995 if haven't found reduction or nested cycle operand, rather than
27996 asserting we must find it.
27998 PR tree-optimization/56381
27999 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
28002 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
28003 Jakub Jelinek <jakub@redhat.com>
28006 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
28007 (swap_optab_enable): Same.
28008 (init_all_optabs): Use argument instead of global.
28009 * tree.h (struct tree_optimization_option): New field target_optabs.
28010 * expr.h (init_all_optabs): Add argument to prototype.
28011 (TREE_OPTIMIZATION_OPTABS): New.
28012 (save_optabs_if_changed): Protoize.
28013 * optabs.h: Declare this_fn_optabs.
28014 * optabs.c (save_optabs_if_changed): New.
28015 Declare this_fn_optabs.
28016 (init_optabs): Add argument to init_all_optabs() call.
28017 * function.c (invoke_set_current_function_hook): Handle per
28019 * function.h (struct function): New field optabs.
28020 * config/mips/mips.c (mips_set_mips16_mode): Handle when
28021 optimization_current_node has changed.
28022 * target-globals.h (save_target_globals_default_opts): Protoize.
28023 * target-globals.c (save_target_globals_default_opts): New.
28025 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28028 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
28029 registers %fr12 and %fr12R as call used.
28032 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
28033 and HImode, require all displacements to be an integer multiple of
28035 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
28036 only allow QImode and HImode when reload is in progress and strict is
28037 true. Likewise for symbolic addresses. Use base14_operand to check
28038 displacements in REG+BASE addresses.
28040 2013-02-18 Richard Biener <rguenther@suse.de>
28042 PR tree-optimization/56366
28043 * tree-vect-loop.c (get_initial_def_for_induction): Properly
28044 handle sign-conversion of outer-loop initial induction value.
28046 2013-02-18 Richard Biener <rguenther@suse.de>
28048 PR middle-end/56349
28049 * cfghooks.c (merge_blocks): If we merge a latch into another
28050 block adjust references to it.
28051 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
28052 (verify_loop_structure): Verify that a recorded latch is in fact
28055 2013-02-18 Richard Biener <rguenther@suse.de>
28057 PR tree-optimization/56321
28058 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
28059 order SSA name release and virtual operand unlinking.
28061 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28063 * config/microblaze/microblaze.md (save_stack_block): Define.
28064 (restore_stack_block): Likewise.
28066 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28068 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
28069 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
28070 * config/microblaze/microblaze.c (microblaze_option_override):
28071 Bail out early for PIC modes when target does not support PIC.
28073 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28075 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
28076 Replace with a microblaze version.
28077 (microblaze_trampoline_init): Adapt for microblaze.
28078 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
28081 2013-02-16 Jakub Jelinek <jakub@redhat.com>
28082 Dodji Seketeli <dodji@redhat.com>
28085 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
28086 (instrument_mem_region_access): Do not forget to always put
28087 instrumentation of the of 'base' and 'base + len' in a "if (len !=
28088 0) statement, even for cases where either 'base' or 'base + len'
28089 are not instrumented -- because they have been previously
28090 instrumented. Simplify the logic by putting all the statements
28091 instrument 'base + len' inside a sequence, and then insert that
28092 sequence right before the current insertion point. Then, to
28093 instrument 'base + len', just get an iterator on that statement.
28094 And do not forget to update the pointer to iterator the function
28095 received as argument.
28097 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
28099 PR rtl-optimization/56348
28100 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
28102 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
28104 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
28105 (clean_graph_dump_file): Pass base to start_graph_dump.
28107 2013-02-14 Richard Henderson <rth@redhat.com>
28110 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
28112 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
28114 * collect2-aix.h: Define F_LOADONLY.
28116 2013-02-14 Richard Biener <rguenther@suse.de>
28119 * varasm.c (output_constant_def_1): Get the decl representing
28120 the constant as argument.
28121 (output_constant_def): Wrap output_constant_def_1.
28122 (make_decl_rtl): Use output_constant_def_1 with the decl
28123 representing the constant.
28124 (build_constant_desc): Optionally re-use a decl already
28125 representing the constant.
28126 (tree_output_constant_def): Adjust.
28128 2013-02-14 Dodji Seketeli <dodji@redhat.com>
28131 * asan.c (instrument_builtin_call): Really put the length of the
28132 second source argument into src1_len.
28134 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28136 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
28137 argument. If it is false, don't create edge from then_bb to
28139 (insert_if_then_before_iter): Pass true to it.
28140 (build_check_stmt): Pass false to it.
28141 (transform_statements): Flush hash table only on extended basic
28142 block boundaries, rather than at the beginning of every bb.
28143 Don't flush hash table on nonfreeing_call_p calls.
28144 * tree-flow.h (nonfreeing_call_p): New prototype.
28145 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
28147 2013-02-13 David S. Miller <davem@davemloft.net>
28149 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
28151 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
28154 * ira.c (max_regno_before_ira): Move from ...
28156 (fix_reg_equiv_init): Use max_regno_before_ira instead of
28159 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28161 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
28163 2013-02-13 Richard Biener <rguenther@suse.de>
28166 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
28167 globals in MEM_REFs.
28169 2013-02-13 Richard Biener <rguenther@suse.de>
28171 * loop-init.c (loop_optimizer_init): Clear loop state when
28172 re-initializing preserved loops.
28173 * loop-unswitch.c (unswitch_single_loop): Return whether
28174 we unswitched the loop. Do not verify loop state here.
28175 (unswitch_loops): When we unswitched a loop discover new loops.
28177 2013-02-13 Kostya Serebryany <kcc@google.com>
28179 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
28181 * sanitizer.def: Rename __asan_init to __asan_init_v1.
28183 2013-02-12 Dodji Seketeli <dodji@redhat.com>
28185 Avoid instrumenting duplicated memory access in the same basic block
28186 * Makefile.in (asan.o): Add new dependency on hash-table.h
28187 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
28188 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
28189 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
28190 (free_mem_ref_resources, has_mem_ref_been_instrumented)
28191 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
28192 (get_mem_ref_of_assignment): New functions.
28193 (get_mem_refs_of_builtin_call): Extract from
28194 instrument_builtin_call and tweak a little bit to make it fit with
28196 (instrument_builtin_call): Use the new
28197 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
28198 of is_gimple_builtin_call.
28199 (instrument_derefs, instrument_mem_region_access): Insert the
28200 instrumented memory reference into the hash table.
28201 (maybe_instrument_assignment): Renamed instrument_assignment into
28202 this, and change it to advance the iterator when instrumentation
28203 actually happened and return true in that case. This makes it
28204 homogeneous with maybe_instrument_assignment, and thus give a
28205 chance to callers to be more 'regular'.
28206 (transform_statements): Clear the memory reference hash table
28207 whenever we enter a new BB, when we cross a function call, or when
28208 we are done transforming statements. Use
28209 maybe_instrument_assignment instead of instrumentation. No more
28210 need to special case maybe_instrument_assignment and advance the
28211 iterator after calling it; it's now handled just like
28212 maybe_instrument_call. Update comment.
28214 2013-02-13 Richard Biener <rguenther@suse.de>
28216 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
28217 Fix loop discovery code.
28219 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
28221 PR inline-asm/56148
28222 * lra-constraints.c (process_alt_operands): Match early clobber
28223 operand with itself. Check conflicts with earlyclobber only if
28224 the operand is not reloaded. Prefer to reload conflicting operand
28225 if earlyclobber and matching operands are the same.
28227 2013-02-12 Richard Biener <rguenther@suse.de>
28230 * lto-streamer-out.c (write_symbol): Do not output symbols
28231 for hard register variables.
28233 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
28236 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
28237 (umulsidi3_insn, mulsidi3_insn): New insns.
28239 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
28241 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
28242 (struct tune_params): Add vec_costs field.
28243 * config/arm/arm.c (arm_builtin_vectorization_cost)
28244 (arm_add_stmt_cost): New functions.
28245 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
28246 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
28247 (arm_default_vec_cost): New struct of type cpu_vec_costs.
28248 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
28249 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
28250 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
28251 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
28253 2013-02-12 Richard Biener <rguenther@suse.de>
28256 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
28257 decls again if possible.
28259 2013-02-12 Richard Biener <rguenther@suse.de>
28261 PR middle-end/56288
28262 * tree-ssa.c (verify_ssa_name): Fix check, move
28263 SSA_NAME_IN_FREE_LIST check up.
28265 2013-02-12 Jakub Jelinek <jakub@redhat.com>
28266 Steven Bosscher <steven@gcc.gnu.org>
28268 PR rtl-optimization/56151
28269 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
28270 equal to op0 or op1, and last_insn pattern is CODE operation
28271 with MEM dest and one of the operands matches that MEM.
28273 2013-02-11 Sriraman Tallam <tmsriram@google.com>
28275 * doc/extend.texi: Document Function Multiversioning and "default"
28276 parameter string to target attribute.
28277 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
28278 target attribute parameter is "default".
28279 (ix86_compare_version_priority): Remove checks for target attribute.
28280 (ix86_mangle_function_version_assembler_name): Change error to sorry.
28281 Remove check for target attribute equal to NULL. Add assert.
28282 (ix86_generate_version_dispatcher_body): Change error to sorry.
28284 2013-02-11 Iain Sandoe <iain@codesourcery.com>
28285 Jack Howarth <howarth@bromo.med.uc.edu>
28286 Patrick Marlier <patrick.marlier@gmail.com>
28289 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
28290 define ENDFILE_SPEC as TM_DESTRUCTOR.
28291 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
28293 2013-02-11 Alexander Potapenko <glider@google.com>
28294 Jack Howarth <howarth@bromo.med.uc.edu>
28295 Jakub Jelinek <jakub@redhat.com>
28298 * config/darwin.c (cdtor_record): Rename ctor_record.
28299 (sort_cdtor_records): Rename sort_ctor_records.
28300 (finalize_dtors): New routine to sort destructors by
28301 priority before use in assemble_integer.
28302 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28304 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28306 PR rtl-optimization/56275
28307 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28308 offset is non-negative and less than cmode size before
28309 calling simplify_subreg.
28311 2013-02-11 Richard Biener <rguenther@suse.de>
28313 PR tree-optimization/56264
28314 * cfgloop.h (fix_loop_structure): Adjust prototype.
28315 * loop-init.c (fix_loop_structure): Return the number of
28316 newly discovered loops.
28317 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28318 are discovered, do a full loop-closed SSA rewrite.
28320 2013-02-11 Richard Biener <rguenther@suse.de>
28322 PR tree-optimization/56273
28323 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28325 (check_array_ref): Fix missing newline in dumps.
28326 (search_for_addr_array): Likewise.
28328 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28330 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28332 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28335 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28337 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28339 PR rtl-optimization/56246
28340 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28342 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28343 constraints are satisfied.
28345 2013-02-08 Jeff Law <law@redhat.com>
28348 * emit-rtl.c (reg_is_parm_p): New function.
28349 * regs.h (reg_is_parm_p): New prototype.
28350 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28351 callee-clobbered registers.
28353 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28356 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28357 If there is no implicit builtin declaration, just return NULL.
28359 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28361 * config/i386/sse.md (FMAMODEM): New mode iterator.
28362 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28363 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28365 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28367 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28368 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28369 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28371 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28373 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28374 (microblaze*-*-elf): Likewise.
28375 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28377 * config/microblaze/microblaze-c.c: Add builtin defines for
28378 _LITTLE_ENDIAN and _BIG_ENDIAN.
28379 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28380 add to TARGET_DEFAULT flags.
28381 Expand ASM_SPEC and LINK_SPEC.
28382 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28383 * config/microblaze/microblaze.md: Update extendsidi2 and
28384 movdi_internal instructions to use low-order / high-order reg
28386 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28387 options and inversemask / mask of LITTLE_ENDIAN.
28388 * config/microblaze/t-microblaze: Expand multilib options to
28389 include mlittle-endian (le) and update exceptions patterns.
28391 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28393 PR rtl-optimization/56195
28394 * lra-constraints.c (get_reload_reg): Don't reuse regs
28395 if they have smaller mode than requested, if they have
28396 wider mode than requested, try to return a SUBREG.
28398 PR tree-optimization/56250
28399 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28400 if type is unsigned and code isn't MULT_EXPR.
28402 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28404 PR tree-optimization/56064
28405 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28406 bits according to mode.
28407 * fixed-value.h (fixed_from_double_int)
28408 (const_fixed_from_double_int): Adjust comments.
28410 2013-02-08 Richard Biener <rguenther@suse.de>
28413 * lto-streamer.h (struct data_in): Remove current_file, current_line
28414 and current_col members.
28415 * lto-streamer-out.c (lto_output_location): Stream changed bits
28416 en-block for efficiency.
28417 * lto-streamer-in.c (clear_line_info): Remove.
28418 (lto_input_location): Cache current file, line and column
28419 globally via local statics. Read changed bits en-block.
28420 (input_function): Do not call clear_line_info.
28421 (lto_read_body): Likewise.
28422 (lto_input_toplevel_asms): Likewise.
28424 2013-02-08 Michael Matz <matz@suse.de>
28426 PR tree-optimization/52448
28427 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28428 (nt_call_phase): New static.
28429 (add_or_mark_expr): Only mark accesses with newer phase than any
28431 (nonfreeing_call_p): New.
28432 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28433 (nt_fini_block): Keep blocks marked as visited.
28434 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28436 2013-02-08 Richard Biener <rguenther@suse.de>
28438 * ira.c (ira): Free broken dominator information.
28440 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28442 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28444 2013-02-08 Marek Polacek <polacek@redhat.com>
28446 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28448 2013-02-08 Richard Biener <rguenther@suse.de>
28450 PR middle-end/56181
28451 * cfgloop.h (flow_loops_find): Adjust.
28452 (bb_loop_header_p): Declare.
28453 * cfgloop.c (bb_loop_header_p): New function split out from ...
28454 (flow_loops_find): ... here. Adjust function signature,
28455 support incremental loop structure update.
28456 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28457 * cfgloopmanip.c (fix_loop_structure): Move ...
28458 * loop-init.c (fix_loop_structure): ... here.
28459 (apply_loop_flags): Split out from ...
28460 (loop_optimizer_init): ... here.
28461 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28462 in incremental mode, only remove dead loops here.
28464 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28467 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28468 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28469 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28470 (*round<mode>3.libgcc): New insns for fixed-modes.
28471 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28472 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28473 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28474 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28475 implementations. Define to __builtin_avr_absFX,
28476 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28477 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28478 __builtin_avr_countlsFX, respectively.
28479 * config/avr/avr-c.c (target.h): Include it.
28480 (enum avr_builtin_id): New enum.
28481 (avr_resolve_overloaded_builtin): New static function.
28482 (avr_register_target_pragmas): Use it to set
28483 targetm.resolve_overloaded_builtin.
28484 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28485 tree nodes used by DEF_BUILTIN.
28486 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28487 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28488 <AVR_BUILTIN_xxBITS>: Same.
28490 2013-02-08 Richard Biener <rguenther@suse.de>
28492 * cfgloop.c (verify_loop_structure): Properly handle
28493 a loop exiting to another loop header.
28494 * ira-int.h (ira_loops): Remove.
28495 * ira.c (ira_loops): Remove.
28496 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28497 (do_reload): Use loop_optimizer_finalize.
28498 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28499 number_of_loops to access the loop tree.
28500 (more_one_region_p): Likewise.
28501 (finish_loop_tree_nodes): Likewise.
28502 (rebuild_regno_allocno_maps): Likewise.
28503 (mark_loops_for_removal): Likewise.
28504 (mark_all_loops_for_removal): Likewise.
28505 (remove_unnecessary_regions): Likewise.
28506 (ira_build): Likewise.
28507 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28509 2013-02-08 Richard Biener <rguenther@suse.de>
28511 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28512 * ipa-pure-const.c (analyze_function): Avoid calling
28513 mark_irreducible_loops twice.
28514 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28516 2013-02-07 David S. Miller <davem@davemloft.net>
28518 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28520 * var-tracking.c (vt_add_function_parameter): Test the presence of
28521 HAVE_window_save properly and do not remap argument registers when
28522 we have a leaf function.
28524 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28527 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28529 * config/i386/i386.c (ix86_print_operand): Ditto.
28531 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28533 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28535 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28537 PR rtl-optimization/56225
28538 * lra-constraints.c (process_alt_operands): Check that reload hard
28539 reg can hold value for strict_low_part.
28541 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28544 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28545 dwarf2out_end_function.
28546 (in_first_function_p, maybe_at_text_label_p,
28547 first_loclabel_num_not_at_text_label): New variables.
28548 (dwarf2out_var_location): In the first function find out
28549 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28550 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28553 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28555 PR rtl-optimization/56178
28556 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28557 SUBREG of a register. Tidy up related block of code.
28558 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28559 note if the source is a register or a SUBREG of a register.
28561 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28564 * config/rs6000/rs6000.md (ptrm): New mode attr.
28565 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28566 call_value_indirect_aix<pttrsize>,
28567 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28570 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28572 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28573 if -bnortl. Convert to strcmp and strncmp.
28575 2013-02-07 Alan Modra <amodra@gmail.com>
28578 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28579 addresses won't wrap when offsetting.
28580 (rs6000_secondary_reload): Provide secondary reloads needed for
28581 wrapping LO_SUM addresses.
28583 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28585 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28586 MACH, just __MACH__.
28588 2013-02-06 Richard Biener <rguenther@suse.de>
28590 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28591 instead of calling fix_loop_structure.
28593 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28595 PR middle-end/56217
28596 * omp-low.c (use_pointer_for_field): Return false if
28597 lower_send_shared_vars doesn't generate any copy-out code.
28599 2013-02-06 Tom de Vries <tom@codesourcery.com>
28601 PR rtl-optimization/56131
28602 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28603 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28604 of the label is NULL. Add comment.
28606 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28608 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28611 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28612 (STATIC_LIBTSAN_LIBS): Likewise.
28613 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28614 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28615 is defined, don't add anything else beyond that.
28616 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28617 (LINK_COMMAND_SPEC): Use them.
28619 PR tree-optimization/56205
28620 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28621 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28622 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28624 2013-02-05 Richard Biener <rguenther@suse.de>
28626 PR tree-optimization/53342
28627 PR tree-optimization/53185
28628 * tree-vectorizer.h (vect_check_strided_load): Remove.
28629 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28630 not disallow peeling for vectorized strided loads.
28631 (vect_check_strided_load): Make static and simplify.
28632 (vect_analyze_data_refs): Adjust.
28633 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28634 correctly when vectorizing strided loads.
28636 2013-02-05 Richard Biener <rguenther@suse.de>
28638 * doc/install.texi: Refer to ISL, not PPL.
28640 2013-02-05 Jan Hubicka <jh@suse.cz>
28642 PR tree-optimization/55789
28643 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28645 2013-02-05 Jan Hubicka <jh@suse.cz>
28647 PR tree-optimization/55789
28648 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28649 the dead call anyway.
28651 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28654 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28656 2013-02-04 Alexander Potapenko <glider@google.com>
28657 Jack Howarth <howarth@bromo.med.uc.edu>
28658 Jakub Jelinek <jakub@redhat.com>
28661 * config/darwin.c (sort_ctor_records): Stabilized qsort
28662 on constructor priority by using original position.
28663 (finalize_ctors): New routine to sort constructors by
28664 priority before use in assemble_integer.
28665 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28667 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28670 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28671 about visibility on artificial decls.
28672 * config/sol2.c (solaris_assemble_visibility): Likewise.
28674 2013-02-04 Kai Tietz <ktietz@redhat.com>
28677 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28678 argument and improve checking of return-argument types for 16-byte
28680 (ix86_function_value_1): Add additional valtype argument on call
28681 of function_value_64.
28682 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28683 handling infunction_value_64 function.
28685 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28687 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28689 2013-02-04 Richard Biener <rguenther@suse.de>
28691 PR tree-optimization/56188
28692 * tree-ssa-structalias.c (label_visit): Consider case with
28693 initially non-empty points-to set.
28694 (perform_var_substitution): Dump node mapping and clean up.
28696 2013-02-04 Richard Guenther <rguenther@suse.de>
28699 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28700 node prevail as last resort.
28701 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28702 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28704 2013-02-04 Richard Biener <rguenther@suse.de>
28706 PR tree-optimization/56113
28707 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28709 (equiv_class_lookup_or_add): ... this.
28710 (label_visit): Adjust and fix error in previous patch.
28711 (perform_var_substitution): Adjust.
28713 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28715 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28716 * config/sh/sh4-300.md: Likewise.
28717 * config/sh/sh4a.md: Likewise.
28718 * config/sh/constraints.md: Likewise.
28719 * config/sh/sh.md: Likewise.
28720 * config/sh/netbsd-elf.h: Likewise.
28721 * config/sh/predicates.md: Likewise.
28722 * config/sh/sh-protos.h: Likewise.
28723 * config/sh/ushmedia.h: Likewise.
28724 * config/sh/linux.h: Likewise.
28725 * config/sh/sh.c: Likewise.
28726 * config/sh/superh.h: Likewise.
28727 * config/sh/elf.h: Likewise.
28728 * config/sh/sh4.md: Likewise.
28729 * config/sh/sh.h: Likewise.
28731 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28733 * config/pa/constraints.md: Adjust unused letters. Change "T"
28734 constraint to match_test floating_point_store_memory_operand().
28735 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28736 (base14_operand): New.
28737 (floating_point_store_memory_operand): New.
28738 (integer_store_memory_operand): Revise to use base14_operand and
28739 reg_plus_base_memory_operand.
28740 (move_dest_operand): Allow symbolic_memory_operands.
28741 (symbolic_memory_operand): Check for LO_SOM.
28742 (symbolic_operand): Change default case to break.
28743 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28744 CONST_DOUBLE values to be reloaded by putting them into memory when
28745 the destination is a floating point register.
28746 (movdf): Remove code to handle CONST_DOUBLE.
28748 (reload_indf_r1): New.
28749 (reload_insf_r1): New.
28750 Consistently use "Q" and "T" constraints with integer and floating
28751 point move instructions, respectively.
28752 (movdi): Remove FAIL.
28753 Change predicate for source operand unamed DImode move from
28754 general_operand to move_src_operand.
28755 (umulsidi3): Change predicate for destination operand to
28757 Likewise for similar unamed patterns.
28758 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28759 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28760 (hppa_legitimize_address): Simplify mask calculation.
28761 (pa_emit_move_sequence): Revised handling of secondary reloads from
28762 REG+D addresses for floating point loads and stores. Directly handle
28763 loading CONST0_RTX (mode) to a floating point register.
28764 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28765 to floating point registers. Don't restrict secondary reloads to
28766 floating point registers to integer modes. Revise some comments and
28768 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28769 (pa_legitimate_address_p): New.
28770 (pa_legitimize_reload_address): New.
28771 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28772 (STRICT_REG_OK_FOR_BASE_P): New.
28773 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28774 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28776 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28777 Andrew Dixie <andrewd@gentrack.com>
28779 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28782 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28784 * expmed.c (extract_bit_field_1): Pass the full width of the
28785 structure to get_best_reg_extraction_insn.
28787 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28790 * configure.ac (use_cxa_atexit): Add AIX.
28791 * configure: Regenerate.
28793 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28795 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28798 * final.c (need_profile_function): New variable.
28799 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28800 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28801 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28802 notes, targetm.asm_out.function_prologue doesn't emit anything,
28803 HAVE_prologue and profiler should be emitted before prologue,
28804 set need_profile_function instead of emitting it.
28805 (final_scan_insn): If need_profile_function, emit
28806 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28807 NOTE_INSN_FUNCTION_BEG note.
28809 2013-02-01 Richard Henderson <rth@redhat.com>
28811 * config/rs6000/rs6000.md (smulditi3): New.
28814 * config/alpha/alpha.md (umulditi3): New.
28816 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28818 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28819 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28821 2013-02-01 Richard Biener <rguenther@suse.de>
28823 PR tree-optimization/56113
28824 * tree-ssa-structalias.c (label_visit): Reduce work for
28825 single-predecessor nodes.
28827 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28829 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28830 range isn't testing for zero.
28832 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28834 PR middle-end/56113
28835 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28837 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28838 Nick Clifton <nickc@redhat.com>
28840 * config/v850/constraints.md (Q): Define as a memory constraint.
28841 * config/v850/predicates.md (label_ref_operand): New predicate.
28842 (e3v5_shift_operand): New predicate.
28843 (ior_operator): New predicate.
28844 * config/v850/t-v850: Add e3v5 multilib.
28845 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28846 (v850_gen_movdi): Prototype.
28847 * config/v850/v850.c: Add support for e3v5 architecture.
28848 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28850 (construct_save_jarl): Add e3v5 long JARL support.
28851 (v850_adjust_insn_length): New function. Adjust length of call
28852 insns when using e3v5 instructions.
28853 (v850_gen_movdi): New function: Generate instructions to move a
28855 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28856 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28857 (TARGET_USE_FPU): Enable for e3v5.
28858 (CONST_OK_FOR_W): New macro.
28859 (ADJUST_INSN_LENGTH): Define.
28860 * config/v850/v850.md (UNSPEC_LOOP): Define.
28861 (attr cpu): Add v850e3v5.
28862 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28863 (movdi): New pattern.
28864 (movdi_internal): New pattern.
28865 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28866 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28867 (cstoresf4): Likewise.
28868 (cstoredf4): Likewise.
28869 (insv): New pattern.
28870 (rotlso3_a): New pattern.
28871 (rotlsi3_b): New pattern
28872 (rotlsi3_v850e3v5): New pattern.
28873 (doloop_begin): New pattern.
28874 (fix_loop_counter): New pattern.
28875 (doloop_end): New pattern.
28876 (branch_normal): Add e3v5 long branch support.
28877 (branch_invert): Likewise.
28878 (branch_z_normal): Likewise.
28879 (branch_z_invert): Likewise.
28880 (branch_nz_normal): Likewise.
28881 (branch_nz_invert): Likewise.
28882 (call_internal_short): Add e3v5 register-indirect JARL support.
28883 (call_internal_long): Likewise.
28884 (call_value_internal_short): Likewise.
28885 (call_value_internal_long): Likewise.
28886 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28887 (mloop): New option.
28888 * config.gcc: Add support for configuring v840e3v5 target.
28889 * doc/invoke.texi: Document new v850 specific command line options.
28891 2013-01-31 Paul Koning <ni1d@arrl.net>
28895 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28896 children if parent is a class.
28897 (prune_unused_types_prune): Don't add DW_AT_declaration.
28899 2013-01-31 Richard Biener <rguenther@suse.de>
28901 PR tree-optimization/56157
28902 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28903 match up operand with SLP child.
28905 2013-01-31 Jason Merrill <jason@redhat.com>
28908 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28909 parameters the first time.
28910 (gen_scheduled_generic_parms_dies): Check completeness here.
28912 2013-01-31 Richard Biener <rguenther@suse.de>
28914 PR middle-end/53073
28915 * common.opt (faggressive-loop-optimizations): New flag,
28916 enabled by default.
28917 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28918 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28919 infer_loop_bounds_from_undefined by it.
28921 2013-01-31 Richard Biener <rguenther@suse.de>
28923 PR tree-optimization/56150
28924 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28925 visit virtual operands.
28926 (find_uses_to_rename_bb): Likewise.
28928 2013-01-31 Richard Biener <rguenther@suse.de>
28930 PR tree-optimization/56150
28931 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28932 mixed store non-store stmts.
28934 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28937 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28938 LIBASAN_EARLY_SPEC is defined.
28939 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28940 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28942 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28945 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28946 invalid args instead of ICEing on it.
28947 (ix86_valid_target_attribute_tree): Return error_mark_node if
28948 ix86_valid_target_attribute_inner_p failed.
28949 (ix86_valid_target_attribute_p): Return false only if
28950 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28951 target("default") attribute.
28952 (sorted_attr_string): Change argument from const char * to tree,
28953 merge in all target attribute arguments rather than just one.
28954 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28955 instead of free. Avoid using strcat.
28956 (ix86_mangle_function_version_assembler_name): Mangle
28957 target("default") as if no target attribute is present. Adjust
28958 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28959 instead of xmalloc and XDELETEVEC instead of free.
28960 (ix86_function_versions): Don't return true if one of the decls
28961 doesn't have target attribute. If they don't and one of the decls
28962 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28963 sorted_attr_string caller. Use XDELETEVEC instead of free.
28964 (ix86_supports_function_versions): Remove.
28965 (make_name): Fix up formatting.
28966 (make_dispatcher_decl): Remove resolver_name and its initialization.
28967 Avoid leaking memory.
28968 (is_function_default_version): Return true if there is
28969 target("default") attribute rather than no target attribute at all.
28970 (make_resolver_func): Avoid leaking memory.
28971 (ix86_generate_version_dispatcher_body): Likewise.
28972 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28973 * target.def (supports_function_versions): Remove.
28974 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28975 * doc/tm.texi: Regenerated.
28977 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28979 PR rtl-optimization/56144
28980 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28981 for values with side effects.
28983 2013-01-30 Richard Biener <rguenther@suse.de>
28985 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28986 (sparseset_pop): Likewise.
28987 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28988 to be able to use quick_push in the worker loop.
28990 2013-01-30 Marek Polacek <polacek@redhat.com>
28992 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28994 2013-01-30 Richard Biener <rguenther@suse.de>
28997 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28999 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
29001 PR tree-optimization/56064
29002 * fixed-value.c (fixed_from_double_int): New function.
29003 * fixed-value.h (fixed_from_double_int): New prototype.
29004 (const_fixed_from_double_int): New static inline function.
29005 * fold-const.c (native_interpret_fixed): New static function.
29006 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
29007 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
29008 (native_encode_fixed): New static function.
29009 (native_encode_expr) <FIXED_CST>: Use it.
29010 (native_interpret_int): Move double_int worker code to...
29011 * double-int.c (double_int::from_buffer): ...this new static method.
29012 * double-int.h (double_int::from_buffer): Prototype it.
29014 2013-01-30 Richard Biener <rguenther@suse.de>
29016 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
29017 New pointer-map and obstack.
29018 (init_alias_vars): Allocate pointer-map and obstack.
29019 (delete_points_to_sets): Free them.
29020 (find_what_var_points_to): Cache result.
29021 (find_what_p_points_to): Adjust for changed interface of
29022 find_what_var_points_to.
29023 (compute_points_to_sets): Likewise.
29024 (ipa_pta_execute): Likewise.
29026 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29028 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
29029 * configure: Regenerate.
29030 * config.in: Regenerate.
29031 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
29032 #nobits/#progbits if supported.
29034 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
29037 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
29038 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
29039 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
29041 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29043 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
29044 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
29046 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29048 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
29049 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
29051 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29053 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
29055 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
29056 * config/arm/cortex-a7.md: New bypasses using
29057 arm_mac_accumulator_is_result.
29059 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29061 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
29062 (cortex_a7_neon_mla): Likewise.
29063 (cortex_a7_fpfmad): New reservation.
29064 (cortex_a7_fpmacs): Use ffmas and update required units.
29065 (cortex_a7_fpmuld): Update required units and latency.
29066 (cortex_a7_fpmacd): Likewise.
29067 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
29068 (cortex_a7_neon). Likewise.
29069 (bypass) Update participating units.
29071 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29073 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
29074 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
29076 * config/arm/vfp11.md (vfp_farith): Use ffmas.
29077 (vfp_fmul): Use ffmad.
29078 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
29079 (cortex_r4_fmacd): Use ffmad.
29080 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
29081 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
29082 (cortex_a9_fmacd): Use ffmad.
29083 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
29084 (cortex_a8_vfp_macd): Use ffmad.
29085 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
29086 (cortex_a5_fpmacd): Use ffmad.
29087 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
29088 (cortex_a15_vfp_macd): Use ffmad.
29089 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29091 2013-01-29 Jason Merrill <jason@redhat.com>
29094 * varasm.c (default_assemble_visibility): Don't warn about
29095 visibility on artificial decls.
29097 2013-01-29 Richard Biener <rguenther@suse.de>
29099 PR tree-optimization/56113
29100 * tree-ssa-structalias.c (equiv_class_lookup): Also return
29102 (label_visit): Free duplicate bitmaps and record the leader instead.
29103 (perform_var_substitution): Adjust.
29105 2013-01-29 Richard Biener <rguenther@suse.de>
29107 PR tree-optimization/55270
29108 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
29109 the CFG, schedule loops for fixup.
29111 2013-01-29 Nick Clifton <nickc@redhat.com>
29113 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
29116 2013-01-28 Leif Ekblad <leif@rdos.net>
29118 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
29119 * config/i386/i386.h (TARGET_RDOS): New macro.
29120 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
29121 * config/i386/i386.c (ix86_option_override_internal): For 64bit
29122 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
29123 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
29124 DEFAULT_LARGE_SECTION_THRESHOLD.
29125 * config/i386/i386.md (R14_REG, R15_REG): New constants.
29126 * config/i386/rdos.h: New file.
29127 * config/i386/rdos64.h: New file.
29129 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
29132 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
29135 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29137 PR rtl-optimization/56117
29138 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
29139 call cselib_lookup_from_insn on the MEM before calling
29140 add_insn_mem_dependence.
29142 2013-01-28 Richard Biener <rguenther@suse.de>
29144 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
29145 to a stmt that didn't have one.
29146 (copy_phis_for_bb): Likewise for PHI arguments.
29147 (copy_debug_stmt): Likewise for debug stmts.
29149 2013-01-28 Richard Biener <rguenther@suse.de>
29151 PR tree-optimization/56034
29152 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
29153 (partition_builtin_p): Adjust.
29154 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
29155 it is the last partition.
29156 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
29157 up the vertex for the definition.
29158 (classify_partition): Classify whether a partition is a
29159 PKIND_REDUCTION, thus has uses outside of the loop.
29160 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
29161 Merge all PKIND_REDUCTION partitions into the last partition.
29162 (tree_loop_distribution): Seed partitions from reductions as well.
29164 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29166 PR tree-optimization/56125
29167 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
29168 pow(x,c) into sqrt(x) * powi(x, n/2) or
29169 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
29170 optimizing for size.
29171 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
29172 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
29175 PR tree-optimization/56094
29176 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
29177 to UNKNOWN_LOCATION while gimplifying expr.
29179 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
29182 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
29183 operand 0 in movabs insn template for -masm=intel asm alternative.
29184 (*movabs<mode>_2): Ditto for operand 1.
29186 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
29189 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
29192 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29194 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
29197 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29199 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
29200 Undef to avoid warning.
29202 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29204 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
29205 * configure: Regenerate.
29207 2013-01-25 Jakub Jelinek <jakub@redhat.com>
29209 PR tree-optimization/56098
29210 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
29211 for stmts with volatile ops.
29212 (cond_store_replacement): Don't optimize if assign has volatile ops.
29213 (cond_if_else_store_replacement_1): Don't optimize if either
29214 then_assign or else_assign have volatile ops.
29215 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
29218 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29220 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
29222 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29224 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
29225 missing ':' in asm example.
29227 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
29229 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
29230 entries into lane and laneq entries.
29231 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
29232 Remove AdvSIMD scalar modes.
29233 (aarch64_sq<r>dmulh_laneq<mode>): New.
29234 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
29236 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
29237 builtin implementations to relfect changes in RTL in aarch64-simd.md.
29238 * config/aarch64/iterators.md (VCOND): New.
29241 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29244 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
29245 Add NULL LIBNAME argument to existing definitions.
29246 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
29247 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
29248 * config/avr/avr.c (DEF_BUILTIN): Same.
29249 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
29250 (avr_expand_builtin): Expand to a vanilla call if a libgcc
29251 implementation is available (DECL_ASSEMBLER_NAME is set).
29252 (avr_fold_absfx): New static function.
29253 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
29254 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
29255 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
29256 AVR_BUILTIN_ABSLLK.
29257 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
29258 (abshk, absk, abslk, absllk): Provide as static inline functions.
29260 2013-01-25 Marek Polacek <polacek@redhat.com>
29262 PR tree-optimization/56035
29263 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
29265 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
29267 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
29268 (*movtf_internal_rex64): Add (!o,C) alternative
29269 (*movxf_internal_rex64): Ditto.
29270 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
29272 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
29274 * doc/invoke.texi: fix typo.
29275 * doc/objc.texi: fix typo.
29277 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
29279 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
29280 for the first two alternatives.
29282 2013-01-24 Diego Novillo <dnovillo@google.com>
29284 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
29285 (ggc-zone.o): Remove.
29286 * configure.ac: Remove option --with-gc.
29287 * configure: Re-generate.
29288 * doc/install.texi: Remove documentation for --with-gc.
29289 * gengtype.c (write_enum_defn): Remove. Update all users.
29290 (write_Types_process_field): Remove generation of gt_e_* argument.
29291 (output_type_enum): Remove. Update all users.
29292 (write_enum_defn): Remove. Update all users.
29293 (enum alloc_zone): Remove. Update all users.
29294 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
29295 * ggc-common.c (ggc_splay_alloc): Remove first argument.
29296 Update all callers.
29297 (struct ptr_data): Remove field TYPE. Update all users.
29298 (gt_pch_note_object): Remove argument TYPE. Update all users.
29299 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29301 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29302 (struct alloc_zone): Remove.
29303 (ggc_internal_alloc_zone_stat): Remove.
29304 (ggc_internal_cleared_alloc_zone_stat): Remove.
29305 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29306 (ggc_pch_count_object): Remove last argument. Update all users.
29307 (ggc_pch_alloc_object): Remove last argument. Update all users.
29308 (struct alloc_zone): Remove.
29309 * ggc-zone.c: Remove.
29310 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29311 (struct alloc_zone): Remove.
29312 (ggc_alloc_typed_stat): Remove.
29313 (ggc_alloc_typed): Remove.
29314 (ggc_splay_alloc): Remove first argument.
29315 (rtl_zone): Remove. Update all users.
29316 (tree_zone): Remove. Update all users.
29317 (tree_id_zone): Remove. Update all users.
29318 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29319 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29320 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29321 * tree-ssanames.c: Remove references to zone allocator in comments.
29323 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29325 * config/avr/avr.c (avr_out_fract): Make register numbers that
29326 might be outside of source operand signed.
29328 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29330 * config/i386/constraints.md (Yf): New constraint.
29331 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29332 of f constraint to conditionaly disable x87 register preferences.
29333 (*movdf_internal): Ditto.
29334 (*movsf_internal): Ditto.
29336 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29338 PR inline-asm/55934
29339 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29340 that have operands with impossible constraints.
29341 Add a FIXME for a speed-up opportunity.
29342 * lra-constraints.c (process_alt_operands): Verify that a class
29343 selected from constraints on asms is valid for the operand mode.
29344 (curr_insn_transform): Remove incorrect comment.
29346 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29348 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29349 TOC operand is a valid symbol ref in the constant pool.
29351 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29353 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29355 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29358 * config/avr/stdfix.h: New file.
29359 * t-avr (stdfix-gcc.h): New rule to build it.
29360 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29362 2013-01-23 Kostya Serebryany <kcc@google.com>
29364 * config/darwin.h: remove dependency on
29365 CoreFoundation (asan on Mac OS).
29367 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29370 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29371 instead of cmpdi_operand for first comparison operand.
29372 Don't assert that comparison operands aren't both constants.
29374 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29376 * doc/install.texi (Downloading the Source): Update references to
29377 downloading separate components.
29379 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29381 * doc/extend.texi (__int128): Improve grammar.
29383 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29386 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29387 alternative to (o,r).
29388 (*movdi_internal_rex64): Remove (!o,n) alternative.
29389 (DImode immediate->memory splitter): Remove.
29390 (DImode immediate->memory peephole2): Remove.
29391 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29392 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29393 alternative to (!o,*r).
29394 (*movtf_internal_sse): New pattern.
29395 (*movxf_internal_rex64): New pattern.
29396 (*movxf_internal): Disable for TARGET_64BIT.
29397 (*movdf_internal_rex64): Remove (!o,F) alternative.
29399 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29401 PR middle-end/56074
29402 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29403 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29404 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29405 stmt locations where LOCATION_LOCUS of the stmt location is
29406 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29409 * config/i386/i386.md (UNSPEC_STOS): New.
29410 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29411 *strsetqi_1): Add UNSPEC_STOS.
29413 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29416 * doc/invoke.texi: Remove left over -Wsynth example.
29418 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29420 PR tree-optimization/56051
29421 * fold-const.c (fold_binary_loc): Don't fold
29422 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29423 a narrowing conversion, or widening conversion from signed
29426 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29428 PR rtl-optimization/56023
29429 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29430 dependent on debug instruction.
29432 2013-01-21 Martin Jambor <mjambor@suse.cz>
29434 PR middle-end/56022
29435 * function.c (allocate_struct_function): Call
29436 invoke_set_current_function_hook earlier.
29438 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29440 * reload1.c (init_reload): Only initialize reload_obstack
29441 during the first call.
29443 2013-01-21 Marek Polacek <polacek@redhat.com>
29445 * cfgloop.c (verify_loop_structure): Fix up grammar.
29447 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29449 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29450 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29452 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29455 * config/arm/marvell-pj4.md: Update copyright year.
29456 Fix up use of alu to alu_reg and simple_alu_imm.
29458 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29460 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29462 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29465 * lra-constraints.c (curr_insn_transform): Don't reuse original
29466 insn for secondary memory move when memory mode should be different.
29468 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29470 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29471 atomic_storedi_1): New patterns.
29473 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29475 btver2 pipeline descriptions.
29476 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29478 * config/i386/i386.md (btver2_decode): New type attributes.
29479 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29481 * config/i386/btver2.md: New file describing btver2 pipelines.
29483 2013-01-19 Andrew Pinski <apinski@cavium.com>
29485 PR tree-optimization/52631
29486 * tree-ssa-sccvn (visit_use): Before looking up the original
29487 statement, try looking up the simplified expression.
29489 2013-01-19 Anthony Green <green@moxielogic.com>
29491 * config/moxie/moxie.c (moxie_expand_prologue): Set
29492 current_function_static_stack_size.
29494 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29496 PR tree-optimization/56029
29497 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29498 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29500 2013-01-18 Sharad Singhai <singhai@google.com>
29502 PR tree-optimization/55995
29503 * dumpfile.c (dump_loc): Print location only if available.
29504 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29506 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29509 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29510 secondary memory move.
29511 (inherit_reload_reg): Use rclass instead of cl for
29512 check_secondary_memory_needed_p.
29514 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29516 PR middle-end/56015
29517 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29518 the case where writing real complex part of target modifies op1.
29520 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29522 * config/aarch64/aarch64-simd.md
29523 (aarch64_vcond_internal<mode>): Handle unordered cases.
29524 * config/aarch64/iterators.md (v_cmp_result): New.
29526 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29527 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29529 * config/arm/marvell-pj4.md: New file.
29530 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29531 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29532 (generic_vfp): Likewise.
29533 * config/arm/arm-cores.def: Add marvell-pj4.
29534 * config/arm/arm-tune.md: Regenerate.
29535 * config/arm/arm-tables.opt: Regenerate.
29536 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29537 * doc/invoke.texi: Document marvell-pj4.
29539 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29541 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29543 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29548 * var-tracking.c (negative_power_of_two_p): New.
29549 (global_get_addr_cache, local_get_addr_cache): New.
29550 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29551 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29553 (vt_stack_offset_p): Remove.
29554 (vt_canon_true_dep): Always canonicalize loc's address.
29555 (clobber_overlapping_mems): Make sure we have a MEM.
29556 (local_get_addr_clear_given_value): New.
29557 (val_reset): Clear local cached entries.
29558 (compute_bb_dataflow): Create and release the local cache.
29559 Disable duplicate MEMs clobbering.
29560 (emit_notes_in_bb): Clobber MEMs likewise.
29561 (vt_emit_notes): Create and release the local cache.
29562 (vt_initialize, vt_finalize): Create and release the global
29563 cache, respectively.
29564 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29566 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29568 PR libmudflap/53359
29569 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29570 not found in the symtab.
29572 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29575 PR rtl-optimization/55547
29576 PR rtl-optimization/53827
29579 * alias.c (offset_overlap_p): New, factored out of...
29580 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29581 the conservative special case for symbolic constants. Don't
29582 adjust zero sizes on alignment.
29584 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29586 PR rtl-optimization/52573
29587 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29588 REG_UNUSED for the same register.
29590 2013-01-17 Richard Biener <rguenther@suse.de>
29591 Marek Polacek <polacek@redhat.com>
29593 PR rtl-optimization/55833
29594 * loop-unswitch.c (unswitch_loops): Move loop verification...
29595 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29596 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29597 Set it to true when we're removing a loop from hierarchy tree in
29598 an irreducible region.
29599 (fix_bb_placements): Adjust caller.
29600 (fix_loop_placements): Likewise.
29602 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29604 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29605 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29606 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29607 Remove ID. Adjust comments.
29608 * config/avr/avr-c.c (avr_builtin_name): Remove.
29609 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29610 * config/avr/avr.c (avr_tolower): New static function.
29611 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29612 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29613 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29616 2013-01-17 Jan Hubicka <jh@suse.cz>
29618 PR tree-optimization/55273
29619 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29621 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29624 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29625 store through atomic_store<mode>_1.
29626 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29628 2013-01-17 Martin Jambor <mjambor@suse.cz>
29630 PR tree-optimizations/55264
29631 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29632 for virtual methods.
29633 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29634 virtual methods before inlining is over.
29635 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29637 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29640 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29642 PR rtl-optimization/56005
29643 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29644 pending reads for prefetch.
29646 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29648 * config/aarch64/aarch64.md
29649 (*cstoresi_neg_uxtw): New pattern.
29650 (*cmovsi_insn_uxtw): New pattern.
29651 (*<optab>si3_uxtw): New pattern.
29652 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29653 (*<optab>si3_insn_uxtw): New pattern.
29654 (*bswapsi2_uxtw): New pattern.
29656 2013-01-16 Richard Biener <rguenther@suse.de>
29658 * tree-inline.c (tree_function_versioning): Remove set but
29659 never used variable.
29661 2013-01-16 Richard Biener <rguenther@suse.de>
29663 PR tree-optimization/55964
29664 * tree-flow.h (rename_variables_in_loop): Remove.
29665 (rename_variables_in_bb): Likewise.
29666 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29667 (copy_loop_before): Adjust and delete update-ssa status.
29668 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29669 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29670 (rename_variables_in_loop): Remove.
29671 (slpeel_update_phis_for_duplicate_loop): Likewise.
29672 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29673 use available cfg machinery instead of duplicating it.
29674 Update PHI nodes and perform poor-mans SSA update here.
29675 (slpeel_tree_peel_loop_to_edge): Adjust.
29677 2013-01-16 Richard Biener <rguenther@suse.de>
29679 PR tree-optimization/54767
29680 PR tree-optimization/53465
29681 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29682 (vrp_visit_phi_node): For PHI arguments coming via backedges
29683 drop all symbolical range information.
29684 (execute_vrp): Compute backedges.
29686 2013-01-16 Richard Biener <rguenther@suse.de>
29688 * doc/install.texi: Update CLooG and ISL requirements to
29691 2013-01-16 Christian Bruel <christian.bruel@st.com>
29694 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29695 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29696 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29698 2013-01-16 DJ Delorie <dj@redhat.com>
29700 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29701 (UNSPECV_SP_SWITCH_E): New.
29702 (sp_switch_1): Change to an unspec.
29703 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29706 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29708 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29709 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29710 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29711 (expand_mem_thread_fence): Ditto.
29712 (expand_mem_signal_fence): Ditto.
29713 (expand_atomic_load): Ditto.
29714 (expand_atomic_store): Ditto.
29716 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29718 PR rtl-optimization/55547
29719 PR rtl-optimization/53827
29722 * alias.c (memrefs_conflict_p): Set sizes to negative after
29725 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29728 * function.c (thread_prologue_and_epilogue_insns): Always
29729 add crtl->drap_reg to set_up_by_prologue.set, even if
29730 stack_realign_drap is false.
29732 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29734 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29735 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29736 *call): Fix indention.
29738 2013-01-15 Tom de Vries <tom@codesourcery.com>
29741 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29744 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29746 PR rtl-optimization/55153
29747 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29749 2013-01-15 Martin Jambor <mjambor@suse.cz>
29751 PR tree-optimization/55920
29752 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29753 accesses as grp_to_be_debug_replaced.
29755 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29757 PR tree-optimization/55920
29758 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29759 there is non-useless type conversion needed from debug rhs to lhs,
29760 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29762 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29763 Mikael Pettersson <mikpe@it.uu.se>
29766 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29768 (ASM_OUTPUT_CASE_LABEL): Remove.
29769 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29770 * final.c (shorten_branches): Update alignment of labels before
29771 jump tables if CASE_VECTOR_SHORTEN_MODE.
29773 2013-01-15 Richard Biener <rguenther@suse.de>
29776 * system.h: Do not include gmp.h for building host tools.
29778 2013-01-15 Richard Biener <rguenther@suse.de>
29780 PR middle-end/55882
29781 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29782 account for bitpos when computing alignment.
29784 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29786 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29787 (ix86_target_macros_internal): Likewise.
29789 * config/i386/i386.c (m_CORE2I7): Removed.
29790 (m_CORE_HASWELL): New macro.
29791 (m_CORE_ALL): Likewise.
29792 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29793 (initial_ix86_arch_features): Likewise.
29794 (processor_target_table): Initializations for Core avx2.
29795 (cpu_names): New names "core-avx2".
29796 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29797 PROCESSOR_CORE_HASWELL.
29798 (ix86_issue_rate): New case.
29799 (ia32_multipass_dfa_lookahead): Likewise.
29800 (ix86_sched_init_global): Likewise.
29802 * config/i386/i386.h (TARGET_HASWELL): New macro.
29803 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29804 (processor_type): New PROCESSOR_HASWELL.
29806 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29808 PR tree-optimization/55955
29809 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29810 *SHIFT_EXPR and *ROTATE_EXPR codes.
29812 PR tree-optimization/48766
29813 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29814 -ftrapv disable -fwrapv.
29816 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29819 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29820 etc. to 1 and not to __flash.
29821 Use LL suffix for __INT24_MAX__ with -mint8.
29822 Use ULL suffix for __UINT24_MAX__ with -mint8.
29824 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29826 * config/avr/avr-arch.h
29827 (struct base_arch_s): Use typedef avr_arch_t instead.
29828 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29829 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29830 * config/avr/avr.c: Same.
29831 * config/avr/avr-devices.c: Same.
29832 * config/avr/driver-avr.c: Same.
29833 * config/avr/gen-avr-mmcu-texi.c: Same.
29834 * config/avr/avr-mcus.def: Adjust comment.
29836 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29838 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29839 * config/aarch64/iterators.md (VALLDI): New.
29841 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29842 Andi Kleen <ak@linux.intel.com>
29845 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29846 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29849 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29851 * config/avr/avr-stdint.h: Remove trailing blanks.
29852 * config/avr/avr-log.h: Same.
29853 * config/avr/avr-arch.h: Same.
29854 * config/avr/avr-devices.c: Same.
29855 * config/avr/avr-dimode.md: Same.
29856 * config/avr/predicates.md: Same.
29857 * config/avr/avr-c.c: Same. And fix typo.
29859 * config/avr/avr-protos.h: Same. And:
29860 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29861 (init_cumulative_args): Rename to avr_init_cumulative_args.
29862 (expand_prologue): Rename to avr_expand_prologue.
29863 (expand_epilogue): Rename to avr_expand_epilogue.
29864 (adjust_insn_length): Rename to avr_adjust_insn_length.
29865 (notice_update_cc): Rename to avr_notice_update_cc.
29866 (final_prescan_insn): Rename to avr_final_prescan_insn.
29867 * config/avr/avr.c: Same.
29868 * config/avr/avr.h: Same.
29869 * config/avr/avr.md: Remove trailing blanks.
29870 (prologue): Use avr_expand_prologue.
29871 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29873 2013-01-14 Richard Biener <rguenther@suse.de>
29875 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29876 verify_location, collect_subblocks): New functions.
29877 (verify_gimple_in_cfg): Verify that locations only reference
29878 BLOCKs in the functions BLOCK tree.
29880 2013-01-14 Richard Biener <rguenther@suse.de>
29882 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29884 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29886 (insert_out_of_ssa_copy_on_edge): Likewise.
29887 (rewrite_close_phi_out_of_ssa): Likewise.
29888 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29890 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29891 propagated constants.
29892 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29895 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29897 * config/avr/avr-modes.def: Add GPL copyright notice.
29899 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29901 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29902 MEMMODEL_MASK to determine memory model.
29903 (atomic_store<mode>): Ditto from operands[2].
29904 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29906 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29909 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29910 (fold_gimple_assign): Don't call unshare_expr here.
29911 (fold_ctor_reference): Call unshare_expr.
29913 2013-01-13 Terry Guo <terry.guo@arm.com>
29915 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29916 * doc/fragments.texi: Document MULTILIB_REUSE.
29917 * gcc.c (multilib_reuse): New internal spec.
29918 (set_multilib_dir): Also search multilib from multilib_reuse.
29919 * genmultilib (tmpmultilib3): Refactor code.
29920 (tmpmultilib4): Ditto.
29921 (multilib_reuse): New multilib argument.
29923 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29925 * Makefile.in: Update copyright.
29927 2013-01-12 Tom de Vries <tom@codesourcery.com>
29929 PR middle-end/55890
29930 * calls.c (expand_call): Check if arg_nr is valid.
29932 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29934 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29935 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29936 documentation. Add missing '__' in front of
29937 __builtin_ia32_packssdw256.
29939 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29942 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29943 NO_REGS for larl operands.
29944 (s390_reload_larl_operand): Use s390_load_address instead of
29947 2013-01-11 Richard Biener <rguenther@suse.de>
29949 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29950 (verify_node_sharing): ... here.
29951 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29953 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29955 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29956 Substitute TREECHECKING.
29957 * configure: Regenerate.
29958 * Makefile.in (TREECHECKING): New.
29960 2013-01-11 Richard Guenther <rguenther@suse.de>
29962 PR tree-optimization/44061
29963 * tree-vrp.c (extract_range_basic): Compute zero as
29964 value-range for __builtin_constant_p of function parameters.
29966 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29968 Update copyright years.
29970 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29972 PR rtl-optimization/55672
29973 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29974 const to be eliminable.
29976 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29978 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29979 * configure: Regenerate.
29981 2013-01-10 Richard Biener <rguenther@suse.de>
29983 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29985 2013-01-10 Richard Biener <rguenther@suse.de>
29988 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29989 locations for virtual PHI arguments.
29990 (rewrite_update_phi_arguments): Likewise.
29992 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29994 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29997 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29999 PR tree-optimization/55921
30000 * tree-complex.c (expand_complex_asm): New function.
30001 (expand_complex_operations_1): Call it for GIMPLE_ASM.
30003 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30006 * config/s390/s390.c (s390_symref_operand_p)
30007 (s390_loadrelative_operand_p): Merge the two functions.
30008 (s390_check_qrst_address, print_operand_address): Add parameters
30009 to s390_loadrelative_operand_p invokation.
30010 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
30011 (s390_reload_larl_operand, s390_secondary_reload): Use
30012 s390_loadrelative_operand_p instead of s390_symref_operand_p.
30013 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
30015 2013-01-09 Mike Stump <mikestump@comcast.net>
30017 * dse.c (record_store): Remove unnecessary assert.
30019 2013-01-09 Jan Hubicka <jh@suse.cz>
30021 PR tree-optimization/55569
30022 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
30023 * cfgloop.h (scale_loop_profile): Likewise.
30025 2013-01-09 Jan Hubicka <jh@suse.cz>
30028 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
30030 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
30032 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
30034 PR middle-end/55114
30035 * expr.h (maybe_emit_group_store): Declare.
30036 * expr.c (maybe_emit_group_store): New function.
30037 * builtins.c (expand_builtin_int_roundingfn): Call it.
30038 (expand_builtin_int_roundingfn_2): Likewise.
30040 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
30042 PR rtl-optimization/55829
30043 * lra-constraints.c (match_reload): Add code for absent output.
30044 (curr_insn_transform): Add code for reloads of matched inputs
30047 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
30049 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
30050 attribute of movddup insn to DF.
30051 (*vec_interleave_lowv2df): Ditto.
30052 (vec_dupv2df): Ditto.
30054 2013-01-09 Jan Hubicka <jh@suse.cz>
30056 PR tree-optimiation/55875
30057 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
30058 EVERY_ITERATION parameter.
30059 (number_of_iterations_exit): Check if exit is executed every iteration.
30060 (idx_infer_loop_bounds): Similarly here.
30061 (n_of_executions_at_most): Simplify
30062 to only test for cases where statement is dominated by the
30063 particular bound; handle correctly the "postdominance" test.
30064 (scev_probably_wraps_p): Use max loop iterations info
30065 as a global bound first.
30067 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
30068 Nick Clifton <nickc@redhat.com>
30070 * config/v850/v850.md (cbranchsf4): New pattern.
30071 (cstoresf4): New pattern.
30072 (cbranchdf4): New pattern.
30073 (cstoredf4): New pattern.
30074 (movsicc): Disallow floating point comparisons.
30075 (cmpsf_le_insn): Fix order of operators.
30076 (cmpsf_lt_insn): Likewise.
30077 (cmpsf_eq_insn): Likewise.
30078 (cmpdf_le_insn): Likewise.
30079 (cmpdf_lt_insn): Likewise.
30080 (cmpdf_eq_insn): Likewise.
30081 (cmpsf_ge_insn): Use LE comparison.
30082 (cmpdf_ge_insn): Likewise.
30083 (cmpsf_gt_insn): Use LT comparison.
30084 (cmpdf_gt_insn): Likewise.
30085 (cmpsf_ne_insn): Delete pattern.
30086 (cmpdf_ne_insn): Delete pattern.
30087 * config/v850/v850.c (v850_gen_float_compare): Use
30088 gen_cmpdf_eq_insn for NE comparison.
30089 (v850_float_z_comparison_operator)
30090 (v850_float_nz_comparison_operator): Move from here ...
30091 * config/v850/predicates.md: ... to here. Move GT and GE
30092 comparisons into v850_float_z_comparison_operator.
30093 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
30095 (v850_float_nz_comparison_operator): Likewise.
30097 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30099 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
30100 with calls to gen_insvsi/gen_insvdi.
30102 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30104 * config/i386/i386.c (initial_ix86_tune_features): Set up
30105 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
30107 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
30108 Jakub Jelinek <jakub@redhat.com>
30110 PR tree-optimization/48189
30111 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
30112 If nitercst is 0, don't predict the exit edge.
30114 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30116 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
30117 in asm_fprintf with reg_names.
30118 (aarch64_print_operand_address): Likewise.
30119 (aarch64_return_addr): Likewise.
30120 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
30122 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30124 * config/pa/pa.h (VAL_U6_BITS_P): Define.
30125 (INT_U6_BITS): Likewise.
30126 * config/pa/predicates.md (uint6_operand): New predicate.
30127 (shift5_operand, shift6_operand): Likewise.
30128 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
30130 (lshrdi3): Use shift6_operand.
30131 (shrpsi4, shrpdi4): New insn patterns.
30132 (extzv): Delete expander.
30133 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
30134 predicates in unamed zero extract patterns. Tighten common constraint.
30135 (extv): Delete expander.
30136 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
30137 predicates in unamed sign extract patterns. Tighten common constraint.
30138 (insv): Delete expander.
30139 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
30140 predicates in unamed insert patterns. Tighten common constraint.
30141 Change uint32_operand predicate to uint6_operand predicate in unamed
30142 DImode pattern to insert constant values of type 1...1xxxx.
30144 2013-01-04 Jan Hubicka <jh@suse.cz>
30146 PR tree-optimization/55823
30147 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
30150 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30151 Uros Bizjak <ubizjak@gmail.com>
30153 PR rtl-optimization/55845
30154 * df-problems.c (can_move_insns_across): Stop scanning at
30155 volatile_insn_p source instruction or give up if
30156 across_from .. across_to range contains any volatile_insn_p
30159 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30161 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
30162 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
30164 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
30165 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
30167 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30170 * asan.c (asan_clear_shadow): New function.
30171 (asan_emit_stack_protection): Use it.
30173 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30175 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
30176 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
30177 with tab instead of space.
30179 2013-01-08 Nick Clifton <nickc@redhat.com>
30181 * config/rl78/rl78.c (rl78_expand_prologue): Always select
30182 register bank 0 at the start of an interrupt handler.
30183 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
30186 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30188 * config/aarch64/aarch64-simd.md
30189 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
30190 (aarch64_simd_bsl): Likewise.
30191 (aarch64_vcond_internal<mode>): Likewise.
30192 (vcond<mode><mode>): Likewise.
30193 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
30194 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
30196 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30198 * config/aarch64/aarch64-builtins.c
30199 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
30201 2013-01-08 Martin Jambor <mjambor@suse.cz>
30204 * tree-sra.c (analyze_access_subtree): Return true also after
30205 potentially creating a debug-only replacement.
30207 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30209 PR middle-end/55890
30210 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
30212 PR tree-optimization/54120
30213 * tree-vrp.c (range_fits_type_p): Don't allow
30214 src_precision < precision from signed vr to unsigned_p
30215 if vr->min or vr->max is negative.
30216 (simplify_float_conversion_using_ranges): Test can_float_p
30217 against CODE_FOR_nothing.
30219 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30220 Richard Biener <rguenther@suse.de>
30222 PR middle-end/55851
30223 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
30224 types instead of just INTEGER_TYPE types.
30226 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
30228 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
30231 2013-01-07 Steve Ellcey <sellcey@mips.com>
30234 * config/mips/mips.opt: Change mad to mmad to match documentation.
30236 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30239 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
30240 .progmemx.data now.
30242 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30245 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
30246 (avr_addrspace_t): Add .section_name field.
30247 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
30249 (avr_addrspace): Same. Initialize .section_name. Remove last
30250 NULL entry. Put __memx into .progmemx.data.
30251 (progmem_section_prefix): Remove.
30252 (avr_asm_init_sections): No need to initialize progmem_section.
30253 (avr_asm_named_section): Use avr_addrspace[].section_name to get
30254 section name prefix.
30255 (avr_asm_select_section): Ditto. And use get_unnamed_section to
30256 retrieve the progmem section.
30257 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
30258 boundary to run over avr_addrspace[].
30259 (avr_register_target_pragmas): Ditto.
30261 2013-01-06 Jakub Jelinek <jakub@redhat.com>
30263 * varasm.c (output_constant_def_contents): For asan_protect_global
30264 protected strings, adjust DECL_ALIGN if needed, before testing for
30266 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
30267 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
30269 (output_object_block): For asan protected decls, emit asan padding
30270 after their contents.
30271 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
30272 (asan_finish_file): Test it here instead.
30274 2013-01-07 Nick Clifton <nickc@redhat.com>
30275 Matthias Klose <doko@debian.org>
30276 Doug Kwan <dougkwan@google.com>
30277 H.J. Lu <hongjiu.lu@intel.com>
30280 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
30282 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
30284 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
30286 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
30288 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
30290 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30293 * doc/install.texi (Cross-Compiler-Specific Options): Document
30296 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
30298 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
30299 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30300 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30301 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30302 vqmovun_high_s64): Fix source operand number and update copyright.
30304 2013-01-07 Richard Biener <rguenther@suse.de>
30306 PR middle-end/55890
30307 * gimple.h (gimple_call_builtin_p): New overload.
30308 * gimple.c (validate_call): New function.
30309 (gimple_call_builtin_p): Likewise.
30310 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30311 Use gimple_call_builtin_p.
30312 (find_func_clobbers): Likewise.
30313 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30314 (strlen_optimize_stmt): Likewise.
30316 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30318 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30319 (vld1q_dup_*): Likewise.
30320 (vld1_*): Likewise.
30321 (vld1q_*): Likewise.
30322 (vld1_lane_*): Likewise.
30323 (vld1q_lane_*): Likewise.
30325 2013-01-07 Richard Biener <rguenther@suse.de>
30327 * lto-streamer.h (LTO_minor_version): Bump to 2.
30329 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30331 * config/aarch64/aarch64-protos.h
30332 (aarch64_const_double_zero_rtx_p): Rename to...
30333 (aarch64_float_const_zero_rtx_p): ...this.
30334 (aarch64_float_const_representable_p): New.
30335 (aarch64_output_simd_mov_immediate): Likewise.
30336 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30337 move immediate case.
30338 * config/aarch64/aarch64.c
30339 (aarch64_const_double_zero_rtx_p): Rename to...
30340 (aarch64_float_const_zero_rtx_p): ...this.
30341 (aarch64_print_operand): Allow printing of new constants.
30342 (aarch64_valid_floating_const): New.
30343 (aarch64_legitimate_constant_p): Check for valid floating-point
30345 (aarch64_simd_valid_immediate): Likewise.
30346 (aarch64_vect_float_const_representable_p): New.
30347 (aarch64_float_const_representable_p): Likewise.
30348 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30349 (aarch64_output_simd_mov_immediate): New.
30350 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30351 (*movdf_aarch64): Likewise.
30352 * config/aarch64/constraints.md (Ufc): New.
30353 (Y): call aarch64_float_const_zero_rtx.
30354 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30356 2013-01-07 Richard Biener <rguenther@suse.de>
30358 PR tree-optimization/55888
30359 PR tree-optimization/55862
30360 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30361 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30362 not if it is contained therein.
30364 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30366 * config/avr/t-avr: Typo.
30368 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30371 * config/avr/t-avr: Don't automatically rebuild
30372 $(srcdir)/config/avr/t-multilib
30373 $(srcdir)/config/avr/avr-tables.opt
30374 $(srcdir)/doc/avr-mmcu.texi
30375 (avr-mcus): New phony target to build them on request.
30376 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30377 * avr/avr-mcus.def: Adjust comments.
30379 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30381 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30383 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30385 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30387 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30389 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30391 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30393 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30394 to generate profiling.
30395 * config/rs6000/aix64.h (LIB_SPEC): Same.
30397 2013-01-04 Andrew Pinski <apinski@cavium.com>
30399 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30401 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30403 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30405 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30407 (ix86_expand_move): Ditto.
30408 (ix86_zero_extend_to_Pmode): Ditto.
30409 (ix86_expand_call): Ditto.
30410 (ix86_expand_special_args_builtin): Ditto.
30411 (ix86_expand_builtin): Ditto.
30413 2013-01-04 Richard Biener <rguenther@suse.de>
30415 PR tree-optimization/55862
30416 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30417 translating them through PHI nodes.
30419 2013-01-04 Martin Jambor <mjambor@suse.cz>
30421 PR tree-optimization/55755
30422 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30423 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30425 2013-01-04 Marek Polacek <polacek@redhat.com>
30427 PR middle-end/55859
30428 * opts.c (default_options_optimization): Clarify error message.
30430 2013-01-04 Richard Biener <rguenther@suse.de>
30432 PR middle-end/55863
30433 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30436 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30439 * config/pa/pa.md (movsi): Revert previous change.
30440 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30443 2013-01-03 Richard Henderson <rth@redhat.com>
30445 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30446 after eliminating TLS symbols.
30448 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30451 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30452 * graphite-poly.c (debug_gmp_value): Likewise.
30454 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30457 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30458 selected code model, define __code_mode_small__, __code_model_medium__,
30459 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30460 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30461 xchg temporary register with %k. Declare temporary register as
30463 [__x86_64__]: For medium and large code models, preserve %rbx register.
30465 2013-01-03 Richard Biener <rguenther@suse.de>
30467 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30468 (dump_subscript): Adjust.
30469 (finalize_ddr_dependent): Do not dump redundant info.
30470 (analyze_siv_subscript): Adjust.
30471 (subscript_dependence_tester): Likewise.
30472 (compute_affine_dependence): Likewise.
30474 2013-01-03 Richard Biener <rguenther@suse.de>
30477 2013-01-03 Richard Biener <rguenther@suse.de>
30479 PR tree-optimization/55857
30480 * tree-vect-stmts.c (vectorizable_load): Do not setup
30481 re-alignment for invariant loads.
30483 2013-01-02 Richard Biener <rguenther@suse.de>
30485 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30486 invariant load do not generate a vector load from the scalar location.
30488 2013-01-03 Richard Biener <rguenther@suse.de>
30490 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30491 for not vectorizing.
30492 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30493 not build INDIRECT_REFs, call get_name once only.
30494 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30495 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30497 2013-01-03 Richard Biener <rguenther@suse.de>
30499 PR tree-optimization/55857
30500 * tree-vect-stmts.c (vectorizable_load): Do not setup
30501 re-alignment for invariant loads.
30503 2013-01-03 Richard Biener <rguenther@suse.de>
30506 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30507 prefer a built-in decl.
30509 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30511 * gcc.c (process_command): Update copyright notice dates.
30512 * gcov.c (print_version): Likewise.
30513 * gcov-dump.c (print_version): Likewise.
30515 PR rtl-optimization/55838
30516 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30517 iv0.step, iv1.step and step.
30519 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30520 Marc Glisse <marc.glisse@inria.fr>
30522 PR tree-optimization/55832
30523 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30524 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30525 integer_{one,zero}_node.
30527 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30530 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30531 * var-tracking.c (reverse_op): Don't add reverse ops to
30532 VALUEs that have already
30533 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30535 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30537 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30539 2013-01-02 Teresa Johnson <tejohnson@google.com>
30541 * dumpfile.c (dump_loc): Print filename with location.
30542 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30543 new location_t parameter to emit complete unroll message with
30544 new dump framework.
30545 (canonicalize_loop_induction_variables): Compute loops location
30546 and pass to try_unroll_loop_completely.
30547 * loop-unroll.c (report_unroll_peel): New function.
30548 (peel_loops_completely): Use new dump format with location
30549 for main dumpfile message, and invoke report_unroll_peel on success.
30550 (decide_unrolling_and_peeling): Ditto.
30551 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30552 by report_unroll_peel.
30553 (decide_peel_completely): Ditto.
30554 (decide_unroll_constant_iterations): Ditto.
30555 (decide_unroll_runtime_iterations): Ditto.
30556 (decide_peel_simple): Ditto.
30557 (decide_unroll_stupid): Ditto.
30558 * cfgloop.c (get_loop_location): New function.
30559 * cfgloop.h (get_loop_location): Declare.
30561 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30563 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30566 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30568 PR middle-end/55198
30569 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30570 BLKmode objects when EXPAND_MEMORY is specified.
30572 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30574 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30576 (fold_builtin_cpu): Do not share cpu model decls across statements.
30578 2013-01-02 Jason Merrill <jason@redhat.com>
30581 * tree.c (build_array_type_1): Revert earlier change.
30583 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30585 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30587 * config/aarch64/aarch64-tune.md: Re-generate.
30589 2013-01-02 Richard Biener <rguenther@suse.de>
30591 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30592 invariant load do not generate a vector load from the scalar location.
30594 2013-01-02 Richard Biener <rguenther@suse.de>
30597 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30598 * configure: Regenerate.
30600 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30602 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30603 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30604 (expand_builtin_int_roundingfn_2): Keep the original target around
30605 for the fallback case.
30607 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30609 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30610 to be clear for sign changes.
30612 2013-01-01 Jan Hubicka <jh@suse.cz>
30614 * ipa-inline-analysis.c: Fix formatting.
30616 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30618 PR tree-optimization/55831
30619 * tree-vect-loop.c (get_initial_def_for_induction): Use
30620 gsi_after_labels instead of gsi_start_bb.
30622 Copyright (C) 2013 Free Software Foundation, Inc.
30624 Copying and distribution of this file, with or without modification,
30625 are permitted in any medium without royalty provided the copyright
30626 notice and this notice are preserved.