1 2013-09-17 Richard Biener <rguenther@suse.de>
3 PR tree-optimization/58432
4 * tree-loop-distribution.c (tree_loop_distribution): Also
5 scan PHIs for outside loop uses and seed a partition from them.
7 2013-09-17 Bin Cheng <bin.cheng@arm.com>
9 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
10 (restructure_reference): Call backtrace_base_for_ref.
12 2013-09-17 Alan Modra <amodra@gmail.com>
15 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11 patch.
17 2013-09-16 DJ Delorie <dj@redhat.com>
19 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
20 vregs location for RL78/G10.
21 (rl78_expand_prologue): Avoid SEL on G10.
22 (rl78_expand_epilogue): Likewise.
23 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
24 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
25 __RL78_G10__ when appropriate.
26 (ASM_SPEC): Pass -mg10 along to the assembler.
27 * config/rl78/rl78.md (sel_rb): Disable for G10.
28 * config/rl78/rl78.opt: Add -mg10 option.
29 * config/rl78/t-rl78: Add -mg10 multilib.
31 2013-09-16 Xinliang David Li <davidxl@google.com>
33 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
34 * omp-low.c (omp_max_vf): Ditto.
35 (expand_omp_simd): Ditto.
36 * tree-vectorizer.c (vectorize_loops): Ditto.
37 (gate_vect_slp): Ditto.
38 (gate_increase_alignment): Ditto.
39 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
40 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
41 (gate_tree_vectorize): Name change.
42 (tree_vectorize): Ditto.
43 (pass_vectorize::gate): Call new function.
44 (pass_vectorize::execute): Ditto.
45 opts.c: O3 default setting change.
46 (finish_options): Check new flag.
47 * doc/invoke.texi: Document new flags.
48 * common.opt: New flags.
50 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
52 * doc/tm.texi.in (Cond Exec Macros): Remove node.
53 (Condition Code): Don't reference it.
54 * doc/tm.texi: Regenerate.
56 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
59 * lra-constraints.c (undo_optional_reloads): Consider all optional
60 reload even if it did not get a hard reg.
62 2013-09-16 Teresa Johnson <tejohnson@google.com>
64 * dumpfile.c (dump_loc): Remove newline emission.
65 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
66 emission to dump_printf_loc calls where missing.
67 (vect_mark_for_runtime_alias_test): Ditto.
68 (vect_analyze_data_ref_dependence): Ditto.
69 (vect_analyze_data_ref_dependences): Ditto.
70 (vect_slp_analyze_data_ref_dependence): Ditto.
71 (vect_slp_analyze_data_ref_dependences): Ditto.
72 (vect_compute_data_ref_alignment): Ditto.
73 (vect_update_misalignment_for_peel): Ditto.
74 (vect_verify_datarefs_alignment): Ditto.
75 (vector_alignment_reachable_p): Ditto.
76 (vect_get_data_access_cost): Ditto.
77 (vect_enhance_data_refs_alignment): Ditto.
78 (vect_find_same_alignment_drs): Ditto.
79 (vect_analyze_data_refs_alignment): Ditto.
80 (vect_analyze_group_access): Ditto.
81 (vect_analyze_data_ref_access): Ditto.
82 (vect_analyze_data_ref_accesses): Ditto.
83 (vect_prune_runtime_alias_test_list): Ditto.
84 (vect_analyze_data_refs): Ditto.
85 (vect_create_addr_base_for_vector_ref): Ditto.
86 (vect_create_data_ref_ptr): Ditto.
87 (vect_grouped_store_supported): Ditto.
88 (vect_grouped_load_supported): Ditto.
89 * value-prof.c (check_counter): Ditto.
90 (check_ic_target): Ditto.
91 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
92 (vect_recog_widen_mult_pattern): Ditto.
93 (vect_recog_widen_sum_pattern): Ditto.
94 (vect_recog_over_widening_pattern): Ditto.
95 (vect_recog_widen_shift_pattern): Ditto.
96 (vect_recog_rotate_pattern): Ditto.
97 (vect_recog_vector_vector_shift_pattern): Ditto.
98 (vect_recog_divmod_pattern): Ditto.
99 (vect_recog_mixed_size_cond_pattern): Ditto.
100 (vect_recog_bool_pattern): Ditto.
101 (vect_pattern_recog_1): Ditto.
102 (vect_pattern_recog): Ditto.
103 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
104 (vect_is_simple_iv_evolution): Ditto.
105 (vect_analyze_scalar_cycles_1): Ditto.
106 (vect_get_loop_niters): Ditto.
107 (vect_analyze_loop_1): Ditto.
108 (vect_analyze_loop_form): Ditto.
109 (vect_analyze_loop_operations): Ditto.
110 (vect_analyze_loop_2): Ditto.
111 (vect_analyze_loop): Ditto.
112 (report_vect_op): Ditto.
113 (vect_is_slp_reduction): Ditto.
114 (vect_is_simple_reduction_1): Ditto.
115 (vect_get_known_peeling_cost): Ditto.
116 (vect_estimate_min_profitable_iters): Ditto.
117 (vect_model_reduction_cost): Ditto.
118 (vect_model_induction_cost): Ditto.
119 (get_initial_def_for_induction): Ditto.
120 (vect_create_epilog_for_reduction): Ditto.
121 (vectorizable_reduction): Ditto.
122 (vectorizable_induction): Ditto.
123 (vectorizable_live_operation): Ditto.
124 (vect_loop_kill_debug_uses): Ditto.
125 (vect_transform_loop): Ditto.
126 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
127 (vect_stmt_relevant_p): Ditto.
128 (process_use): Ditto.
129 (vect_mark_stmts_to_be_vectorized): Ditto.
130 (vect_model_simple_cost): Ditto.
131 (vect_model_promotion_demotion_cost): Ditto.
132 (vect_model_store_cost): Ditto.
133 (vect_get_store_cost): Ditto.
134 (vect_model_load_cost): Ditto.
135 (vect_get_load_cost): Ditto.
136 (vect_init_vector_1): Ditto.
137 (vect_get_vec_def_for_operand): Ditto.
138 (vect_finish_stmt_generation): Ditto.
139 (vectorizable_call): Ditto.
140 (vectorizable_conversion): Ditto.
141 (vectorizable_assignment): Ditto.
142 (vectorizable_shift): Ditto.
143 (vectorizable_operation): Ditto.
144 (vectorizable_store): Ditto.
145 (vectorizable_load): Ditto.
146 (vectorizable_condition): Ditto.
147 (vect_analyze_stmt): Ditto.
148 (vect_transform_stmt): Ditto.
149 (vect_is_simple_use): Ditto.
150 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
151 (vect_can_advance_ivs_p): Ditto.
152 (vect_update_ivs_after_vectorizer): Ditto.
153 (vect_do_peeling_for_loop_bound): Ditto.
154 (vect_gen_niters_for_prolog_loop): Ditto.
155 (vect_update_inits_of_drs): Ditto.
156 (vect_create_cond_for_alias_checks): Ditto.
157 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
158 (vect_build_slp_tree_1): Ditto.
159 (vect_supported_load_permutation_p): Ditto.
160 (vect_analyze_slp_instance): Ditto.
161 (vect_analyze_slp): Ditto.
162 (vect_make_slp_decision): Ditto.
163 (vect_detect_hybrid_slp): Ditto.
164 (vect_bb_vectorization_profitable_p): Ditto.
165 (vect_slp_analyze_bb_1): Ditto.
166 (vect_update_slp_costs_according_to_vf): Ditto.
167 (vect_get_mask_element): Ditto.
168 (vect_transform_slp_perm_load): Ditto.
169 (vect_schedule_slp_instance): Ditto.
170 (vect_schedule_slp): Ditto.
171 (vect_slp_transform_bb): Ditto.
172 * profile.c (read_profile_edge_counts): Ditto.
173 (compute_branch_probabilities): Ditto.
174 * coverage.c (get_coverage_counts): Ditto.
176 2013-09-16 Diego Novillo <dnovillo@google.com>
178 * tree-core.h: Add missing comment lines from refactoring
181 2013-09-16 Jan Hubicka <jh@suse.cz>
183 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
184 abstract functions; for static functions check the presence
187 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
189 * config/aarch64/aarch64-simd-builtins.def (fma): New.
190 * config/aarch64/aarch64-simd.md
191 (aarch64_mla_elt<mode>): New.
192 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
193 (aarch64_mls_elt<mode>): Likewise.
194 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
195 (aarch64_fma4_elt<mode>): Likewise.
196 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
197 (aarch64_fma4_elt_to_128v2df): Likewise.
198 (aarch64_fma4_elt_to_64df): Likewise.
199 (fnma<mode>4): Likewise.
200 (aarch64_fnma4_elt<mode>): Likewise.
201 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
202 (aarch64_fnma4_elt_to_128v2df): Likewise.
203 (aarch64_fnma4_elt_to_64df): Likewise.
204 * config/aarch64/iterators.md (VDQSF): New.
205 * config/aarch64/arm_neon.h
206 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
207 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
209 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
211 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
212 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
213 (aarch64_mul3_elt_to_128df): Likewise.
214 (aarch64_mul3_elt_to_64v2df): Likewise.
215 * config/aarch64/iterators.md (VEL): Also handle DFmode.
217 (VMUL_CHANGE_NLANES) Likewise.
220 * config/aarch64/arm_neon.h
221 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
223 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
225 * config/aarch64/arm_neon.h
226 (vcvtx_high_f32_f64): Fix parameters.
228 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
229 Uros Bizjak <ubizjak@gmail.com>
231 * config/alpha.c: Include tree-ssanames.h.
233 2013-09-16 Richard Biener <rguenther@suse.de>
235 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
236 (dot_rdg_1): Handle control_dd.
237 (create_edge_for_control_dependence): New function.
238 (create_rdg_edges): Add control dependences if asked for.
239 (build_rdg): Likewise.
240 (generate_loops_for_partition): If there are not necessary
241 control stmts remove all their dependencies.
242 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
243 (distribute_loop): Pass on control dependences.
244 (tree_loop_distribution): Compute control dependences and remove
245 restriction on number of loop nodes.
247 2013-09-16 Jakub Jelinek <jakub@redhat.com>
249 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
252 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
254 * cse.c (try_const_anchors): Punt on CC modes.
256 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
258 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
260 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
263 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
266 2013-09-14 Iain Sandoe <iain@codesourcery.com>
269 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
271 (darwin_objc1_section): Likewise.
272 (darwin_file_end): Emit Image Info section when required.
274 2013-09-14 Jan Hubicka <jh@suse.cz>
276 * tree-into-ssa.c (gate_into_ssa): New.
277 (pass_data_build_ssa): Use it.
278 * cgraph.h (expand_thunk): Update prototype.
279 * cgraphunit.c (analyze_function): Expand thunks early.
280 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
281 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
282 set lowered flag; do not add new function.
283 (assemble_thunks_and_aliases): Update.
284 * tree-ssa.c (gate_init_datastructures): New gate.
285 (pass_data_init_datastructures): Use it.
287 2013-09-14 Iain Sandoe <iain@codesourcery.com>
290 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
291 xmm register set described in the psABI.
293 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
295 * dwarf2out.c (should_emit_struct_debug): Add check
296 for type_decl variable is not NULL.
298 2013-09-13 Jacek Caban <jacek@codeweavers.com>
300 * config.gcc: Use new winnt-c.c target hooks
301 * config/t-winnt: New file
302 * config/winnt-c.c: New file
303 * doc/tm.texi.in: Document new hook
304 * doc/tm.texi: Regenerated
306 2013-09-13 Jan Hubicka <jh@suse.cz>
309 * ipa-inline.c (check_callers): New function.
310 (check_caller_edge): Remove.
311 (want_inline_function_to_all_callers_p): Also permit alises that are
313 (inline_to_all_callers): Terminate the walk when devirtualization
316 2013-09-13 Jan Hubicka <jh@suse.cz>
318 * ipa-inline-analysis.c (struct growth_data): Add node.
319 (do_estimate_growth_1): Fix detection of recursion.
321 2013-09-13 Jakub Jelinek <jakub@redhat.com>
323 PR tree-optimization/58392
324 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
325 to avoid shadowing of outer loop variable. If
326 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
327 replace_by_duplicate_decl simduid of loops that have it set and
328 set dest_cfun->has_simduid_loops and/or
329 dest_cfun->has_force_vect_loops.
330 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
331 instead of maybe_lookup_decl.
332 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
333 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
334 call. Copy over force_vect and copy and remap simduid. Set
335 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
336 (copy_cfg_body): Remove blocks_to_copy argument. Use
337 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
338 caller. Don't set cfun->has_simduid_loops and/or
339 cfun->has_force_vect_loops here.
340 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
342 (expand_call_inline, tree_function_versioning): Adjust copy_body
345 2013-09-13 Martin Jambor <mjambor@suse.cz>
348 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
349 the assert if the edge was a speculative one.
351 2013-09-13 Richard Biener <rguenther@suse.de>
353 * tree-data-ref.h (known_dependences_p): Move here ...
354 * tree-loop-distribution.c (known_dependences_p): ... from here.
355 (dump_rdg_component, debug_rdg_component): Remove.
357 (generate_loops_for_partition): Use gimple_uid instead of
358 relying on matching stmt visit order.
359 (rdg_build_partitions): Take starting stmt vector.
360 (ldist_gen): Merge into ...
361 (distribute_loop): ... this function. Do not compute starting
363 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
365 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
367 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
368 Set type attribute correctly. Set predicable_short_it attribute.
369 (cmpsi_shiftsi): Remove %? from output template.
371 2013-09-13 Richard Biener <rguenther@suse.de>
373 * tree-loop-distribution.c (struct rdg_component,
374 rdg_defs_used_in_other_loops_p, free_rdg_components,
375 rdg_build_components): Remove.
376 (stmts_from_loop): Do not record virtual PHIs.
377 (generate_loops_for_partition): Skip virtual PHIs.
378 (build_rdg_partition_for_component): Rename to ...
379 (build_rdg_partition_for_vertex): ... this and adjust.
380 (rdg_build_partitions): Take a vector of starting vertices
381 instead of components. Remove unnecessary leftover handling.
382 (ldist_gen): Do not build components or record other stores.
383 (distribute_loop): Do not distribute loops containing stmts
386 2013-09-13 Christian Bruel <christian.bruel@st.com>
389 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
391 2013-09-13 Kai Tietz <ktietz@redhat.com>
393 * config.gcc: Separate cases for mingw and cygwin targets,
394 and add 64-bit cygwin target case.
396 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
397 dll-export inline-functions.
398 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
401 2013-09-13 Jeff Law <law@redhat.com>
405 2013-09-06 Jeff Law <law@redhat.com>
407 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
408 edge implied equivalences into successor phis.
410 2013-09-12 DJ Delorie <dj@redhat.com>
412 * config/rl78/rl78-virt.md: Change from | to \; for asm line
415 2013-09-12 Brooks Moses <bmoses@google.com>
418 * Makefile.in: Do not install driver binaries in $(target)/bin.
420 2013-09-12 DJ Delorie <dj@redhat.com>
422 * config/rl78/rl78.opt (mrelax): New.
423 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
424 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
426 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
428 (rl78_expand_epilogue): Likewise.
430 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
433 * lra-eliminations.c (remove_reg_equal_offset_note): New.
434 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
435 pointer elimination with using remove_reg_equal_offset_note.
437 2013-09-12 DJ Delorie <dj@redhat.com>
439 * config/msp430/: New port.
440 * config.gcc (msp430): Added.
441 * doc/invoke.texi: Document MSP430 options.
442 * doc/install.texi: Document msp430-elf
443 * doc/md.texi: Document msp430-elf
444 * doc/contrib.texi: Document msp430-elf
446 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
449 2013-09-12 Martin Jambor <mjambor@suse.cz>
452 * ipa-prop.c (remove_described_reference): Give up if the edge in the
453 reference descriptor is NULL.
454 (ipa_edge_removal_hook): If owning a reference descriptor, set its
457 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
459 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
460 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
461 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
462 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
463 make_temp_ssa_name): move to tree-ssanames.h
464 * tree-ssa-alias.h: Move prototype.
465 * tree-ssa.h: Include tree-ssanames.h.
466 * tree-ssanames.c (FREE_SSANAMES): Move to here.
467 * tree-ssanames.h: New. Move items from tree-flow*.h
468 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
470 2013-09-12 Richard Biener <rguenther@suse.de>
472 PR tree-optimization/58404
473 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
474 propagate non-invariant addresses into dereferences wrapped
475 in component references.
477 2013-09-12 Richard Biener <rguenther@suse.de>
479 PR tree-optimization/58402
480 * passes.def: Move pass_late_warn_uninitialized later.
482 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
484 * tree-ssa.h: New. Move content from tree-flow.h and
486 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
487 Move prototypes belonging to tree-ssa.c.
488 * tree-flow-inline.h (redirect_edge_var_map_def,
489 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
491 * gimple.h: Adjust prototypes.
492 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
494 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
495 * tree.h: Move prototype to tree-ssa.h.
496 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
497 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
498 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
499 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
500 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
501 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
502 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
503 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
504 graphite-blocking.c, graphite-clast-to-gimple.c,
505 graphite-dependences.c, graphite-interchange.c,
506 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
507 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
508 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
509 ipa-reference.c, ipa-split.c, ipa-utils.c,
510 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
511 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
512 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
513 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
514 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
515 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
516 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
517 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
518 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
519 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
520 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
521 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
522 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
523 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
524 value-prof.c, var-tracking.c,
525 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
527 2013-09-12 Richard Biener <rguenther@suse.de>
529 PR tree-optimization/58396
530 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
531 (build_rdg): Take a loop-nest parameter, fix memleaks.
532 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
534 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
536 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
539 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
541 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
542 The FMA instruction names should have a 'v' prefix.
544 2013-09-12 Richard Biener <rguenther@suse.de>
546 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
547 (dot_rdg): Use popen instead of system in optional code.
548 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
549 (already_processed_vertex_p): Adjust.
550 (has_anti_or_output_dependence, predecessor_has_mem_write,
551 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
552 rdg_flag_uses): Remove.
553 (rdg_flag_vertex): Simplify.
554 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
556 (build_rdg_partition_for_component): Process the first vertex
558 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
560 2013-09-12 Alan Modra <amodra@gmail.com>
562 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
564 2013-09-11 DJ Delorie <dj@redhat.com>
565 Nick Clifton <nickc@redhat.com>
567 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
568 (rl78_stack_based_mem): New.
569 * config/rl78/constraints.md (Iv08): New.
578 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
580 (movsi): Change from expand to insn-and-split.
581 (ashrsi3): Clobber AX.
585 * config/rl78/rl78.md (CC_REG): Fix.
586 (addsi3): Allow memory and immediate operands.
587 (addsi3_internal): Split into...
588 (addsi3_internal_virt): ...new, and ...
589 (addsi3_internal_real): ...new.
591 (subsi3_internal_virt): New.
592 (subsi3_internal_real): New.
593 (mulsi3): Add memory operand.
594 (mulsi3_rl78): Likewise.
595 (mulsi3_g13): Likewise.
596 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
597 (cbranchqi4_real): Add more constraint options.
598 (cbranchhi4_real): Expand pattern.
599 (cbranchhi4_real_signed): New.
600 (cbranchhi4_real_inverted): New.
601 (cbranchsi4_real_lt): New.
602 (cbranchsi4_real_ge): New.
603 (cbranchsi4_real_signed): New.
604 (cbranchsi4_real): New.
606 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
608 (lshrsi3_virt): Likewise.
609 (ashlsi3_virt): Likewise.
610 (cbranchqi4_virt_signed): New.
611 (cbranchhi4_virt_signed): New.
612 (cbranchsi4_virt): New.
613 * config/rl78/rl78.c: Whitespace fixes throughout.
614 (move_elim_pass): New.
615 (pass_data_rl78_move_elim): New.
616 (pass_rl78_move_elim): New.
617 (make_pass_rl78_move_elim): New.
618 (rl78_devirt_info): Run devirt earlier.
619 (rl78_move_elim_info): New.
620 (rl78_asm_file_start): Register it.
621 (rl78_split_movsi): New.
622 (rl78_as_legitimate_address): Allow virtual base registers when
624 (rl78_addr_space_convert): Remove spurious debug stuff.
625 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
626 (rl78_print_operand): More cases for not printing '#'.
627 (rl78_expand_compare): Remove most of the logic.
628 (content_memory): New.
629 (clear_content_memory): New.
630 (get_content_index): New.
631 (get_content_name): New.
632 (display_content_memory): New.
633 (update_content): New.
634 (record_content): New.
635 (already_contains): New.
636 (insn_ok_now): Re-recog insns with virtual registers.
637 (add_postponed_content_update): New.
638 (process_postponed_content_update): New.
639 (gen_and_emit_move): New.
640 (transcode_memory_rtx): Record new location content.
641 Use gen_and_emit_move.
642 (force_into_acc): New.
643 (move_to_acc): Use gen_and_emit_move.
644 (move_from_acc): Likewise.
645 (move_acc_to_reg): Likewise.
646 (move_to_x): Likewise.
647 (move_to_hl): Likewise.
648 (move_to_de): Likewise.
649 (rl78_alloc_physical_registers_op1): Record location content.
650 (has_constraint): New.
651 (rl78_alloc_physical_registers_op2): Record location content.
653 (rl78_alloc_physical_registers_ro1): Likewise.
654 (rl78_alloc_physical_registers_cmp): Likewise.
655 (rl78_alloc_physical_registers_umul): Likewise.
656 (rl78_alloc_address_registers_macax): New.
657 (rl78_alloc_physical_registers): Initialize and set location
658 content memory as needed.
659 (rl78_reorg): Make sure split2 is called.
660 (rl78_rtx_costs): New.
662 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
664 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
665 for (not (neg ...)) and (neg (not ...)) cases.
667 2013-09-11 Richard Biener <rguenther@suse.de>
670 * passes.def: Split critical edges before late uninit warning passes.
671 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
673 2013-09-11 Jakub Jelinek <jakub@redhat.com>
675 PR tree-optimization/58385
676 * fold-const.c (build_range_check): If both low and high are NULL,
677 use omit_one_operand_loc to preserve exp side-effects.
679 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
681 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
683 2013-09-11 Richard Biener <rguenther@suse.de>
685 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
686 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
687 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
688 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
689 stmts_from_loop, known_dependences_p, build_empty_rdg,
690 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
691 * tree-loop-distribution.c: ... here.
692 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
693 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
694 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
695 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
696 * tree-loop-distribution.c: ... here.
697 * tree-loop-distribution.c: Include gimple-pretty-print.h.
698 (struct partition_s): Add loops member.
699 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
700 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
701 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
703 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
704 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
705 Sergey Lega <sergey.s.lega@intel.com>
706 Anna Tikhonova <anna.tikhonova@intel.com>
707 Ilya Tocar <ilya.tocar@intel.com>
708 Andrey Turetskiy <andrey.turetskiy@intel.com>
709 Ilya Verbin <ilya.verbin@intel.com>
710 Kirill Yukhin <kirill.yukhin@intel.com>
711 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
713 * config/i386/constraints.md (k): New.
715 * config/i386/i386.c (const regclass_map): Add new mask registers.
716 (dbx_register_map): Ditto.
717 (dbx64_register_map): Ditto.
718 (svr4_dbx_register_map): Ditto.
719 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
721 (ix86_preferred_reload_class): Disable constants for mask registers.
722 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
723 (ix86_hard_regno_mode_ok): Support new mask registers.
724 (x86_order_regs_for_local_alloc): Ditto.
725 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
726 (FIXED_REGISTERS): Add new mask registers.
727 (CALL_USED_REGISTERS): Ditto.
728 (REG_ALLOC_ORDER): Ditto.
729 (VALID_MASK_REG_MODE): New.
730 (FIRST_MASK_REG): Ditto.
731 (LAST_MASK_REG): Ditto.
732 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
733 (MAYBE_MASK_CLASS_P): New.
734 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
735 (REG_CLASS_CONTENTS): Ditto.
737 (ANY_MASK_REG_P): Ditto.
738 (HI_REGISTER_NAMES): Add new mask registers.
739 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
740 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
742 (attribute "type"): Add mskmov, msklog.
743 (attribute "length_immediate"): Support them.
744 (attribute "memory"): Ditto.
745 (attribute "prefix_0f"): Ditto.
746 (*movhi_internal): Support new mask registers.
747 (*movqi_internal): Ditto.
748 (define_split): Split out clobber pattern is a logic
749 insn on mask registers.
750 (*k<logic><mode>): New.
751 (*andhi_1): Extend to support mask regs.
752 (*andqi_1): Extend to support mask regs.
754 (define_split): Split and-not to and and not if operands
756 (*<code><mode>_1): Separate HI mode to new pattern...
758 (*<code>qi_1): Extend to support mask regs.
763 (*one_cmpl<mode>2_1): Remove HImode and handle it...
764 (*one_cmplhi2_1): ...Here, now with mask registers support.
765 (*one_cmplqi2_1): Support new mask registers.
766 (HI/QImode arithmetics splitter): Don't split if mask registers
768 (HI/QImode not splitter): Ditto.
769 * config/i386/predicated.md (mask_reg_operand): New.
770 (general_reg_operand): Ditto.
772 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
774 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
775 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
777 2013-09-10 Jeff Law <law@redhat.com>
779 PR tree-optimization/58380
780 * tree-ssa-threadupdate.c (thread_block): Recognize another case
781 of threading through a buried loop header.
783 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
784 return value for single successor case.
786 2013-09-10 Jan Hubicka <jh@suse.cz>
788 * ipa-devirt.c (ipa_devirt): Enable with LTO.
790 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
793 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
794 support conditional execution.
795 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
797 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
799 * lra.c (lra): Clear lra_optional_reload_pseudos before every
801 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
802 Check destination too to check move insn.
803 (undo_optional_reloads): Add check that the original peudo did not
804 changed its allocation and the optional reload was inherited on last
805 inheritance pass. Break loop after deciding to keep optional reload.
806 (lra_undo_inheritance): Add check that inherited pseudo still in
809 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
811 * config/aarch64/aarch64.md (generic_sched): New.
812 * config/aarch64/aarch64-generic.md (load): Make conditional
813 on generic_sched attribute.
816 2013-09-10 Jan Hubicka <jh@suse.cz>
818 * lto-cgraph.c: Include ipa-utils.h.
819 (compute_ltrans_boundary): Also add possible targets into the boundary.
821 2013-09-10 Jan Hubicka <jh@suse.cz>
823 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
824 VAR_DECL of vtable rather than full expression.
826 2013-09-10 Jan Hubicka <jh@suse.cz>
827 Paolo Carlini <paolo.carlini@oracle.com>
829 * cgraphunit.c (analyze_functions): Save input_location, set it
830 to UNKNOWN_LOCATION and restore it at the end.
832 2013-09-10 Martin Jambor <mjambor@suse.cz>
834 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
835 represented by a thunk.
837 2013-09-10 Jeff Law <law@redhat.com>
839 PR tree-optimization/58343
840 * tree-ssa-threadupdate.c (thread_block): Identify and disable
841 jump threading requests through loop headers buried in the middle
842 of a jump threading path.
844 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
845 in return value/type.
847 2013-09-10 Jakub Jelinek <jakub@redhat.com>
849 PR rtl-optimization/58365
850 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
851 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
854 2013-09-10 Richard Biener <rguenther@suse.de>
856 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
857 * tree-data-ref.c (create_rdg_vertices): Collect all data
858 references, signal failure to the caller, use data-ref API.
859 (build_rdg): Compute data references only once. Maintain lifetime
860 of data references and data dependences from within RDG.
861 (free_rdg): Free dependence relations.
862 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
863 inventing extra dependences.
864 (distribute_loop): Update for RDG API changes.
866 2013-09-10 Kai Tietz <ktietz@redhat.com>
868 * doc/invoke.texi (fms-extensions): Document changed
869 behavior for ms-abi targets.
870 * config/i386/i386.c (ix86_option_override_internal):
871 Set default value of option -fms-extension for ms-abi targets.
873 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
875 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
877 2013-09-10 Alan Modra <amodra@gmail.com>
880 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
882 2013-09-10 Alan Modra <amodra@gmail.com>
884 * config/rs6000/predicates.md (add_cint_operand): New.
885 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
886 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
887 using add_cint_operand.
888 (largetoc_high_plus_aix): Likewise.
890 2013-09-09 Jakub Jelinek <jakub@redhat.com>
892 PR tree-optimization/58364
893 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
894 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
895 the current range can't be an unconditional true or false.
897 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
899 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
901 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
903 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
905 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
908 * doc/invoke.texi (-Wdelete-incomplete): Document it.
910 2013-09-09 Ian Bolton <ian.bolton@arm.com>
912 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
913 NO_REGS for immediate that can't be moved directly into FP_REGS.
915 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
917 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
918 comparison with negated operand.
919 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
922 2013-09-09 Richard Biener <rguenther@suse.de>
925 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
926 of a subloop record all its block as affecting loop-closed SSA form.
928 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
930 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
931 of an rtx/bitpos pair.
932 (store_fixed_bit_field): Update accordingly.
934 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
936 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
938 * builtins.c (expand_errno_check): Likewise.
939 * dwarf2cfi.c (init_return_column_size): Likewise.
940 * except.c (sjlj_mark_call_sites): Likewise.
941 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
942 * lra-constraints.c (emit_inc): Likewise.
943 * ree.c (combine_set_extension): Likewise.
944 * regmove.c (fixup_match_2): Likewise.
945 * reload1.c (inc_for_reload): Likewise.
947 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
949 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
950 (force_to_mode, simplify_shift_const_1, simplify_comparison):
951 Use gen_int_mode with the mode of the associated simplify_* call.
952 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
953 * expmed.c (expand_shift_1): Likewise.
954 * function.c (instantiate_virtual_regs_in_insn): Likewise.
955 * loop-iv.c (iv_number_of_iterations): Likewise.
956 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
957 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
959 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
961 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
962 of the associated expand_* call.
963 (asan_emit_stack_protection): Likewise.
964 * builtins.c (round_trampoline_addr): Likewise.
965 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
966 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
967 (emit_store_flag): Likewise.
968 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
969 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
971 * function.c (instantiate_virtual_regs_in_insn): Likewise.
972 * ifcvt.c (noce_try_store_flag_constants): Likewise.
973 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
974 * modulo-sched.c (generate_prolog_epilog): Likewise.
975 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
976 (expand_ctz, expand_ffs, expand_unop): Likewise.
978 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
980 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
981 of the associated gen_rtx_* call.
982 * caller-save.c (init_caller_save): Likewise.
983 * combine.c (find_split_point, make_extraction): Likewise.
984 (make_compound_operation): Likewise.
985 * dwarf2out.c (mem_loc_descriptor): Likewise.
986 * explow.c (plus_constant, probe_stack_range): Likewise.
987 * expmed.c (expand_mult_const): Likewise.
988 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
989 * reload1.c (init_reload): Likewise.
990 * valtrack.c (cleanup_auto_inc_dec): Likewise.
991 * var-tracking.c (adjust_mems): Likewise.
992 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
993 rather than gen_rtx_fmt_ee.
995 2013-09-09 Jan Hubicka <jh@suse.cz>
998 * value-prof.c (gimple_ic): Copy also abnormal edges.
1000 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
1002 * asan.c (asan_shadow_cst): Use gen_int_mode.
1004 2013-09-08 Jan Hubicka <jh@suse.cz>
1006 * ipa-profile.c: Add toplevel comment.
1007 (ipa_propagate_frequency_1): Be more conservative when profile is read.
1008 (contains_hot_call_p): New function.
1009 (ipa_propagate_frequency): Set frequencies based on counts when
1011 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
1012 profile; do not tamper with profile after inlining if it is read.
1014 2013-09-08 Jan Hubicka <jh@suse.cz>
1016 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
1019 2013-09-08 Jan Hubicka <jh@suse.cz>
1021 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
1024 2013-09-08 Jeff Law <law@redhat.com>
1027 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
1030 2013-09-08 Andi Kleen <ak@linux.intel.com>
1032 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
1034 2013-09-08 Jan Hubicka <jh@suse.cz>
1036 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
1038 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
1040 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
1041 for non-debug insns.
1042 * lra.c (new_insn_reg): Take the containing insn as a parameter.
1043 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
1044 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
1047 2013-09-08 Jan Hubicka <jh@suse.cz>
1049 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
1050 targets and devirtualize to BUILT_IN_UNREACHABLE.
1051 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
1052 * ipa.c (walk_polymorphic_call_targets): New function.
1053 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
1054 functions; use the new timevar.
1055 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
1056 was removed from the program.
1057 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
1058 not consider it in the walk when its vtable is dead.
1059 (possible_polymorphic_call_targets_1): Pass anonymous flag to
1061 (devirt_variable_node_removal_hook): New function.
1062 (possible_polymorphic_call_targets): Also register
1063 devirt_variable_node_removal_hook.
1064 (ipa_devirt): Do not do non-speculative devirtualization.
1065 (gate_ipa_devirt): One execute if devirtualizing speculatively.
1067 2013-09-08 Jan Hubicka <jh@suse.cz>
1069 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
1070 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
1071 varpool_remove_variable_insertion_hook): Declare.
1072 * varpool.c (varpool_node_hook_list): New structure.
1073 (first_varpool_node_removal_hook,
1074 first_varpool_variable_insertion_hook): New variables.
1075 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
1076 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
1077 varpool_remove_variable_insertion_hook,
1078 varpool_call_variable_insertion_hooks): New functions.
1079 (varpool_remove_node): Use it.
1081 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
1084 * diagnostic.c (diagnostic_build_prefix): When s.file is
1085 "<built-in>" don't output line and column numbers.
1087 2013-09-06 Jan Hubicka <jh@suse.cz>
1089 * cgraphunit.c (expand_thunk): Get body before touching arguments.
1090 * lto-streamer-out.c: Stream thunks, too.
1091 * lto-streamer-in.c (input_function): Pop cfun here
1092 (lto_read_body): Instead of here.
1094 2013-09-06 Caroline Tice <cmtice@google.com>
1096 * doc/install.texi: Add documentation for the --enable-vtable-verify
1097 and the --disable-libvtv configure options.
1099 2013-09-06 Jeff Law <law@redhat.com>
1101 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
1102 edge implied equivalences into successor phis.
1104 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
1106 * resource.c (mark_referenced_resources): Handle COND_EXEC.
1108 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
1110 * resource.c (mark_target_live_regs): Compute resources taking
1111 into account if a call is predicated or not.
1113 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1115 * toplev.c (output_stack_usage): Be prepared for suffixes created by
1116 the compiler in the function names.
1118 2013-09-06 Jan Hubicka <jh@suse.cz>
1121 * ipa-inline.c (has_caller_p): New function.
1122 (want_inline_function_to_all_callers_p): Use it.
1123 (sum_callers, inline_to_all_callers): Break out from ...
1124 (ipa_inline): ... here.
1126 2013-09-06 Jan Hubicka <jh@suse.cz>
1128 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
1129 only when AVX is enabled.
1131 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1133 * config/aarch64/aarch64.md
1134 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
1135 is fpsimd_<load/store>2.
1136 (load_pair<mode>): Likewise.
1137 (store_pair<mode>): Likewise.
1139 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1141 * config/arm/types.md (type): Add "mrs" type.
1142 * config/aarch64/aarch64.md
1143 (aarch64_load_tp_hard): Make type "mrs".
1145 (load_tp_hard): Make type "mrs".
1146 * config/arm/cortex-a15.md: Update with new attributes.
1147 * config/arm/cortex-a5.md: Update with new attributes.
1148 * config/arm/cortex-a53.md: Update with new attributes.
1149 * config/arm/cortex-a7.md: Update with new attributes.
1150 * config/arm/cortex-a8.md: Update with new attributes.
1151 * config/arm/cortex-a9.md: Update with new attributes.
1152 * config/arm/cortex-m4.md: Update with new attributes.
1153 * config/arm/cortex-r4.md: Update with new attributes.
1154 * config/arm/fa526.md: Update with new attributes.
1155 * config/arm/fa606te.md: Update with new attributes.
1156 * config/arm/fa626te.md: Update with new attributes.
1157 * config/arm/fa726te.md: Update with new attributes.
1159 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1161 * config/aarch64/aarch64.md
1162 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
1163 (*movtf_aarch64): Likewise.
1165 (thumb1_movdi_insn): Use "multiple" for type where more than one
1166 instruction is used for a move.
1167 (*arm32_movhf): Likewise.
1168 (*thumb_movdf_insn): Likewise.
1170 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1172 * config/arm/types.md (type): Rename fcpys to fmov.
1174 (*arm_movsi_vfp): Rename type fcpys as fmov.
1175 (*thumb2_movsi_vfp): Likewise
1176 (*movhf_vfp_neon): Likewise
1177 (*movhf_vfp): Likewise
1178 (*movsf_vfp): Likewise
1179 (*thumb2_movsf_vfp): Likewise
1180 (*movsfcc_vfp): Likewise
1181 (*thumb2_movsfcc_vfp): Likewise
1182 * config/aarch64/aarch64-simd.md
1183 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
1184 * config/aarch64/aarch64.md
1185 (*movsi_aarch64): Replace type mov_reg with fmovs.
1186 (*movdi_aarch64): Likewise
1187 (*movsf_aarch64): Likewise
1188 (*movdf_aarch64): Likewise
1190 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
1191 * config/arm/iwmmxt.md
1192 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
1193 * config/arm/arm1020e.md: Update with new attributes.
1194 * config/arm/cortex-a15-neon.md: Update with new attributes.
1195 * config/arm/cortex-a5.md: Update with new attributes.
1196 * config/arm/cortex-a53.md: Update with new attributes.
1197 * config/arm/cortex-a7.md: Update with new attributes.
1198 * config/arm/cortex-a8-neon.md: Update with new attributes.
1199 * config/arm/cortex-a9.md: Update with new attributes.
1200 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1201 * config/arm/cortex-r4f.md: Update with new attributes.
1202 * config/arm/marvell-pj4.md: Update with new attributes.
1203 * config/arm/vfp11.md: Update with new attributes.
1205 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1207 * config/aarch64/aarch64.md
1208 (*madd<mode>): Fix type attribute.
1209 (*maddsi_uxtw): Likewise.
1210 (*msub<mode>): Likewise.
1211 (*msubsi_uxtw): Likewise.
1212 (<su_optab>maddsidi4): Likewise.
1213 (<su_optab>msubsidi4): Likewise.
1215 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1217 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
1218 * config/arm/arm.md (core_cycles): Remove fdiv.
1219 * config/arm/vfp.md:
1220 (*sqrtsf2_vfp): Update for attribute changes.
1221 (*sqrtdf2_vfp): Likewise.
1222 * config/aarch64/aarch64.md:
1223 (sqrt<mode>2): Update for attribute changes.
1224 * config/arm/arm1020e.md: Update with new attributes.
1225 * config/arm/cortex-a15-neon.md: Update with new attributes.
1226 * config/arm/cortex-a5.md: Update with new attributes.
1227 * config/arm/cortex-a53.md: Update with new attributes.
1228 * config/arm/cortex-a7.md: Update with new attributes.
1229 * config/arm/cortex-a8-neon.md: Update with new attributes.
1230 * config/arm/cortex-a9.md: Update with new attributes.
1231 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1232 * config/arm/cortex-r4f.md: Update with new attributes.
1233 * config/arm/marvell-pj4.md: Update with new attributes.
1234 * config/arm/vfp11.md: Update with new attributes.
1236 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1238 * config/arm/types.md
1239 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
1240 * config/aarch64/aarch64.md
1241 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
1243 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
1244 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
1245 (float<GPI:mode><GPF:mode>2): Likewise.
1247 (*truncsisf2_vfp): Update with new attributes.
1248 (*truncsidf2_vfp): Likewise.
1249 (fixuns_truncsfsi2): Likewise.
1250 (fixuns_truncdfsi2): Likewise.
1251 (*floatsisf2_vfp): Likewise.
1252 (*floatsidf2_vfp): Likewise.
1253 (floatunssisf2): Likewise.
1254 (floatunssidf2): Likewise.
1255 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
1256 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
1257 * config/arm/arm1020e.md: Update with new attributes.
1258 * config/arm/cortex-a15-neon.md: Update with new attributes.
1259 * config/arm/cortex-a5.md: Update with new attributes.
1260 * config/arm/cortex-a53.md: Update with new attributes.
1261 * config/arm/cortex-a7.md: Update with new attributes.
1262 * config/arm/cortex-a8-neon.md: Update with new attributes.
1263 * config/arm/cortex-a9.md: Update with new attributes.
1264 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1265 * config/arm/cortex-r4f.md: Update with new attributes.
1266 * config/arm/marvell-pj4.md: Update with new attributes.
1267 * config/arm/vfp11.md: Update with new attributes.
1269 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1271 * config/aarch64/arm_neon.h
1272 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
1273 (vqtbx<1,2,3,4><q>_s8): Likewise.
1275 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1277 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
1278 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
1279 (add<mode>3): Add type attribute.
1280 (add<mode>3): Likewise.
1281 (usadd<mode>3): Likewise.
1282 (ssadd<mode>3): Likewise.
1283 (sub<mode>3): Likewise.
1284 (sub<mode>3): Likewise.
1285 (ussub<mode>3): Likewise.
1286 (sssub<mode>3): Likewise.
1287 (ssmulsa3): Likewise.
1288 (usmulusa3): Likewise.
1289 (arm_usatsihi): Likewise.
1291 (*movdi_vfp): Add types for all instructions.
1292 (*movdi_vfp_cortexa8): Likewise.
1293 (*movhf_vfp_neon): Likewise.
1294 (*movhf_vfp): Likewise.
1295 (*movdf_vfp): Likewise.
1296 (*thumb2_movdf_vfp): Likewise.
1297 (*thumb2_movdfcc_vfp): Likewise.
1298 * config/arm/arm.md: Add type attribute to all insn patterns.
1299 (*thumb1_adddi3): Add type attribute.
1300 (*arm_adddi3): Likewise.
1301 (*adddi_sesidi_di): Likewise.
1302 (*adddi_zesidi_di): Likewise.
1303 (*thumb1_addsi3): Likewise.
1304 (addsi3_compare0): Likewise.
1305 (*addsi3_compare0_scratch): Likewise.
1306 (*compare_negsi_si): Likewise.
1307 (cmpsi2_addneg): Likewise.
1308 (*addsi3_carryin_<optab>): Likewise.
1309 (*addsi3_carryin_alt2_<optab>): Likewise.
1310 (*addsi3_carryin_clobercc_<optab>): Likewise.
1311 (*subsi3_carryin): Likewise.
1312 (*subsi3_carryin_const): Likewise.
1313 (*subsi3_carryin_compare): Likewise.
1314 (*subsi3_carryin_compare_const): Likewise.
1315 (*arm_subdi3): Likewise.
1316 (*thumb_subdi3): Likewise.
1317 (*subdi_di_zesidi): Likewise.
1318 (*subdi_di_sesidi): Likewise.
1319 (*subdi_zesidi_di): Likewise.
1320 (*subdi_sesidi_di): Likewise.
1321 (*subdi_zesidi_ze): Likewise.
1322 (thumb1_subsi3_insn): Likewise.
1323 (*arm_subsi3_insn): Likewise.
1324 (*anddi3_insn): Likewise.
1325 (*anddi_zesidi_di): Likewise.
1326 (*anddi_sesdi_di): Likewise.
1327 (*ne_zeroextracts): Likewise.
1328 (*ne_zeroextracts): Likewise.
1329 (*ite_ne_zeroextr): Likewise.
1330 (*ite_ne_zeroextr): Likewise.
1331 (*anddi_notdi_di): Likewise.
1332 (*anddi_notzesidi): Likewise.
1333 (*anddi_notsesidi): Likewise.
1334 (andsi_notsi_si): Likewise.
1335 (thumb1_bicsi3): Likewise.
1336 (*iordi3_insn): Likewise.
1337 (*iordi_zesidi_di): Likewise.
1338 (*iordi_sesidi_di): Likewise.
1339 (*thumb1_iorsi3_insn): Likewise.
1340 (*xordi3_insn): Likewise.
1341 (*xordi_zesidi_di): Likewise.
1342 (*xordi_sesidi_di): Likewise.
1343 (*arm_xorsi3): Likewise.
1344 (*andsi_iorsi3_no): Likewise.
1345 (*smax_0): Likewise.
1346 (*smax_m1): Likewise.
1347 (*arm_smax_insn): Likewise.
1348 (*smin_0): Likewise.
1349 (*arm_smin_insn): Likewise.
1350 (*arm_umaxsi3): Likewise.
1351 (*arm_uminsi3): Likewise.
1352 (*minmax_arithsi): Likewise.
1353 (*minmax_arithsi_): Likewise.
1354 (*satsi_<SAT:code>): Likewise.
1355 (arm_ashldi3_1bit): Likewise.
1356 (arm_ashrdi3_1bit): Likewise.
1357 (arm_lshrdi3_1bit): Likewise.
1358 (*arm_negdi2): Likewise.
1359 (*thumb1_negdi2): Likewise.
1360 (*arm_negsi2): Likewise.
1361 (*thumb1_negsi2): Likewise.
1362 (*negdi_extendsid): Likewise.
1363 (*negdi_zero_extend): Likewise.
1364 (*arm_abssi2): Likewise.
1365 (*thumb1_abssi2): Likewise.
1366 (*arm_neg_abssi2): Likewise.
1367 (*thumb1_neg_abss): Likewise.
1368 (one_cmpldi2): Likewise.
1369 (extend<mode>di2): Likewise.
1370 (*compareqi_eq0): Likewise.
1371 (*arm_extendhisi2addsi): Likewise.
1372 (*arm_movdi): Likewise.
1373 (*thumb1_movdi_insn): Likewise.
1374 (*arm_movt): Likewise.
1375 (*thumb1_movsi_insn): Likewise.
1376 (pic_add_dot_plus_four): Likewise.
1377 (pic_add_dot_plus_eight): Likewise.
1378 (tls_load_dot_plus_eight): Likewise.
1379 (*thumb1_movhi_insn): Likewise.
1380 (*thumb1_movsf_insn): Likewise.
1381 (*movdf_soft_insn): Likewise.
1382 (*thumb_movdf_insn): Likewise.
1383 (cbranchsi4_insn): Likewise.
1384 (cbranchsi4_scratch): Likewise.
1385 (*negated_cbranchsi4): Likewise.
1386 (*tbit_cbranch): Likewise.
1387 (*tlobits_cbranch): Likewise.
1388 (*tstsi3_cbranch): Likewise.
1389 (*cbranchne_decr1): Likewise.
1390 (*addsi3_cbranch): Likewise.
1391 (*addsi3_cbranch_scratch): Likewise.
1392 (*arm_cmpdi_insn): Likewise.
1393 (*arm_cmpdi_unsig): Likewise.
1394 (*arm_cmpdi_zero): Likewise.
1395 (*thumb_cmpdi_zero): Likewise.
1396 (*deleted_compare): Likewise.
1397 (*mov_scc): Likewise.
1398 (*mov_negscc): Likewise.
1399 (*mov_notscc): Likewise.
1400 (*cstoresi_eq0_thumb1_insn): Likewise.
1401 (cstoresi_nltu_thumb1): Likewise.
1402 (cstoresi_ltu_thu): Likewise.
1403 (thumb1_addsi3_addgeu): Likewise.
1404 (*arm_jump): Likewise.
1405 (*thumb_jump): Likewise.
1406 (*check_arch2): Likewise.
1407 (arm_casesi_internal): Likewise.
1408 (thumb1_casesi_dispatch): Likewise.
1409 (*arm_indirect_jump): Likewise.
1410 (*thumb1_indirect_jump): Likewise.
1412 (*and_scc): Likewise.
1413 (*ior_scc): Likewise.
1414 (*compare_scc): Likewise.
1415 (*cond_move): Likewise.
1416 (*cond_arith): Likewise.
1417 (*cond_sub): Likewise.
1418 (*cmp_ite0): Likewise.
1419 (*cmp_ite1): Likewise.
1420 (*cmp_and): Likewise.
1421 (*cmp_ior): Likewise.
1422 (*ior_scc_scc): Likewise.
1423 (*ior_scc_scc_cmp): Likewise.
1424 (*and_scc_scc): Likewise.
1425 (*and_scc_scc_cmp): Likewise.
1426 (*and_scc_scc_nod): Likewise.
1427 (*negscc): Likewise.
1428 (movcond_addsi): Likewise.
1429 (movcond): Likewise.
1430 (*ifcompare_plus_move): Likewise.
1431 (*if_plus_move): Likewise.
1432 (*ifcompare_move_plus): Likewise.
1433 (*if_move_plus): Likewise.
1434 (*ifcompare_arith_arith): Likewise.
1435 (*if_arith_arith): Likewise.
1436 (*ifcompare_arith_move): Likewise.
1437 (*if_arith_move): Likewise.
1438 (*ifcompare_move_arith): Likewise.
1439 (*if_move_arith): Likewise.
1440 (*ifcompare_move_not): Likewise.
1441 (*if_move_not): Likewise.
1442 (*ifcompare_not_move): Likewise.
1443 (*if_not_move): Likewise.
1444 (*ifcompare_shift_move): Likewise.
1445 (*if_shift_move): Likewise.
1446 (*ifcompare_move_shift): Likewise.
1447 (*if_move_shift): Likewise.
1448 (*ifcompare_shift_shift): Likewise.
1449 (*ifcompare_not_arith): Likewise.
1450 (*ifcompare_arith_not): Likewise.
1451 (*if_arith_not): Likewise.
1452 (*ifcompare_neg_move): Likewise.
1453 (*if_neg_move): Likewise.
1454 (*ifcompare_move_neg): Likewise.
1455 (*if_move_neg): Likewise.
1456 (prologue_thumb1_interwork): Likewise.
1457 (*cond_move_not): Likewise.
1458 (*sign_extract_onebit): Likewise.
1459 (*not_signextract_onebit): Likewise.
1460 (stack_tie): Likewise.
1461 (align_4): Likewise.
1462 (align_8): Likewise.
1463 (consttable_end): Likewise.
1464 (consttable_1): Likewise.
1465 (consttable_2): Likewise.
1466 (consttable_4): Likewise.
1467 (consttable_8): Likewise.
1468 (consttable_16): Likewise.
1469 (*thumb1_tablejump): Likewise.
1470 (prefetch): Likewise.
1471 (force_register_use): Likewise.
1472 (thumb_eh_return): Likewise.
1473 (load_tp_hard): Likewise.
1474 (load_tp_soft): Likewise.
1475 (tlscall): Likewise.
1476 (*arm_movtas_ze): Likewise.
1477 (*arm_rev): Likewise.
1478 (*arm_revsh): Likewise.
1479 (*arm_rev16): Likewise.
1480 * config/arm/thumb2.md
1481 (*thumb2_smaxsi3): Likewise.
1482 (*thumb2_sminsi3): Likewise.
1483 (*thumb32_umaxsi3): Likewise.
1484 (*thumb2_uminsi3): Likewise.
1485 (*thumb2_negdi2): Likewise.
1486 (*thumb2_abssi2): Likewise.
1487 (*thumb2_neg_abss): Likewise.
1488 (*thumb2_movsi_insn): Likewise.
1489 (tls_load_dot_plus_four): Likewise.
1490 (*thumb2_movhi_insn): Likewise.
1491 (*thumb2_mov_scc): Likewise.
1492 (*thumb2_mov_negs): Likewise.
1493 (*thumb2_mov_negs): Likewise.
1494 (*thumb2_mov_nots): Likewise.
1495 (*thumb2_mov_nots): Likewise.
1496 (*thumb2_movsicc_): Likewise.
1497 (*thumb2_movsfcc_soft_insn): Likewise.
1498 (*thumb2_indirect_jump): Likewise.
1499 (*thumb2_and_scc): Likewise.
1500 (*thumb2_ior_scc): Likewise.
1501 (*thumb2_ior_scc_strict_it): Likewise.
1502 (*thumb2_cond_move): Likewise.
1503 (*thumb2_cond_arith): Likewise.
1504 (*thumb2_cond_ari): Likewise.
1505 (*thumb2_cond_sub): Likewise.
1506 (*thumb2_negscc): Likewise.
1507 (*thumb2_movcond): Likewise.
1508 (thumb2_casesi_internal): Likewise.
1509 (thumb2_casesi_internal_pic): Likewise.
1510 (*thumb2_alusi3_short): Likewise.
1511 (*thumb2_mov<mode>_shortim): Likewise.
1512 (*thumb2_addsi_short): Likewise.
1513 (*thumb2_subsi_short): Likewise.
1514 (thumb2_addsi3_compare0): Likewise.
1515 (*thumb2_cbz): Likewise.
1516 (*thumb2_cbnz): Likewise.
1517 (*thumb2_one_cmplsi2_short): Likewise.
1518 (*thumb2_negsi2_short): Likewise.
1519 (*orsi_notsi_si): Likewise.
1520 * config/arm/arm1020e.md: Update with new attributes.
1521 * config/arm/arm1026ejs.md: Update with new attributes.
1522 * config/arm/arm1136jfs.md: Update with new attributes.
1523 * config/arm/arm926ejs.md: Update with new attributes.
1524 * config/arm/cortex-a15.md: Update with new attributes.
1525 * config/arm/cortex-a5.md: Update with new attributes.
1526 * config/arm/cortex-a53.md: Update with new attributes.
1527 * config/arm/cortex-a7.md: Update with new attributes.
1528 * config/arm/cortex-a8.md: Update with new attributes.
1529 * config/arm/cortex-a9.md: Update with new attributes.
1530 * config/arm/cortex-m4.md: Update with new attributes.
1531 * config/arm/cortex-r4.md: Update with new attributes.
1532 * config/arm/fa526.md: Update with new attributes.
1533 * config/arm/fa606te.md: Update with new attributes.
1534 * config/arm/fa626te.md: Update with new attributes.
1535 * config/arm/fa726te.md: Update with new attributes.
1537 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1539 * config/aarch64/aarch64-simd.md
1540 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1541 <vwx> iterator to ensure correct register choice.
1542 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1543 (aarch64_sqdmull_n<mode>): Likewise.
1544 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1545 * config/aarch64/arm_neon.h
1546 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1547 (vml<as><q>_n_<su>16): Likewise.
1548 (vml<as>l_high_lane<q>_<su>16): Likewise.
1549 (vml<as>l_high_n_<su>16): Likewise.
1550 (vml<as>l_lane<q>_<su>16): Likewise.
1551 (vml<as>l_n_<su>16): Likewise.
1552 (vmul<q>_lane<q>_<su>16): Likewise.
1553 (vmul<q>_n_<su>16): Likewise.
1554 (vmull_lane<q>_<su>16): Likewise.
1555 (vmull_n_<su>16): Likewise.
1556 (vmull_high_lane<q>_<su>16): Likewise.
1557 (vmull_high_n_<su>16): Likewise.
1558 (vqrdmulh<q>_n_s16): Likewise.
1560 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1562 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1563 have the correct lane parameter.
1565 2013-09-06 Richard Biener <rguenther@suse.de>
1567 * cfganal.c (control_dependences::~control_dependences):
1568 Properly free all of the vector.
1570 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1573 * config/i386/i386.c (ix86_conditional_register_usage):
1574 Proper initialize extended SSE registers.
1576 2013-09-06 Jan Hubicka <jh@suse.cz>
1578 PR tree-optimization/58311
1579 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1581 2013-09-06 Jan Hubicka <jh@suse.cz>
1583 * Makefile.in (tree-sra.o): Update dependencies.
1584 * tree-sra.c: Include ipa-utils.h
1585 (scan_function): Use recursive_call_p.
1586 (has_caller_p): New function.
1587 (cgraph_for_node_and_aliases): Count also callers of aliases.
1589 2013-09-06 Jan Hubicka <jh@suse.cz>
1592 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1593 * tree-tailcall.c: Include ipa-utils.h.
1594 (find_tail_calls): Use it.
1595 * ipa-pure-const.c (check_call): Likewise.
1596 * ipa-utils.c (recursive_call_p): New function.
1597 * ipa-utils.h (recursive_call_p): Dclare.
1598 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1599 (symtab_semantically_equivalent_p): New function.
1600 * Makefile.in (tree-tailcall.o): Update dependencies.
1602 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1604 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1607 2013-09-06 Richard Biener <rguenther@suse.de>
1609 * lto-streamer.h (lto_global_var_decls): Remove.
1610 * Makefile.in (OBJS): Remove lto-symtab.o.
1611 (lto-symtab.o): Remove.
1612 (GTFILES): Remove lto-symtab.c
1613 * lto-symtab.c: Move to lto/
1615 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1617 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1618 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1619 (UNSPEC_FPINT_RINT): New constant definitions.
1620 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1621 definition with 2 attributes.
1622 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1623 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1626 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1628 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
1629 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1630 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1631 attribute to "z196_alone".
1632 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1635 2013-09-06 Richard Biener <rguenther@suse.de>
1637 * basic-block.h (class control_dependences): New.
1638 * tree-ssa-dce.c (control_dependence_map): Remove.
1640 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1641 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1642 find_pdom, find_control_dependence, find_all_control_dependences):
1644 (mark_control_dependent_edges_necessary,
1645 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1646 tree_dce_done, perform_tree_ssa_dce): Adjust.
1647 * cfganal.c (set_control_dependence_map_bit,
1648 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1649 find_all_control_dependences): Move from tree-ssa-dce.c and
1650 implement as methods of control_dependences class.
1651 (control_dependences::control_dependences): New.
1652 (control_dependences::~control_dependences): Likewise.
1653 (control_dependences::get_edges_dependent_on): Likewise.
1654 (control_dependences::get_edge): Likewise.
1656 2013-09-04 Jan Hubicka <jh@suse.cz>
1658 * tree.c (types_same_for_odr): Drop overactive check.
1659 * ipa-devirt.c (hash_type_name): Likewise.
1661 2013-09-04 Jan Hubicka <jh@suse.cz>
1663 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1664 (analyze_functions): ... here.
1666 2013-09-04 Jan Hubicka <jh@suse.cz>
1669 * cgraphunit.c (analyze_functions): Clear AUX fields
1670 after processing; initialize assembler name has.
1672 2013-09-05 Jeff Law <law@redhat.com>
1674 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1675 from thread_around_empty_block. Record threading path into PATH.
1676 Recurse if threading through the initial block is successful.
1677 (thread_across_edge): Corresponding changes to slightly simplify.
1679 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1681 * config/aarch64/aarch64.md
1682 (type): Remove frecpe, frecps, frecpx.
1683 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1684 fix to be a TARGET_SIMD instruction.
1685 (aarch64_frecps): Remove.
1686 * config/aarch64/aarch64-simd.md
1687 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1688 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1690 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1691 Sofiane Naci <sofiane.naci@arm.com>
1693 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
1694 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
1695 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1696 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1697 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1698 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1699 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1700 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1701 into "clz, "rbit". Rename "shift" to "shift_imm".
1702 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
1703 changes. Update for attribute changes all occurrences of arlo_* and
1705 * config/arm/arm-fixed.md: Update for attribute changes
1706 all occurrences of arlo_* types.
1707 * config/arm/thumb2.md: Update for attribute changes all occurrences
1709 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1710 (cortexa7_older_only): Likewise.
1711 (cortexa7_younger): Likewise.
1712 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1713 (1020alu_shift_op): Likewise.
1714 (1020alu_shift_reg_op): Likewise.
1715 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1716 (alu_shift_op): Likewise.
1717 (alu_shift_reg_op): Likewise.
1718 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
1719 (11_alu_shift_op): Likewise.
1720 (11_alu_shift_reg_op): Likewise.
1721 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1722 (9_alu_shift_reg_op): Likewise.
1723 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1725 (cortex_a15_alu_shift): Likewise.
1726 (cortex_a15_alu_shift_reg): Likewise.
1727 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1729 (cortex_a5_alu_shift): Likewise.
1730 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1732 (cortex_a53_alu_shift): Likewise.
1733 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1735 (cortex_a7_alu_reg): Likewise.
1736 (cortex_a7_alu_shift): Likewise.
1737 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1739 (cortex_a8_alu_shift): Likewise.
1740 (cortex_a8_alu_shift_reg): Likewise.
1741 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
1742 (cortex_a9_dp_shift): Likewise.
1743 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1745 * config/arm/cortex-r4.md
1746 (cortex_r4_alu): Update for attribute changes.
1747 (cortex_r4_mov): Likewise.
1748 (cortex_r4_alu_shift_reg): Likewise.
1749 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
1750 (526_alu_shift_op): Likewise.
1751 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1752 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
1753 (626te_alu_shift_op): Likewise.
1754 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
1755 (726te_alu_shift_op): Likewise.
1756 (726te_alu_shift_reg_op): Likewise.
1757 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1758 (mp626_alu_shift_op): Likewise.
1759 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1760 (pj4_alu_conds): Likewise.
1761 (pj4_shift): Likewise.
1762 (pj4_shift_conds): Likewise.
1763 (pj4_alu_shift): Likewise.
1764 (pj4_alu_shift_conds): Likewise.
1765 * config/aarch64/aarch64.md: Update for attribute change
1766 all occurrences of arlo_* and shift* types.
1768 2013-09-05 Mike Stump <mikestump@comcast.net>
1770 * tree.h: Move documentation for tree_function_decl to tree-core.h
1771 with the declaration.
1773 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1776 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1777 looking for widest mode.
1779 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1781 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1783 2013-09-05 Richard Biener <rguenther@suse.de>
1785 PR tree-optimization/58137
1786 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1787 Do not create vectors of pointers.
1788 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1789 types for the components of the vector initializer.
1790 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1791 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1793 2013-09-05 Martin Jambor <mjambor@suse.cz>
1795 * ipa-prop.c (remove_described_reference): Accept missing references,
1796 return false if that hppens, otherwise return true.
1797 (cgraph_node_for_jfunc): New function.
1798 (try_decrement_rdesc_refcount): Likewise.
1799 (try_make_edge_direct_simple_call): Use them.
1800 (ipa_edge_removal_hook): Remove references from rdescs.
1801 (ipa_edge_duplication_hook): Clone rdescs and their references
1802 when the new edge has the same caller as the old one.
1803 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1804 reference before removing any edges.
1806 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
1808 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1810 * thumb2.md (thumb2_storewb_parisi): New pattern.
1812 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1814 * config/aarch64/aarch64-option-extensions.def: Add
1815 AARCH64_OPT_EXTENSION of 'crc'.
1816 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1817 (AARCH64_ISA_CRC): Ditto.
1818 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1819 description of the CRC extension.
1821 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1823 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1824 * config/rs6000/linux.h: Ditto.
1825 * alpha/linux.h: Ditto.
1826 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1827 no_c99_libc_has_function.
1828 * config/c6x/uclinux-elf.h: Ditto.
1829 * config/lm32/uclinux-elf.h: Ditto.
1830 * config/m68k/uclinux.h: Ditto.
1831 * config/moxie/uclinux.h: Ditto.
1832 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1833 (crisv32-*-linux*, cris-*-linux*): Ditto.
1834 * config/bfin/bfin.c: Include "tm_p.h".
1836 2013-09-05 Richard Biener <rguenther@suse.de>
1838 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1839 check for a definition without a basic-block.
1841 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1842 Sofiane Naci <sofiane.naci@arm.com>
1844 * config/aarch64/aarch64.md
1845 (*movti_aarch64): Rename r_2_f and f_2_r.
1846 (*movsf_aarch64): Likewise.
1847 (*movdf_aarch64): Likewise.
1848 (*movtf_aarch64): Likewise.
1849 (aarch64_movdi_<mode>low): Likewise.
1850 (aarch64_movdi_<mode>high): Likewise.
1851 (aarch64_mov<mode>high_di): Likewise.
1852 (aarch64_mov<mode>low_di): Likewise.
1853 (aarch64_movtilow_tilow): Likewise.
1854 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1855 values to config/arm/types.md
1856 (attribute "conds"): Update for attribute change.
1857 (anddi3_insn): Likewise.
1858 (iordi3_insn): Likewise.
1859 (xordi3_insn): Likewise.
1860 (one_cmpldi2): Likewise.
1861 * config/arm/types.md (type): Add Neon types.
1862 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1863 use "type" attribute.
1864 (movmisalign<mode>_neon_store): Likewise.
1865 (movmisalign<mode>_neon_load): Likewise.
1866 (vec_set<mode>_internal): Likewise.
1867 (vec_setv2di_internal): Likewise.
1868 (vec_extract<mode>): Likewise.
1869 (vec_extractv2di): Likewise.
1870 (add<mode>3_neon): Likewise.
1871 (adddi3_neon): Likewise.
1872 (sub<mode>3_neon): Likewise.
1873 (subdi3_neon): Likewise.
1874 (mul<mode>3_neon): Likewise.
1875 (mul<mode>3add<mode>_neon): Likewise.
1876 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1877 (fma<VCVTF:mode>4)): Likewise.
1878 (fma<VCVTF:mode>4_intrinsic): Likewise.
1879 (fmsub<VCVTF:mode>4)): Likewise.
1880 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1881 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1882 (ior<mode>3): Likewise.
1883 (and<mode>3): Likewise.
1884 (anddi3_neon): Likewise.
1885 (orn<mode>3_neon): Likewise.
1886 (orndi3_neon): Likewise.
1887 (bic<mode>3_neon): Likewise.
1888 (bicdi3_neon): Likewise.
1889 (xor<mode>3): Likewise.
1890 (one_cmpl<mode>2): Likewise.
1891 (abs<mode>2): Likewise.
1892 (neg<mode>2): Likewise.
1893 (umin<mode>3_neon): Likewise.
1894 (umax<mode>3_neon): Likewise.
1895 (smin<mode>3_neon): Likewise.
1896 (smax<mode>3_neon): Likewise.
1897 (vashl<mode>3): Likewise.
1898 (vashr<mode>3_imm): Likewise.
1899 (vlshr<mode>3_imm): Likewise.
1900 (ashl<mode>3_signed): Likewise.
1901 (ashl<mode>3_unsigned): Likewise.
1902 (neon_load_count): Likewise.
1903 (ashldi3_neon_noclobber): Likewise.
1904 (signed_shift_di3_neon): Likewise.
1905 (unsigned_shift_di3_neon): Likewise.
1906 (ashrdi3_neon_imm_noclobber): Likewise.
1907 (lshrdi3_neon_imm_noclobber): Likewise.
1908 (widen_ssum<mode>3): Likewise.
1909 (widen_usum<mode>3): Likewise.
1910 (quad_halves_<code>v4si): Likewise.
1911 (quad_halves_<code>v4sf): Likewise.
1912 (quad_halves_<code>v8hi): Likewise.
1913 (quad_halves_<code>v16qi): Likewise.
1914 (reduc_splus_v2di): Likewise.
1915 (neon_vpadd_internal<mode>): Likewise.
1916 (neon_vpsmin<mode>): Likewise.
1917 (neon_vpsmax<mode>): Likewise.
1918 (neon_vpumin<mode>): Likewise.
1919 (neon_vpumax<mode>): Likewise.
1920 (ss_add<mode>_neon): Likewise.
1921 (us_add<mode>_neon): Likewise.
1922 (ss_sub<mode>_neon): Likewise.
1923 (us_sub<mode>_neon): Likewise.
1924 (neon_vadd<mode>_unspec): Likewise.
1925 (neon_vaddl<mode>): Likewise.
1926 (neon_vaddw<mode>): Likewise.
1927 (neon_vhadd<mode>): Likewise.
1928 (neon_vqadd<mode>): Likewise.
1929 (neon_vaddhn<mode>): Likewise.
1930 (neon_vmul<mode>): Likewise.
1931 (neon_vmla<mode>): Likewise.
1932 (neon_vmlal<mode>): Likewise.
1933 (neon_vmls<mode>): Likewise.
1934 (neon_vmlsl<mode>): Likewise.
1935 (neon_vqdmulh<mode>): Likewise.
1936 (neon_vqdmlal<mode>): Likewise.
1937 (neon_vqdmlsl<mode>): Likewise.
1938 (neon_vmull<mode>): Likewise.
1939 (neon_vqdmull<mode>): Likewise.
1940 (neon_vsub<mode>_unspec): Likewise.
1941 (neon_vsubl<mode>): Likewise.
1942 (neon_vsubw<mode>): Likewise.
1943 (neon_vqsub<mode>): Likewise.
1944 (neon_vhsub<mode>): Likewise.
1945 (neon_vsubhn<mode>): Likewise.
1946 (neon_vceq<mode>): Likewise.
1947 (neon_vcge<mode>): Likewise.
1948 (neon_vcgeu<mode>): Likewise.
1949 (neon_vcgt<mode>): Likewise.
1950 (neon_vcgtu<mode>): Likewise.
1951 (neon_vcle<mode>): Likewise.
1952 (neon_vclt<mode>): Likewise.
1953 (neon_vcage<mode>): Likewise.
1954 (neon_vcagt<mode>): Likewise.
1955 (neon_vtst<mode>): Likewise.
1956 (neon_vabd<mode>): Likewise.
1957 (neon_vabdl<mode>): Likewise.
1958 (neon_vaba<mode>): Likewise.
1959 (neon_vabal<mode>): Likewise.
1960 (neon_vmax<mode>): Likewise.
1961 (neon_vmin<mode>): Likewise.
1962 (neon_vpaddl<mode>): Likewise.
1963 (neon_vpadal<mode>): Likewise.
1964 (neon_vpmax<mode>): Likewise.
1965 (neon_vpmin<mode>): Likewise.
1966 (neon_vrecps<mode>): Likewise.
1967 (neon_vrsqrts<mode>): Likewise.
1968 (neon_vqabs<mode>): Likewise.
1969 (neon_vqneg<mode>): Likewise.
1970 (neon_vcls<mode>): Likewise.
1971 (clz<mode>2): Likewise.
1972 (popcount<mode>2): Likewise.
1973 (neon_vrecpe): Likewise.
1974 (neon_vrsqrte): Likewise.
1975 (neon_vget_lane<mode>_sext_internal): Likewise.
1976 (neon_vget_lane<mode>_zext_internal): Likewise.
1977 (neon_vdup_n<mode>): Likewise.
1978 (neon_vdup_nv2di): Likewise.
1979 (neon_vdpu_lane<mode>_internal): Likewise.
1980 (neon_vswp<mode>): Likewise.
1981 (float<mode><V_cvtto>2): Likewise.
1982 (floatuns<mode><V_cvtto>2): Likewise.
1983 (fix_trunc<mode><V_cvtto>)2): Likewise
1984 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1985 (neon_vcvt<mode>): Likewise.
1986 (neon_vcvtv4sfv4hf): Likewise.
1987 (neon_vcvtv4hfv4sf): Likewise.
1988 (neon_vcvt_n<mode>): Likewise.
1989 (neon_vmovn<mode>): Likewise.
1990 (neon_vqmovn<mode>): Likewise.
1991 (neon_vqmovun<mode>): Likewise.
1992 (neon_vmovl<mode>): Likewise.
1993 (neon_vmul_lane<mode>): Likewise.
1994 (neon_vmull_lane<mode>): Likewise.
1995 (neon_vqdmull_lane<mode>): Likewise.
1996 (neon_vqdmulh_lane<mode>): Likewise.
1997 (neon_vmla_lane<mode>): Likewise.
1998 (neon_vmlal_lane<mode>): Likewise.
1999 (neon_vqdmlal_lane<mode>): Likewise.
2000 (neon_vmls_lane<mode>): Likewise.
2001 (neon_vmlsl_lane<mode>): Likewise.
2002 (neon_vqdmlsl_lane<mode>): Likewise.
2003 (neon_vext<mode>): Likewise.
2004 (neon_vrev64<mode>): Likewise.
2005 (neon_vrev32<mode>): Likewise.
2006 (neon_vrev16<mode>): Likewise.
2007 (neon_vbsl<mode>_internal): Likewise.
2008 (neon_vshl<mode>): Likewise.
2009 (neon_vqshl<mode>): Likewise.
2010 (neon_vshr_n<mode>): Likewise.
2011 (neon_vshrn_n<mode>): Likewise.
2012 (neon_vqshrn_n<mode>): Likewise.
2013 (neon_vqshrun_n<mode>): Likewise.
2014 (neon_vshl_n<mode>): Likewise.
2015 (neon_vqshl_n<mode>): Likewise.
2016 (neon_vqshlu_n<mode>): Likewise.
2017 (neon_vshll_n<mode>): Likewise.
2018 (neon_vsra_n<mode>): Likewise.
2019 (neon_vsri_n<mode>): Likewise.
2020 (neon_vsli_n<mode>): Likewise.
2021 (neon_vtbl1v8qi): Likewise.
2022 (neon_vtbl2v8qi): Likewise.
2023 (neon_vtbl3v8qi): Likewise.
2024 (neon_vtbl4v8qi): Likewise.
2025 (neon_vtbx1v8qi): Likewise.
2026 (neon_vtbx2v8qi): Likewise.
2027 (neon_vtbx3v8qi): Likewise.
2028 (neon_vtbx4v8qi): Likewise.
2029 (neon_vtrn<mode>_internal): Likewise.
2030 (neon_vzip<mode>_internal): Likewise.
2031 (neon_vuzp<mode>_internal): Likewise.
2032 (neon_vld1<mode>): Likewise.
2033 (neon_vld1_lane<mode>): Likewise.
2034 (neon_vld1_dup<mode>): Likewise.
2035 (neon_vld1_dupv2di): Likewise.
2036 (neon_vst1<mode>): Likewise.
2037 (neon_vst1_lane<mode>): Likewise.
2038 (neon_vld2<mode>): Likewise.
2039 (neon_vld2_lane<mode>): Likewise.
2040 (neon_vld2_dup<mode>): Likewise.
2041 (neon_vst2<mode>): Likewise.
2042 (neon_vst2_lane<mode>): Likewise.
2043 (neon_vld3<mode>): Likewise.
2044 (neon_vld3qa<mode>): Likewise.
2045 (neon_vld3qb<mode>): Likewise.
2046 (neon_vld3_lane<mode>): Likewise.
2047 (neon_vld3_dup<mode>): Likewise.
2048 (neon_vst3<mode>): Likewise.
2049 (neon_vst3qa<mode>): Likewise.
2050 (neon_vst3qb<mode>): Likewise.
2051 (neon_vst3_lane<mode>): Likewise.
2052 (neon_vld4<mode>): Likewise.
2053 (neon_vld4qa<mode>): Likewise.
2054 (neon_vld4qb<mode>): Likewise.
2055 (neon_vld4_lane<mode>): Likewise.
2056 (neon_vld4_dup<mode>): Likewise.
2057 (neon_vst4<mode>): Likewise.
2058 (neon_vst4qa<mode>): Likewise.
2059 (neon_vst4qb<mode>): Likewise.
2060 (neon_vst4_lane<mode>): Likewise.
2061 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2062 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2063 (neon_vec_<US>mult_lo_<mode>): Likewise.
2064 (neon_vec_<US>mult_hi_<mode>): Likewise.
2065 (neon_vec_<US>shiftl_<mode>): Likewise.
2066 (neon_unpack<US>_<mode>): Likewise.
2067 (neon_vec_<US>mult_<mode>): Likewise.
2068 (vec_pack_trunc_<mode>): Likewise.
2069 (neon_vec_pack_trunk_<mode>): Likewise.
2070 (neon_vabd<mode>_2): Likewise.
2071 (neon_vabd<mode>_3): Likewise.
2072 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
2073 (thumb2_movsi_vfp): Likewise.
2074 (movdi_vfp): Likewise.
2075 (movdi_vfp_cortexa8): Likewise.
2076 (movhf_vfp_neon): Likewise.
2077 (movhf_vfp): Likewiwse.
2078 (movsf_vfp): Likewiwse.
2079 (thumb2_movsf_vfp): Likewiwse.
2080 (movdf_vfp): Likewise.
2081 (thumb2_movdf_vfp): Likewise.
2082 (movsfcc_vfp): Likewise.
2083 (thumb2_movsfcc_vfp): Likewise.
2084 (movdfcc_vfp): Likewise.
2085 (thumb2_movdfcc_vfp): Likewise.
2086 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
2087 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
2088 (v10_v2c): Likewise.
2089 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
2091 (cortex_a15_neon_int_2): Likewise.
2092 (cortex_a15_neon_int_3): Likewise.
2093 (cortex_a15_neon_int_4): Likewise.
2094 (cortex_a15_neon_int_5): Likewise.
2095 (cortex_a15_neon_vqneg_vqabs): Likewise.
2096 (cortex_a15_neon_vmov): Likewise.
2097 (cortex_a15_neon_vaba): Likewise.
2098 (cortex_a15_neon_vaba_qqq): Likewise.
2099 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2100 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2101 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
2102 scalar_64_32_long_scalar): Likewise.
2103 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2104 (cortex_a15_neon_mla_qqq_8_16): Likewise.
2105 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2106 lotype_qdd_64_32_long): Likewise.
2107 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2108 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2109 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
2110 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2111 (cortex_a15_neon_shift_1): Likewise.
2112 (cortex_a15_neon_shift_2): Likewise.
2113 (cortex_a15_neon_shift_3): Likewise.
2114 (cortex_a15_neon_vshl_ddd): Likewise.
2115 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2116 (cortex_a15_neon_vsra_vrsra): Likewise.
2117 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
2118 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
2119 (cortex_a15_neon_fp_vmul_ddd): Likewise.
2120 (cortex_a15_neon_fp_vmul_qqd): Likewise.
2121 (cortex_a15_neon_fp_vmla_ddd): Likewise.
2122 (cortex_a15_neon_fp_vmla_qqq): Likewise.
2123 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
2124 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
2125 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2126 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2127 (cortex_a15_neon_bp_simple): Likewise.
2128 (cortex_a15_neon_bp_2cycle): Likewise.
2129 (cortex_a15_neon_bp_3cycle): Likewise.
2130 (cortex_a15_neon_vld1_1_2_regs): Likewise.
2131 (cortex_a15_neon_vld1_3_4_regs): Likewise.
2132 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2133 (cortex_a15_neon_vld2_4_regs): Likewise.
2134 (cortex_a15_neon_vld3_vld4): Likewise.
2135 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2136 (cortex_a15_neon_vst1_3_4_regs): Likewise.
2137 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
2138 (cortex_a15_neon_vst3_vst4): Likewise.
2139 (cortex_a15_neon_vld1_vld2_lane): Likewise.
2140 (cortex_a15_neon_vld3_vld4_lane" 10
2141 (cortex_a15_neon_vst1_vst2_lane): Likewise.
2142 (cortex_a15_neon_vst3_vst4_lane): Likewise.
2143 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
2144 (cortex_a15_neon_ldm_2): Likewise.0
2145 (cortex_a15_neon_stm_2): Likewise.
2146 (cortex_a15_neon_mcr): Likewise.
2147 (cortex_a15_neon_mcr_2_mcrr): Likewise.
2148 (cortex_a15_neon_mrc): Likewise.
2149 (cortex_a15_neon_mrrc): Likewise.
2150 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
2152 (cortex_a15_alu_shift): Likewise.
2153 (cortex_a15_alu_shift_reg): Likewise.
2154 (cortex_a15_mult32): Likewise.
2155 (cortex_a15_mult64): Likewise.
2156 (cortex_a15_block): Likewise.
2157 (cortex_a15_branch): Likewise.
2158 (cortex_a15_load1): Likewise.
2159 (cortex_a15_load3): Likewise.
2160 (cortex_a15_store1): Likewise.
2161 (cortex_a15_store3): Likewise.
2162 (cortex_a15_call): Likewise.
2163 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
2164 (cortex_a5_f2r): Likewise.
2165 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
2167 (cortex_a53_f2r): Likewise.
2168 * config/arm/cortex-a7.md
2169 (cortex_a7_branch): Update for attribute change.
2170 (cortex_a7_call): Likewise.
2171 (cortex_a7_alu_imm): Likewise.
2172 (cortex_a7_alu_reg): Likewise.
2173 (cortex_a7_alu_shift): Likewise.
2174 (cortex_a7_mul): Likewise.
2175 (cortex_a7_load1): Likewise.
2176 (cortex_a7_store1): Likewise.
2177 (cortex_a7_load2): Likewise.
2178 (cortex_a7_store2): Likewise.
2179 (cortex_a7_load3): Likewise.
2180 (cortex_a7_store3): Likewise.
2181 (cortex_a7_load4): Likewise.
2182 (cortex_a7_store4): Likewise.
2183 (cortex_a7_fpalu): Likewise.
2184 (cortex_a7_fconst): Likewise.
2185 (cortex_a7_fpmuls): Likewise.
2186 (cortex_a7_neon_mul): Likewise.
2187 (cortex_a7_fpmacs): Likewise.
2188 (cortex_a7_neon_mla: Likewise.
2189 (cortex_a7_fpmuld: Likewise.
2190 (cortex_a7_fpmacd: Likewise.
2191 (cortex_a7_fpfmad: Likewise.
2192 (cortex_a7_fdivs: Likewise.
2193 (cortex_a7_fdivd: Likewise.
2194 (cortex_a7_r2f: Likewise.
2195 (cortex_a7_f2r: Likewise.
2196 (cortex_a7_f_flags: Likewise.
2197 (cortex_a7_f_loads: Likewise.
2198 (cortex_a7_f_loadd: Likewise.
2199 (cortex_a7_f_stores: Likewise.
2200 (cortex_a7_f_stored: Likewise.
2201 (cortex_a7_neon): Likewise.
2202 * config/arm/cortex-a8-neon.md
2203 (cortex_a8_neon_mrc): Update for attribute change.
2204 (cortex_a8_neon_mrrc): Likewise.
2205 (cortex_a8_neon_int_1): Likewise.
2206 (cortex_a8_neon_int_2): Likewise.
2207 (cortex_a8_neon_int_3): Likewise.
2208 (cortex_a8_neon_int_4): Likewise.
2209 (cortex_a8_neon_int_5): Likewise.
2210 (cortex_a8_neon_vqneg_vqabs): Likewise.
2211 (cortex_a8_neon_vmov): Likewise.
2212 (cortex_a8_neon_vaba): Likewise.
2213 (cortex_a8_neon_vaba_qqq): Likewise.
2214 (cortex_a8_neon_vsma): Likewise.
2215 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2216 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2217 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2218 long_scalar): Likewise.
2219 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2220 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2221 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2222 long_scalar_qdd_64_32_long): Likewise.
2223 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2224 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2225 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2226 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2227 (cortex_a8_neon_shift_1): Likewise.
2228 (cortex_a8_neon_shift_2): Likewise.
2229 (cortex_a8_neon_shift_3): Likewise.
2230 (cortex_a8_neon_vshl_ddd): Likewise.
2231 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2232 (cortex_a8_neon_vsra_vrsra): Likewise.
2233 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2234 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2235 (cortex_a8_neon_fp_vsum): Likewise.
2236 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2237 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2238 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2239 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2240 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2241 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2242 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2243 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2244 (cortex_a8_neon_bp_simple): Likewise.
2245 (cortex_a8_neon_bp_2cycle): Likewise.
2246 (cortex_a8_neon_bp_3cycle): Likewise.
2247 (cortex_a8_neon_ldr): Likewise.
2248 (cortex_a8_neon_str): Likewise.
2249 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2250 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2251 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2252 (cortex_a8_neon_vld2_4_regs): Likewise.
2253 (cortex_a8_neon_vld3_vld4): Likewise.
2254 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2255 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2256 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2257 (cortex_a8_neon_vst3_vst4): Likewise.
2258 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2259 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2260 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2261 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2262 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2263 (cortex_a8_neon_mcr): Likewise.
2264 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2265 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
2266 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
2268 (ca9_neon_mrrc): Likewise.
2269 (cortex_a9_neon_int_1): Likewise.
2270 (cortex_a9_neon_int_2): Likewise.
2271 (cortex_a9_neon_int_3): Likewise.
2272 (cortex_a9_neon_int_4): Likewise.
2273 (cortex_a9_neon_int_5): Likewise.
2274 (cortex_a9_neon_vqneg_vqabs): Likewise.
2275 (cortex_a9_neon_vmov): Likewise.
2276 (cortex_a9_neon_vaba): Likewise.
2277 (cortex_a9_neon_vaba_qqq): Likewise.
2278 (cortex_a9_neon_vsma): Likewise.
2279 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2280 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
2281 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2282 long_scalar): Likewise.
2283 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2284 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2285 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2286 long_scalar_qdd_64_32_long): Likewise.
2287 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2288 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2289 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2290 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2291 (cortex_a9_neon_shift_1): Likewise.
2292 (cortex_a9_neon_shift_2): Likewise.
2293 (cortex_a9_neon_shift_3): Likewise.
2294 (cortex_a9_neon_vshl_ddd): Likewise.
2295 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2296 (cortex_a9_neon_vsra_vrsra): Likewise.
2297 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2298 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2299 (cortex_a9_neon_fp_vsum): Likewise.
2300 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2301 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2302 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2303 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2304 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2305 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2306 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2307 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2308 (cortex_a9_neon_bp_simple): Likewise.
2309 (cortex_a9_neon_bp_2cycle): Likewise.
2310 (cortex_a9_neon_bp_3cycle): Likewise.
2311 (cortex_a9_neon_ldr): Likewise.
2312 (cortex_a9_neon_str): Likewise.
2313 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2314 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2315 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2316 (cortex_a9_neon_vld2_4_regs): Likewise.
2317 (cortex_a9_neon_vld3_vld4): Likewise.
2318 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2319 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2320 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2321 (cortex_a9_neon_vst3_vst4): Likewise.
2322 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2323 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2324 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2325 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2326 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2327 (cortex_a9_neon_mcr): Likewise.
2328 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2329 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2330 (cortex_a9_fps): Likewise.
2331 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
2333 (cortex_m4_fmuls): Likewise.
2334 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
2336 (cortex_r4_mrc): Likewise.
2337 * config/arm/iterators.md: Update comment referring to neon_type.
2338 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
2339 (iwmmxt_movsi_insn): Likewise.
2340 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
2342 (pj4_core_to_vfp): Likewise.
2343 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
2345 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
2346 (vfp_fstore): Likewise.
2347 * doc/md.texi: Change references to neon_type to refer to type.
2349 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2351 * tree.h (DECL_BUILT_IN): Fix typo in comment.
2353 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
2355 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
2358 2013-09-04 Easwaran Raman <eraman@google.com>
2361 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
2362 (build_and_add_sum): Use it.
2363 (appears_later_in_bb): Simplify code.
2365 2013-09-04 Teresa Johnson <tejohnson@google.com>
2367 * dumpfile.c (dump_finish): Don't close stderr/stdout.
2369 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
2371 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
2373 2013-09-04 Jan Hubicka <jh@suse.cz>
2375 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
2376 * ipa-devirt.c: Include diganostic.h
2377 (odr_type_d): Add types and types_set.
2378 (hash_type_name): Work for types with vtables during LTO.
2379 (odr_hasher::remove): Fix comment; destroy types_set.
2380 (add_type_duplicate): New function,
2381 (get_odr_type): Use it.
2382 (dump_type_inheritance_graph): Dump type duplicates.
2383 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
2385 * tree.c (types_same_for_odr): Give exact answers on types with
2388 2013-09-04 Dodji Seketeli <dodji@redhat.com>
2390 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
2391 explaining their differences.
2393 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
2395 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
2397 2013-09-03 Jeff Law <law@redhat.com>
2399 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
2400 when not threading through a joiner block. Pass joiner/no joiner
2401 state to register_jump_thread.
2402 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
2403 state from argument rather than implying on path length.
2404 Dump the entire jump thread path into debugging dump.
2405 * tree-flow.h (register_jump_thread): Update prototype.
2407 2013-08-29 Xinliang David Li <davidxl@google.com>
2409 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2410 Remove a trivial gcc_assert.
2412 2013-08-29 Xinliang David Li <davidxl@google.com>
2414 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
2415 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
2416 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2417 Delay base decl alignment adjustment.
2418 * tree-vectorizer.c (vect_destroy_datarefs): New function.
2419 * tree-vectorizer.h: New data structure.
2420 (set_dr_misalignment): New function.
2421 (dr_misalignment): Ditto.
2422 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
2423 (vectorizable_load): Ditto.
2424 (ensure_base_align): New function.
2425 (vectorize_loops): Add dbg_cnt support.
2426 (execute_vect_slp): Ditto.
2427 * dbgcnt.def: New debug counter.
2428 * Makefile: New dependency.
2430 2013-09-03 Meador Inge <meadori@codesourcery.com>
2434 2013-08-30 Meador Inge <meadori@codesourcery.com>
2436 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2438 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
2440 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
2441 function descriptor.
2443 2013-09-03 Richard Biener <rguenther@suse.de>
2445 * tree-affine.c (add_elt_to_tree): Fix association issue,
2446 avoid useless converts and make sure to always return a
2447 properly typed result.
2449 2013-09-03 Richard Biener <rguenther@suse.de>
2452 * fold-const.c (negate_expr_p): Fix division case.
2453 (negate_expr): Likewise.
2455 2013-09-03 Richard Biener <rguenther@suse.de>
2458 * tree-streamer-out.c: Include tm.h.
2459 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2461 2013-09-03 Jan Hubicka <jh@suse.cz>
2463 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2465 2013-09-03 Alan Modra <amodra@gmail.com>
2467 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2468 multilibs to big-endian target and vice versa.
2469 * config/rs6000/t-linux64: Use := assignment on all vars.
2470 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2471 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2472 * config/rs6000/t-linux64le: New file.
2473 * config/rs6000/t-linux64bele: New file.
2474 * config/rs6000/t-linux64lebe: New file.
2476 2013-09-02 Jan Hubicka <jh@suse.cz>
2478 * ipa-inline-transform.c (inline_transform): Do not
2479 optimize_inline_calls when not optimizing.
2481 2013-09-02 Jan Hubicka <jh@suse.cz>
2483 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2484 duplicated nodes for assembler names.
2485 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2486 hard registers from assembler name hash.
2488 2013-09-02 Jan Hubicka <jh@suse.cz>
2490 * ipa-split.c (execute_split_functions): Split externally visible
2491 functions called once.
2493 2013-09-02 Martin Jambor <mjambor@suse.cz>
2496 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2497 linked list. When finding the correct duplicate, also consider also
2498 the caller in additon to its inlined_to node.
2500 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2502 * config/aarch64/aarch64-simd-builtins.def
2503 (dup_lane_scalar): Remove.
2504 * config/aarch64/aarch64-simd.md
2505 (aarch64_simd_dup): Add 'w->w' alternative.
2506 (aarch64_dup_lane<mode>): Allow for VALL.
2507 (aarch64_dup_lane_scalar<mode>): Remove.
2508 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2509 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2510 (aarch64_get_lane_unsigned<mode>): Likewise.
2511 (aarch64_get_lane<mode>): Likewise.
2512 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2513 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2514 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2515 (VCON): Change container of V2SF.
2516 (vswap_width_name): Likewise.
2517 * config/aarch64/arm_neon.h
2518 (__aarch64_vdup_lane_any): New.
2519 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2520 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2521 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2523 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2526 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2527 the source or the destination is a single hard register.
2529 2013-09-02 Richard Biener <rguenther@suse.de>
2532 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2535 2013-09-02 Richard Biener <rguenther@suse.de>
2537 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2538 arithmetic to sizetype.
2540 2013-09-02 Bin Cheng <bin.cheng@arm.com>
2542 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2543 Find auto-increment use both before and after candidate.
2545 2013-09-02 Marek Polacek <polacek@redhat.com>
2547 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2549 2013-09-01 Jan Hubicka <jh@suse.cz>
2551 * Makefile.in: Add ipa-profile.o
2552 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2553 * cgraph.c (struct cgraph_propagate_frequency_data,
2554 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2555 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2556 * cgraph.h (cgraph_propagate_frequency): Remove.
2557 * ipa-inline-analysis.c: Include ipa-utils.h;
2558 drop duplicated cfgloop.h.
2559 (inline_update_callee_summaries): Update.
2560 * ipa-profile.c: New file.
2561 * ipa-utils.h (ipa_propagate_frequency): Declare.
2562 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
2563 data-streamer.h, value-prof.h.
2564 (symtab_remove_unreachable_nodes): Update profile.
2565 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2566 account_time_size, cmp_counts, dump_histogram,
2567 ipa_profile_generate_summary, ipa_profile_write_summary,
2568 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2569 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
2570 Move to ipa-profile.c.
2572 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
2574 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2576 2013-09-01 Jan Hubicka <jh@suse.cz>
2578 * common.opt (fdevirtualize-speculatively): New function.
2579 * invoke.texi (fdevirtualize-speculatively): Document.
2580 * ipa-devirt.c: Include ipa-inline.h
2581 (likely_target_p): New function.
2582 (ipa_devirt): New function.
2583 (gate_ipa_devirt): New function.
2584 (pass_data_ipa_devirt): New static var.
2585 (pass_ipa_devirt): Likewise.
2586 (make_pass_ipa_devirt): New function.
2587 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2588 (common_handle_option): Disable devirtualization when
2589 value range profiling is available.
2590 * passes.def (pass_ipa_devirt): Add.
2591 * timever.def (TV_IPA_DEVIRT): New timevar.
2592 * tree-pass.h (make_pass_ipa_devirt):
2594 2013-09-01 Iain Sandoe <iain@codesourcery.com>
2596 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2597 include sanitize(undefined).
2599 2013-08-31 Diego Novillo <dnovillo@google.com>
2601 * Makefile.in (TREE_CORE_H): Define.
2603 (GTFILES): Add tree-core.h.
2604 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2606 * tree-core.h: New file.
2607 Move all data structures, enum, typedefs, global
2608 declarations and constants from ...
2611 2013-08-31 Jan Hubicka <jh@suse.cz>
2613 * bulitins.c (expand_builtin): Do not early exit for gcov
2614 instrumented functions.
2616 2013-08-31 Marek Polacek <polacek@redhat.com>
2618 * ubsan.c: Include tm_p.h.
2620 2013-08-31 Jan Hubicka <jh@suse.cz>
2622 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2625 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2626 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
2627 * tree-cfg.c (verify_gimple_label): ... here.
2628 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2629 (ipa_merge_profiles): New function.
2630 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2631 (lto_input_function_body): Likewise.
2632 * ipa-utils.h (ipa_merge_profiles): Declare.
2633 * lto-streamer.h (lto_input_function_body): Update prototype.
2634 (emit_label_in_global_context_p): Remove.
2635 * lto-symtab.c: Include ipa-utils.h
2636 (lto_cgraph_replace_node): Use ipa_merge_profiles.
2638 2013-08-31 Jan Hubicka <jh@suse.cz>
2640 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2642 2013-08-31 Jan Hubicka <jh@suse.cz>
2644 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2646 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
2648 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2649 "cmp" RTX before signed_comparison_operator check to account
2652 2013-08-30 Jan Hubicka <jh@suse.cz>
2654 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
2655 (detect_type_change_1): Rename to ...
2656 (detect_type_change): ... this one; early return on non-polymorphic
2658 (detect_type_change_ssa): Add comp_type parameter; update
2659 use of detect_type_change.
2660 (compute_complex_assign_jump_func): Add param_type parameter;
2661 update use of detect_type_change_ssa.
2662 (compute_complex_ancestor_jump_func): Likewise.
2663 (ipa_get_callee_param_type): New function.
2664 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2665 update calls to the jump function computation functions.
2667 2013-08-30 Teresa Johnson <tejohnson@google.com>
2668 Steven Bosscher <steven@gcc.gnu.org>
2670 * cfgrtl.c (fixup_new_cold_bb): New routine.
2671 (commit_edge_insertions): Invoke fixup_partitions.
2672 (find_partition_fixes): New routine.
2673 (fixup_partitions): Ditto.
2674 (verify_hot_cold_block_grouping): Update comments.
2675 (rtl_verify_edges): Invoke find_partition_fixes.
2676 (rtl_verify_bb_pointers): Update comments.
2677 (rtl_verify_bb_layout): Ditto.
2678 * basic-block.h (probably_never_executed_edge_p): Declare.
2679 (fixup_partitions): Ditto.
2680 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2681 * bb-reorder.c (sanitize_hot_paths): New function.
2682 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2684 * predict.c (probably_never_executed_edge_p): New routine.
2685 * cfg.c (check_bb_profile): Add partition insanity warnings.
2687 2013-08-30 Meador Inge <meadori@codesourcery.com>
2689 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2691 2013-08-30 Marek Polacek <polacek@redhat.com>
2693 * Makefile.in (ubsan.o): Add.
2694 (c-family/c-ubsan.o): Add.
2695 (builtins.o): Add ubsan.h dependency.
2696 * ubsan.h: New file.
2697 * ubsan.c: New file.
2698 * common.opt: Add -fsanitize=undefined option.
2699 (flag_sanitize): Add variable.
2700 (fsanitize=): Add option. Add Driver.
2701 (fsanitize=thread): Remove option.
2702 (fsanitize=address): Likewise.
2703 (static-libubsan): New option.
2704 * doc/invoke.texi: Document the new flag and -static-libubsan.
2705 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2706 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2707 * builtin-attrs.def (ATTR_COLD): Define.
2708 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2709 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2710 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2711 * flag-types.h (sanitize_code): New enum.
2712 * opts.c (common_handle_option): Parse command line arguments
2713 of -fsanitize=. Add -fsanitize=unreachable option.
2714 * varasm.c (get_variable_section): Adjust.
2715 (assemble_noswitch_variable): Likewise.
2716 (assemble_variable): Likewise.
2717 (output_constant_def_contents): Likewise.
2718 (categorize_decl_for_section): Likewise.
2719 (place_block_symbol): Likewise.
2720 (output_object_block): Likewise.
2721 * builtins.def: Likewise.
2722 * toplev.c (compile_file): Likewise.
2723 (process_options): Likewise.
2724 * cppbuiltin.c: Likewise.
2725 * tsan.c (tsan_pass): Likewise.
2726 (tsan_gate): Likewise.
2727 (tsan_gate_O0): Likewise.
2728 * cfgexpand.c (partition_stack_vars): Likewise.
2729 (expand_stack_vars): Likewise.
2730 (defer_stack_allocation): Likewise.
2731 (expand_used_vars): Likewise.
2732 * cfgcleanup.c (old_insns_match_p): Likewise.
2733 * asan.c (asan_finish_file): Likewise.
2734 (asan_instrument): Likewise.
2735 (gate_asan): Likewise.
2736 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2737 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2738 (asan_global_struct): Use pointer_sized_int_node instead
2739 calling build_nonstandard_integer_type.
2740 (initialize_sanitizer_builtins): Likewise.
2741 (asan_finish_file): Likewise.
2742 * gcc.c: Document %{%:function(args):X}.
2743 (static_spec_functions): Add sanitize.
2744 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2745 store funcval != NULL there.
2746 (do_spec_1): Adjust handle_spec_function caller.
2747 (handle_braces): Allow %:function(args) as condition.
2748 (sanitize_spec_function): New function.
2749 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2750 (LIBUBSAN_SPEC): Likewise.
2751 (LIBUBSAN_EARLY_SPEC): Likewise.
2752 (SANITIZER_SPEC): Handle libubsan.
2753 (SANITIZER_EARLY_SPEC): Likewise.
2754 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2755 instead of fsanitize=address.
2756 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2757 instead of fsanitize=address*.
2758 * builtins.c: Include ubsan.h.
2759 (fold_builtin_0): Instrument __builtin_unreachable.
2760 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2761 instead of flag_asan.
2762 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2763 (pointer_sized_int_node): Define.
2764 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
2766 2013-08-30 Mike Stump <mikestump@comcast.net>
2768 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2771 2013-08-29 Jan Hubicka <jh@suse.cz>
2773 * cgraph.c (cgraph_function_body_availability): Handle weakref
2775 * passes.def: Remove pass_fixup_cfg.
2776 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2777 track when we need to remove functions.
2778 (gate_ipa_inline): Execute inlining always; add comment why.
2779 (pass_data_ipa_inline): Remove TODO_remove_functions.
2780 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2781 do not produce summaries.
2782 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2783 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2784 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2785 constant pool and vtable.
2787 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2789 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2790 New arm_neon.h's internal macros to specify 64-bit constants.
2791 Avoid using stdint.h's macros.
2793 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2795 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2796 before applying REG_P.
2798 2013-08-30 Jakub Jelinek <jakub@redhat.com>
2800 PR tree-optimization/58277
2801 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2802 after seeing too many stmts with vdef in between dombb and current
2803 bb, invalidate everything.
2805 2013-08-30 Richard Biener <rguenther@suse.de>
2807 * fold-const.c (fold_single_bit_test): Fix overflow test.
2809 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2811 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2812 and which can live in a register, always retrieve the value on entry.
2813 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2814 passed by invisible reference specially.
2815 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2816 (vt_add_function_parameter): Correctly deal with a parameter passed by
2817 invisible reference.
2819 2013-08-30 Jan Hubicka <jh@suse.cz>
2821 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2823 2013-08-30 Richard Biener <rguenther@suse.de>
2825 PR tree-optimization/58228
2826 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2827 allow invariant loads in nested loop vectorization.
2829 2013-08-30 Richard Biener <rguenther@suse.de>
2831 PR tree-optimization/58223
2832 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2833 (has_anti_or_output_dependence): ... this and adjust to also
2834 look for output dependences.
2835 (mark_nodes_having_upstream_mem_writes): Adjust.
2836 (rdg_flag_uses): Likewise.
2838 2013-08-30 Richard Biener <rguenther@suse.de>
2840 PR tree-optimization/58010
2841 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2842 assert that we have a loop-closed PHI.
2844 2013-08-29 Jan Hubicka <jh@suse.cz>
2846 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2847 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2848 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2849 (lto_free_function_in_decl_state_for_node): New function.
2851 2013-08-29 Xinliang David Li <davidxl@google.com>
2853 * loop-unroll.c (report_unroll_peel): Minor message change.
2854 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2855 Emit alignment peeling message with default -fopt-info.
2856 (vect_loop_versioning): Emit loop version info message.
2857 * tree-vectorizer.c (vectorize_loops): Minor message change.
2858 (execute_vect_slp): Ditto.
2860 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2862 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2863 of the clone from the DECL_NAME of the original function.
2865 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2867 * passes.c (register_pass): Add overload.
2868 * tree-pass.h (register_pass): Forward declare it. Add comment.
2870 2013-08-29 Jan Hubicka <jh@suse.cz>
2872 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2873 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2874 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2875 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2876 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2877 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
2878 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
2879 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2881 2013-08-29 Teresa Johnson <tejohnson@google.com>
2883 * dumpfile.c (dump_loc): Output column number.
2884 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2885 * doc/invoke.texi: Document optall -fopt-info flag.
2886 * profile.c (read_profile_edge_counts): Use new dump framework.
2887 (compute_branch_probabilities): Ditto.
2888 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2889 when pass not in any opt group.
2890 * pass_manager.h (pass_manager::get_pass_profile): New method.
2891 * value-prof.c (check_counter): Use new dump framework.
2892 (check_ic_target): Ditto.
2893 * coverage.c (get_coverage_counts): Ditto.
2894 (coverage_init): Setup new dump framework.
2896 2013-08-29 Richard Biener <rguenther@suse.de>
2898 PR tree-optimization/58246
2899 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2900 handle the dominance check inside a basic-block.
2902 2013-08-29 Richard Biener <rguenther@suse.de>
2905 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2906 of default defs that appear in abnormal PHI nodes.
2908 2013-08-29 Richard Biener <rguenther@suse.de>
2910 PR tree-optimization/57685
2911 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2912 single-use operands to avoid exponential complexity.
2914 2013-08-28 Dehao Chen <dehao@google.com>
2916 * ipa-inline.c (edge_badness): Fix integer underflow.
2918 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
2920 * gtm-builtins.def (_ITM_free): Declare leaf.
2922 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2925 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2926 (*tls_local_dynamic_base_64_largepic): Likewise.
2927 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2928 Remove predicate from call operand.
2929 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2930 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2932 2013-08-28 Jeff Law <law@redhat.com>
2934 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2935 checks for the number of predecessors and successors allowed.
2936 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2937 which require copying a joiner block if there is a request which
2938 is a subpath that requires no joiner block copying.
2940 2013-08-28 Jan Hubicka <jh@suse.cz>
2942 * lto-streamer-out.c (DFS_write_tree_body): Drop
2943 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2944 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2945 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2946 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2947 Do not read DECL_ERROR_ISSUED.
2948 (unpack_ts_decl_with_vis_value_fields): Do not read
2950 (lto_input_ts_binfo_tree_pointers): Do not read
2951 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
2952 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2953 write DECL_ERROR_ISSUED..
2954 (pack_ts_decl_with_vis_value_fields): Do not write
2956 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
2957 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
2958 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2959 * tree.h (tree_decl_common): Update comment.
2960 (DECL_ERROR_ISSUED): Remove.
2962 2013-08-28 Jakub Jelinek <jakub@redhat.com>
2965 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2967 2013-08-28 Jan Hubicka <jh@suse.cz>
2969 * builtins.def (free): Declare leaf.
2971 2013-08-27 David Malcolm <dmalcolm@redhat.com>
2974 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2975 * configure: Regenerate.
2977 2013-08-27 Martin Jambor <mjambor@suse.cz>
2979 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2980 (ipa_ancestor_jf_data): Likewise.
2981 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2982 (ipa_get_jf_pass_through_type_preserved): New function.
2983 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2984 (ipa_get_jf_ancestor_type_preserved): New function.
2985 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
2986 (ipa_get_jf_ancestor_result): Likewise.
2987 (propagate_vals_accross_pass_through): Use
2988 ipa_get_jf_pass_through_result to do all the value mappings.
2989 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2990 type_preserved flag.
2991 (ipa_set_jf_cst_copy): New function.
2992 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2993 (ipa_set_jf_arith_pass_through): Likewise.
2994 (ipa_set_ancestor_jf): Likewise.
2995 (compute_complex_assign_jump_func): Set type_preserved instead of
2997 (ipa_compute_jump_functions_for_edge): Likewise.
2998 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2999 (update_jump_functions_after_inlining): Update type_preserved.
3000 Explicitely create jump functions when combining one with pass_through.
3001 (ipa_write_jump_function): Stream the type_preserved flags.
3002 (ipa_read_jump_function): Likewise.
3004 2013-08-27 Jakub Jelinek <jakub@redhat.com>
3005 Aldy Hernandez <aldyh@redhat.com>
3007 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
3008 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
3009 * function.h (struct function): Add has_force_vect_loops and
3011 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
3012 * gimple.c (gimple_build_omp_critical): Add KIND argument and
3014 * gimple.def: Update CLAUSES comments.
3015 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
3016 (gimple_build_omp_for): Add argument to prototype.
3017 (gimple_omp_for_kind): New.
3018 (gimple_omp_for_set_kind): New.
3019 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
3020 GOVD_DATA_SHARE_CLASS.
3021 (enum omp_region_type): Add ORT_SIMD.
3022 (gimple_add_tmp_var): Handle ORT_SIMD.
3023 (gimplify_var_or_parm_decl): Same.
3024 (is_gimple_stmt): Same.
3025 (omp_firstprivatize_variable): Same.
3026 (omp_add_variable): Only use splay_tree_insert if lookup failed.
3027 (omp_notice_variable): Handle ORT_SIMD.
3028 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
3029 (omp_check_private): Handle ORT_SIMD.
3030 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3032 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
3033 Handle OMP_CLAUSE_LASTPRIVATE.
3034 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
3036 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
3037 (gimplify_expr): Handle OMP_SIMD.
3038 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
3039 (expand_GOMP_SIMD_VF): New.
3040 (expand_GOMP_SIMD_LAST_LANE): New.
3041 * internal-fn.def (GOMP_SIMD_LANE): New.
3042 (GOMP_SIMD_VF): New.
3043 (GOMP_SIMD_LAST_LANE): New.
3044 * omp-low.c: Include target.h.
3045 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
3047 (check_omp_nesting_restrictions): Same.
3049 (lower_rec_simd_input_clauses): New.
3050 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
3052 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
3053 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
3054 (expand_omp_build_assign): New.
3055 (expand_omp_for_init_counts): New.
3056 (expand_omp_for_init_vars): New.
3057 (extract_omp_for_update_vars): New.
3058 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
3059 and rewrite accordingly.
3060 (expand_omp_simd): New.
3061 (expand_omp_for): Use expand_omp_simd.
3062 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
3063 (lower_omp_for): Do not lower the body.
3064 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
3066 * tree-flow.h (find_omp_clause): Remove prototype.
3067 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
3068 forcing vectorization of the loop, or if flag_tree_vectorize.
3069 (gate_tree_if_conversion): Similarly.
3070 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
3071 gimple_build_omp_for.
3072 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
3073 * tree-parloops (create_parallel_loop): Pass kind argument to
3074 gimple_build_omp_for.
3075 * tree-pretty-print.c (dump_omp_clause): Add cases for
3076 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
3077 OMP_CLAUSE__SIMDUID_.
3078 (dump_generic_node): Handle OMP_SIMD.
3079 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
3080 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
3081 unroll OMP_SIMD loops here.
3082 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
3083 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
3085 (vect_analyze_data_refs): Handle simd loops.
3086 * tree-vect-loop.c (vectorizable_live_operation): Handle
3088 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
3089 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
3090 (vectorizable_load): Same.
3091 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
3092 (struct simduid_to_vf): New.
3093 (simduid_to_vf::hash): New.
3094 (simduid_to-vf::equal): New.
3095 (struct simd_array_to_simduid): New.
3096 (simd_array_to_simduid::hash): New.
3097 (simd_array_to_simduid::equal): New.
3098 (adjust_simduid_builtins): New.
3099 (struct note_simd_array_uses_struct): New.
3100 (note_simd_array_uses_cb): New.
3101 (note_simd_array_uses): New.
3102 (vectorize_loops): Handle simd hints and adjust simd builtins
3104 * tree-vectorizer.h (struct _stmt_vec_info): Add
3105 simd_lane_access_p field.
3106 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
3107 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
3108 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
3109 (omp_clause_code_name): Same.
3110 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
3111 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
3112 * tree.def (OMP_SIMD): New entry.
3113 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
3114 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
3115 (OMP_CLAUSE_DECL): Adjust range for new clauses.
3116 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
3117 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
3118 (OMP_CLAUSE_LINEAR_STEP): New.
3119 (OMP_CLAUSE_SAFELEN_EXPR): New.
3120 (OMP_CLAUSE__SIMDUID__DECL): New.
3121 (find_omp_clause): New prototype.
3123 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
3125 * config/i386/driver-i386.c (host_detect_local_cpu): Update
3126 Haswell processor detection.
3128 2013-08-27 Christian Widmer <shadow@umbrox.de>
3131 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
3132 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
3133 AVX2 capable processors.
3135 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
3137 * config/aarch64/arm_neon.h: Replace all inline asm implementations
3138 of vget_low_* with implementations in terms of other intrinsics.
3140 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
3143 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
3144 values to -1, 0 and 1.
3146 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
3148 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
3149 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
3150 (<optab><mode>3_insn): Remove.
3151 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
3152 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
3153 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
3154 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
3155 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
3156 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
3157 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
3158 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
3159 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
3160 (ror<mode>3_insn): Likewise.
3161 * config/aarch64/predicates.md (aarch64_simd_register): New.
3163 2013-08-27 Richard Biener <rguenther@suse.de>
3165 PR tree-optimization/57521
3166 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
3167 one edge is non-critical.
3168 (find_phi_replacement_condition): Make sure to use a non-critical
3169 edge. Cleanup and remove old bug workarounds.
3170 (bb_postdominates_preds): Remove.
3171 (if_convertible_loop_p_1): Do not compute post-dominators.
3172 (combine_blocks): Do not free post-dominators.
3173 (main_tree_if_conversion): Likewise.
3174 (pass_data_if_conversion): Add TODO_verify_ssa.
3176 2013-08-27 DJ Delorie <dj@redhat.com>
3178 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
3180 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
3182 * function.c (assign_parm_find_data_types): Set passed_mode and
3183 nominal_mode to the TYPE_MODE of nominal_type for the built
3184 pointer type in case of the struct-pass-by-reference.
3186 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3188 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
3189 (UINT16_TYPE): Change default to "unsigned int".
3191 * config/avr/avr.opt (mfract-convert-truncate): New option.
3192 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
3193 is set, round negative fractional integers according to n1169
3194 when converting to integer types.
3196 2013-08-26 Jan Hubicka <jh@suse.cz>
3198 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
3199 methods can not be called indirectly when their address is not taken.
3201 2013-08-26 Jan Hubicka <jh@suse.cz>
3203 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
3206 2013-08-26 Jan Hubicka <jh@suse.cz>
3208 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
3211 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3213 * reload.c (find_valid_class): Allow classes that do not include
3214 FIRST_PSEUDO_REGISTER - 1.
3216 2013-08-26 Jan Hubicka <jh@suse.cz>
3218 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
3219 fix edge count/frequency when speculation failed; fix type check
3220 for the direct call.
3222 2013-08-26 Jan Hubicka <jh@suse.cz>
3224 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
3226 2013-08-26 Jan Hubicka <jh@suse.cz>
3228 * ipa-inline-transform.c (inline_transform): Be ready for basic block
3229 to be changed by edge redirection.
3231 2013-08-26 Jan Hubicka <jh@suse.cz>
3233 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
3234 formating; add sanity check.
3235 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
3236 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
3237 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
3238 (resolve_noninline_speculation): Update callee keys, too.
3240 2013-08-26 Jan Hubicka <jh@suse.cz>
3242 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
3243 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
3245 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
3247 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
3250 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
3252 * config/i386/i386.c (ix86_debug_options): Remove prototype.
3253 (x86_64_elf_select_section): Ditto.
3254 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
3256 (ix86_pass_by_reference): Ditto.
3257 (output_set_got): Ditto.
3258 (ix86_unary_operator_ok): Ditto.
3259 (ix86_expand_builtin): Ditto.
3261 2013-08-23 Jan Hubicka <jh@suse.cz>
3263 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
3265 2013-08-23 Jan Hubicka <jh@suse.cz>
3267 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
3268 (tree_decl_with_vis): Add FINAL field.
3270 2013-08-23 Jeff Law <law@redhat.com>
3272 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
3273 the debugging dump when the expression is fully redundant.
3275 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3277 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
3278 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
3279 * pretty-print.c (pp_formatted_text_data): Likewise.
3280 (pp_write_text_to_stream): Likewise.
3281 (pp_write_text_as_dot_label_to_stream): Likewise.
3282 (pp_append_r): Likewise.
3283 (pp_format): Likewise.
3284 (pp_flush): Likewise.
3285 (pp_clear_output_area): Likewise.
3286 (pp_append_text): Likewise.
3287 (pp_formatted_text): Likewise.
3288 (pp_remaining_character_count_for_line): Likewise.
3289 (pp_newline): Likewise.
3290 (pp_character): Likewise.
3291 (output_buffer::~output_buffer): Define.
3292 (pretty_printer::~pretty_printer): Destruct output buffer.
3293 * pretty-print.h (output_buffer::~output_buffer): Declare.
3294 (pretty_printer::~pretty_printer): Declare virtual.
3296 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
3299 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
3300 HOST_WIDE_INT_M1U): New macros.
3301 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
3302 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
3303 unsigned -1 for lshift.
3304 * cse.c (cse_insn): Likewise.
3305 * double-int.c (rshift_double, lshift_double): Likewise.
3306 * builtins.c (fold_builtin_bitop): Likewise.
3307 * combine.c (force_to_mode): Likewise.
3308 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
3309 * simplify-rtx.c (simplify_const_unary_operation,
3310 simplify_const_binary_operation): Likewise.
3311 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
3312 check_va_list_escapes): Likewise.
3313 * rtlanal.c (nonzero_bits1): Likewise.
3314 * expmed.c (expand_smod_pow2): Likewise.
3315 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
3317 2013-08-23 Jan Hubicka <jh@suse.cz>
3319 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
3320 as having address taken.
3322 2013-08-23 Jan Hubicka <jh@suse.cz>
3324 * ipa-utils.h (method_class_type): Declare.
3325 * ipa-devirt.c (method_class_type): Export.
3327 * cgraphunit.c (analyze_functions): Do basic devirtualization;
3328 do not walk base classes of anonymous types.
3330 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
3332 PR rtl-optimization/58220
3334 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
3335 handle SEQUENCE insns properly.
3337 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3339 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
3341 (pp_newline_and_indent): Likewise.
3342 (pp_separate_with): Likewise.
3343 * pretty-print.c (pp_newline_and_flush): Define.
3344 (pp_newline_and_indent): Likewise.
3345 (pp_separate_with): Likewise.
3347 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3350 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
3351 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
3353 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
3355 * config/i386/predicates.md (ext_sse_reg_operand): New.
3356 * config/i386/i386.md (*movti_internal): Use
3357 predicate to determine if EVEX is needed.
3358 (*movsi_internal): Ditto.
3359 (*movdf_internal): Ditto.
3360 (*movsf_internal): Ditto.
3361 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
3363 2013-08-23 Jakub Jelinek <jakub@redhat.com>
3365 PR tree-optimization/58209
3366 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
3367 (find_tail_calls): Give up for pointer result types if m is non-NULL.
3368 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
3369 emit POINTER_PLUS_EXPR.
3370 (create_tailcall_accumulator): For pointer result type accumulate in
3373 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
3375 * configure.ac: Add backslashes missing from the last change.
3376 * configure: Regenerate.
3378 2013-08-22 Jan Hubicka <jh@susue.cz>
3380 * ipa.c (function_and_variable_visibility): First remember function
3381 was global and then make it local.
3383 2013-08-22 Julian Brown <julian@codesourcery.com>
3385 * configure.ac: Add aarch64 to list of arches which use "nop" in
3387 * configure: Regenerate.
3389 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3391 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
3392 gnu_libc_has_function.
3393 * config/s390/tpf.h: Likewise.
3395 2013-08-22 Jan Hubicka <jh@susue.cz>
3397 * timevar.c (validate_phases): Add cast.
3399 2013-08-22 Jan Hubicka <jh@susue.cz>
3401 * timevar.c (validate_phases): Use size_t for memory.
3402 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
3404 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
3406 * pretty-print.h (output_buffer::output_buffer): Declare.
3407 (pretty_printer::pretty_printer): Likewise.
3408 (pp_construct): Remove.
3409 * pretty-print.c (output_buffer::output_buffer): Define.
3410 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
3411 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
3412 (print_gimple_expr): Likewise.
3413 (print_gimple_seq): Likewise.
3414 (gimple_dump_bb): Likewise.
3415 * sched-vis.c (dump_value_slim): Likewise.
3416 (dump_insn_slim): Likewise.
3417 (dump_rtl_slim): Likewise.
3418 (str_pattern_slim): Likewise.
3419 * tree-mudflap.c (mf_varname_tree): Likewise.
3420 * graph.c (print_graph_cfg): Likewise.
3421 (start_graph_dump): Likewise.
3422 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
3424 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
3425 pretty printer initialization.
3426 * coretypes.h (diagnostic_context): Remove superflous type alias
3428 (pretty_printer): Likewise. Declare directly as a class.
3429 (pretty_print_info): Remove declaration as class.
3430 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
3431 and pp_clear_output_area.
3432 (asan_add_global): Likewise.
3434 2013-08-22 Jan Hubicka <jh@suse.cz>
3436 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
3437 * ipa-utils.h (update_type_inheritance_graph): Declare.
3438 (possible_polymorphic_call_target_p): Declare.
3439 (possible_polymorphic_call_target_p): New.
3440 * ipa-devirt.c: Update toplevel comments.
3441 (cached_polymorphic_call_targets): Move up.
3442 (odr_type_d): Move ID down.
3443 (polymorphic_type_binfo_p): Update comment.
3444 (odr_hasher::remove): Likewise;
3445 (get_odr_type): Set anonymous_namespace.
3446 (dump_odr_type): Dump it.
3447 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
3448 (maybe_record_node): Record node in cached_polymorphic_call_targets.
3449 (record_binfo): Add comment.
3450 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
3452 (devirt_node_removal_hook): Do not iCE when cache is freed.
3453 (possible_polymorphic_call_target_p): New predicate.
3454 (update_type_inheritance_graph): New function.
3456 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3457 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3458 Sergey Lega <sergey.s.lega@intel.com>
3459 Anna Tikhonova <anna.tikhonova@intel.com>
3460 Ilya Tocar <ilya.tocar@intel.com>
3461 Andrey Turetskiy <andrey.turetskiy@intel.com>
3462 Ilya Verbin <ilya.verbin@intel.com>
3463 Kirill Yukhin <kirill.yukhin@intel.com>
3464 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3466 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3467 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3468 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3469 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3470 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3471 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3472 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3473 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3474 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3475 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3476 OPT_mavx512pf, OPT_mavx512er cases.
3477 * config/i386/constraints.md (v): New constraint.
3478 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3479 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3480 (bit_AVX512CD): New.
3481 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3482 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3483 * config/i386/i386-c.c (ix86_target_macros_internal):
3484 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3486 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3487 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3488 * config/i386/i386.c (regclass_map, dbx_register_map)
3489 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3490 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3491 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3493 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3494 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3495 -mavx512pf options. Fix formatting.
3496 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3497 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3498 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3499 -mavx512cd, -mavx512pf options.
3500 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3501 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3502 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3504 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3505 (ix86_set_reg_reg_cost): Ditto.
3506 (x86_order_regs_for_local_alloc): Ditto.
3507 (MAX_VECT_LEN): Extend to 64-byte.
3508 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3509 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3510 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3511 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3512 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3513 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3514 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3515 (SSE_REG_MODE_P): Support new modes.
3516 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3517 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3518 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3519 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3520 (REG_CLASS_CONTENTS): Add new registers.
3521 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3522 (EXT_REX_SSE_REGNO_P): New.
3523 (HI_REGISTER_NAMES): Add new registers.
3524 * config/i386/i386.md: Define constants for new registers.
3525 (mode): Add new 512-bit modes.
3526 (prefix): Support evex prefix.
3527 (isa): Support avx512f, noavx512f, fma_avx512f.
3528 (ssemodesuffix): Add new 512-bit modes.
3530 (*movxi_internal_avx512f): Ditto.
3531 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3533 (*movsi_internal): Likewise.
3534 (*movdf_internal): Likewise.
3535 (*movsf_internal): Likewise.
3536 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3537 (<code><mode>3): Likewise.
3538 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3540 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3541 with the new constraint "v".
3542 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3544 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3545 with the new constraint "v".
3546 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3547 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3548 (*<plusminus_insn><mode>3): Likewise.
3549 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3550 (*mul<mode>3): Likewise.
3551 (<sse>_vmmul<mode>3): Likewise.
3552 (<sse>_div<mode>3): Likewise.
3553 (<sse>_vmdiv<mode>3): Likewise.
3554 (<sse>_sqrt<mode>2): Likewise.
3555 (<sse>_vmsqrt<mode>2): Likewise.
3556 (*<code><mode>3_finite): Likewise.
3557 (*<code><mode>3) <smaxmin>: Likewise.
3558 (<sse>_vm<code><mode>3): Likewise.
3559 (*<code><mode>3) <any_logic>: Likewise.
3560 (*fma_fmadd_<mode>): Likewise.
3561 (*fma_fmsub_<mode>): Likewise.
3562 (*fma_fnmadd_<mode>): Likewise.
3563 (*fma_fnmsub_<mode>): Likewise.
3564 (*fma_fmaddsub_<mode>): Likewise.
3565 (*fma_fmsubadd_<mode>): Likewise.
3566 (*fmai_fmadd_<mode>): Likewise.
3567 (*fmai_fmsub_<mode>): Likewise.
3568 (*fmai_fnmadd_<mode>): Likewise.
3569 (*fmai_fnmsub_<mode>): Likewise.
3570 (sse_cvtsi2ss): Likewise.
3571 (sse_cvtsi2ssq): Likewise.
3572 (sse_cvtss2si): Likewise.
3573 (sse_cvtss2si_2): Likewise.
3574 (sse_cvtss2siq): Likewise.
3575 (sse_cvtss2siq_2): Likewise.
3576 (sse_cvttss2si): Likewise.
3577 (sse_cvtss2siq_2): Likewise.
3578 (float<sseintvecmodelower><mode>2): Likewise.
3579 (sse2_cvtsd2si_2): Likewise.
3580 (sse2_cvtsd2siq_2): Likewise.
3581 (*<plusminus_insn><mode>3): Likewise.
3582 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3583 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3584 (ashr<mode>3): Likewise.
3585 (<shift_insn><mode>3): Likewise.
3586 (avx2_<code><mode>3): Likewise.
3587 (*avx2_<code><mode>3): Likewise.
3588 (*andnot<mode>3): Likewise.
3589 (*<code><mode>3) <any_logic>: Likewise.
3590 (abs<mode>2): Likewise.
3591 (avx2_permvar<mode>): Likewise.
3592 (avx2_perm<mode>_1): Likewise.
3593 (*avx_vpermilp<mode>): Likewise.
3594 (avx_vpermilvar<mode>3): Likewise.
3595 (avx2_ashrv<mode>): Likewise.
3596 (avx2_<shift_insn>v<mode>): Likewise.
3597 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3599 * doc/rtl.texi: Document XImode.
3601 2013-08-21 Jeff Law <law@redhat.com>
3603 * tree-flow.h (register_jump_thread): Pass vector of edges
3604 instead of each important edge.
3605 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3606 thread path into a vector and pass that to register_jump_thread.
3607 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3608 passed in edge vector to the current 3-edge form.
3611 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3613 * dce.c (fini_dce): Call df_analyze again just in case
3614 delete_unmarked_insns removed anything.
3616 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3618 * reload.h (struct reg_equivs): Rename to ..
3619 (struct reg_equivs_s): .. this.
3621 2013-08-20 Martin Liska <marxin.liska@gmail.com>
3623 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3625 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3627 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
3629 2013-08-21 Jeff Law <law@redhat.com>
3631 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3632 simplify assignments too. If the RHS collapses to a singleton
3633 range, then return the value for the range.
3635 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3637 * config/i386/sse.md (V16): Rename to...
3639 (mov<mode>): Update iterator name.
3640 (*mov<mode>_internal): Ditto.
3641 (push<mode>1): Ditto.
3642 (movmisalign<mode>): Ditto.
3644 2013-08-20 Jan Hubicka <jh@suse.cz>
3647 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3648 entry for direct edges.
3649 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3651 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3653 Revert my last two changes, r201865 and r201864:
3656 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3658 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3659 instances can own GC refs.
3661 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3662 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3663 (gcc::context::gt_pch_nx): Likewise.
3664 (gcc::context::gt_pch_nx): Likewise.
3665 * ggc.h (gt_ggc_mx <T>): New.
3666 (gt_pch_nx_with_op <T>): New.
3667 (gt_pch_nx <T>): New.
3668 * passes.c (opt_pass::gt_ggc_mx): New.
3669 (opt_pass::gt_pch_nx): New.
3670 (opt_pass::gt_pch_nx_with_op): New.
3671 (pass_manager::gt_ggc_mx): New.
3672 (pass_manager::gt_pch_nx): New.
3673 (pass_manager::gt_pch_nx_with_op): New.
3674 (pass_manager::operator new): Use
3675 ggc_internal_cleared_alloc_stat rather than xcalloc.
3676 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3677 (pass_manager::gt_ggc_mx): New.
3678 (pass_manager::gt_pch_nx): New.
3679 (pass_manager::gt_pch_nx_with_op): New.
3680 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3681 (opt_pass::operator new): New.
3682 (opt_pass::gt_ggc_mx): New.
3683 (opt_pass::gt_pch_nx): New.
3684 (opt_pass::gt_pch_nx_with_op): New.
3687 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3689 * Makefile.in (GTFILES): Add context.h.
3690 * context.c (gcc::context::operator new): New.
3691 (gcc::context::gt_ggc_mx): New.
3692 (gcc::context::gt_pch_nx): New.
3693 (gcc::context::gt_pch_nx): New.
3694 * context.h (gcc::context): Add GTY((user)) marking.
3695 (gcc::context::operator new): New.
3696 (gcc::context::gt_ggc_mx): New.
3697 (gcc::context::gt_pch_nx): New.
3698 (gcc::context::gt_pch_nx): New.
3699 (g): Add GTY marking.
3700 (gt_ggc_mx (gcc::context *)): New.
3701 (gt_pch_nx (gcc::context *)): New.
3702 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3703 void *cookie)): New.
3704 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3706 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3708 * dce.c (fini_dce): Call df_analyze again just in case
3709 delete_unmarked_insns removed anything.
3711 2013-08-20 Teresa Johnson <tejohnson@google.com>
3713 PR rtl-optimizations/57451
3714 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3715 from crossing split section boundaries.
3717 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3719 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3720 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3721 with MULTLIB_DEFAULTS.
3723 2013-08-20 Nick Clifton <nickc@redhat.com>
3725 * target.def (narrow_volatile_bitfield): Note that the default
3726 value is false, not !TARGET_STRICT_ALIGN.
3727 * doc/tm.texi: Regenerate.
3729 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3731 Fix LIB_SPEC for systems without libpthread.
3733 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3734 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3736 * config/i386/linux-common.h: Likewise.
3737 * config/mips/linux-common.h: Likewise.
3739 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
3741 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3744 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3746 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3747 instances can own GC refs.
3749 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3750 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3751 (gcc::context::gt_pch_nx): Likewise.
3752 (gcc::context::gt_pch_nx): Likewise.
3753 * ggc.h (gt_ggc_mx <T>): New.
3754 (gt_pch_nx_with_op <T>): New.
3755 (gt_pch_nx <T>): New.
3756 * passes.c (opt_pass::gt_ggc_mx): New.
3757 (opt_pass::gt_pch_nx): New.
3758 (opt_pass::gt_pch_nx_with_op): New.
3759 (pass_manager::gt_ggc_mx): New.
3760 (pass_manager::gt_pch_nx): New.
3761 (pass_manager::gt_pch_nx_with_op): New.
3762 (pass_manager::operator new): Use
3763 ggc_internal_cleared_alloc_stat rather than xcalloc.
3764 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3765 (pass_manager::gt_ggc_mx): New.
3766 (pass_manager::gt_pch_nx): New.
3767 (pass_manager::gt_pch_nx_with_op): New.
3768 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3769 (opt_pass::operator new): New.
3770 (opt_pass::gt_ggc_mx): New.
3771 (opt_pass::gt_pch_nx): New.
3772 (opt_pass::gt_pch_nx_with_op): New.
3774 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3776 * Makefile.in (GTFILES): Add context.h.
3777 * context.c (gcc::context::operator new): New.
3778 (gcc::context::gt_ggc_mx): New.
3779 (gcc::context::gt_pch_nx): New.
3780 (gcc::context::gt_pch_nx): New.
3781 * context.h (gcc::context): Add GTY((user)) marking.
3782 (gcc::context::operator new): New.
3783 (gcc::context::gt_ggc_mx): New.
3784 (gcc::context::gt_pch_nx): New.
3785 (gcc::context::gt_pch_nx): New.
3786 (g): Add GTY marking.
3787 (gt_ggc_mx (gcc::context *)): New.
3788 (gt_pch_nx (gcc::context *)): New.
3789 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3790 void *cookie)): New.
3791 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3793 2013-08-20 Alan Modra <amodra@gmail.com>
3796 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3797 (rs6000_emit_epilogue): Likewise.
3799 2013-08-19 Dehao Chen <dehao@google.com>
3801 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3803 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3804 Jakub Jelinek <jakub@redhat.com>
3806 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3807 (BUILT_IN_FABSD64): Likewise.
3808 (BUILT_IN_FABSD128): Likewise.
3809 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3810 (fold_builtin_1): Likewise.
3811 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3812 destination and source operands.
3813 (*abstd2_fpr): Likewise.
3814 (*nabstd2_fpr): Likewise.
3816 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3818 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3821 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
3823 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3826 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3828 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3829 * builtins.c (default_libc_has_function): New.
3830 (gnu_libc_has_function): Ditto.
3831 (no_c99_libc_has_function): Ditto.
3832 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3833 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3834 (fold_builtin_sincos): Likewise.
3835 (fold_builtin_cexp): Likewise.
3836 * builtins.def (DEF_C94_BUILTIN): Likewise.
3837 (DEF_C99_BUILTIN): Likewise.
3838 (DEF_C99_C90RES_BUILTIN): Likewise.
3839 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3840 definitions to using this define.
3841 * config/darwin-protos.h (darwin_libc_has_function): New.
3842 * config/darwin.c (darwin_libc_has_function): Ditto.
3843 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3844 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3845 * config/darwin.h: Ditto.
3846 * config/elfos.h: Ditto.
3847 * config/freebsd.h: Ditto.
3848 * config/i386/cygming.h: Ditto.
3849 * config/i386/djgpp.h: Ditto.
3850 * config/i386/i386-interix.h: Ditto.
3851 * config/microblaze/microblaze.h: Ditto.
3852 * config/mmix/mmix.h: Ditto.
3853 * config/gnu-user.h: Ditto.
3854 * config/ia64/hpux.h: Ditto.
3855 * config/pa/pa-hpux.h: Ditto.
3856 * config/pdp11/pdp11.h: Ditto.
3857 * config/picochip/picochip.h: Ditto.
3858 * config/linux.h: Ditto.
3859 * config/netbsd.h: Ditto.
3860 * config/openbsd.h: Ditto.
3861 * config/rs6000/aix43.h: Ditto.
3862 * config/rs6000/aix51.h: Ditto.
3863 * config/rs6000/aix52.h: Ditto.
3864 * config/rs6000/aix53.h: Ditto.
3865 * config/rs6000/aix61.h: Ditto.
3866 * config/rs6000/darwin.h: Ditto.
3867 * config/rs6000/linux.h: Ditto.
3868 * config/rs6000/linux64.h: Ditto.
3869 * config/s390/tpf.h: Ditto.
3870 * config/sol2-10.h: Ditto.
3871 * config/sol2.h: Ditto.
3872 * config/vms/vms.h: Ditto.
3873 * config/vxworks.h: Ditto.
3874 * config/linux-android.c (linux_android_libc_has_function):
3875 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3876 * config/linux-protos.h (linux_android_libc_has_function):
3878 * config/i386/i386.c (ix86_libc_has_function): New.
3879 * config/i386/i386-protos.h
3880 (ix86_libc_has_function): New declaration.
3881 * config/i386/i386.md
3882 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3883 ("isinf<mode>2): Likewise.
3884 * convert.c (convert_to_integer): Using new target hook
3885 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3886 TARGET_C99_FUNCTIONS.
3887 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3888 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3889 * coretypes.h (function_class): New enum for different
3890 classes of functions.
3891 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3892 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3893 (TARGET_HAS_SINCOS): Likewise.
3894 (TARGET_LIBC_HAS_FUNCTION): New.
3895 * doc/tm.texi: Regenerated.
3896 * targhooks.h (default_libc_has_function): New declaration.
3897 (no_c99_libc_has_function): Ditto.
3898 (gnu_libc_has_function): Ditto.
3899 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3900 and TARGET_HAS_SINCOS.
3902 2013-08-18 Jan Hubicka <jh@suse.cz>
3904 * Makeifle-in (ipa-devirt.o): New.
3905 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3906 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3907 (analyze_functions): Look into possible targets of polymorphic call.
3908 * dumpfile.c (dump_files): Add type-inheritance dump.
3909 * dumpfile.h (TDI_inheritance): New.
3910 * ipa-devirt.c: New file.
3911 * ipa-utils.h (odr_type_d): Forward declare.
3912 (odr_type): New type.
3913 (build_type_inheritance_graph): Declare.
3914 (possible_polymorphic_call_targets): Declare and introduce inline
3915 variant when only edge is pased.
3916 (dump_possible_polymorphic_call_targets): Likewise.
3917 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3918 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3919 (types_same_for_odr): ... here.
3920 * tree.h (type_in_anonymous_namespace_p): Declare.
3922 2013-08-18 Jakub Jelinek <jakub@redhat.com>
3924 PR tree-optimization/58006
3925 * tree-parloops.c (take_address_of): Don't ICE if get_name
3927 (eliminate_local_variables_stmt): Remove clobber stmts.
3929 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3931 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3932 error message is issued for an alias to undefined symbol.
3934 2013-08-18 Jan Hubicka <jh@suse.cz>
3936 * cgraph.c (cgraph_create_indirect_edge): Discover
3937 polymorphic calls and record basic info into indirect_info.
3938 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3939 devirtualization, ignore objc function calls.
3940 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3941 call with no parm index info.
3942 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3943 * tree.c (virtual_method_call_p): New function.
3944 * tree.h (virtual_method_call_p): Declare.
3946 2013-08-16 Jan Hubicka <jh@suse.cz>
3949 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3951 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
3953 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3956 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3958 * gengtype.c (type_for_name): Add special-case support for
3959 locating types within the "gcc::" namespace.
3960 (open_base_files): Emit a "using namespace gcc" directive.
3962 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3965 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3966 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3968 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3969 array instead of each individual operand as a separate argument.
3970 (emit_fusion_gpr_load): Likewise.
3971 (expand_fusion_gpr_load): Add new function declaration.
3973 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3974 signature to have the operands passed as an array, instead of as
3975 separate arguments. Allow ZERO_EXTEND to be in the memory
3976 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3977 depend on the register live/dead flags when peepholes are run.
3978 (expand_fusion_gpr_load): New function to be called from the
3979 peephole2 pass, to change the register that addis sets to be the
3981 (emit_fusion_gpr_load): Change the calling signature to have the
3982 operands passed as an array, instead of as separate arguments.
3983 Allow ZERO_EXTEND to be in the memory address, and also
3984 SIGN_EXTEND if -mpower8-fusion-sign.
3986 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3988 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3989 adjust the register addis loads up in the peephole2 pass. Do not
3990 depend on the register live/dead state when the peephole pass is done.
3992 2013-08-16 David Malcolm <dmalcolm@redhat.com>
3994 * gengtype.c (create_user_defined_type): Ensure that the kind
3995 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3996 declaration is seen before the GTY((user)) marking.
3998 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
4001 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
4003 2013-08-16 Jan Hubicka <jh@suse.cz>
4005 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
4006 arugment expected_type.
4007 (gimple_fold_call): Use it.
4008 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
4009 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
4010 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
4011 (try_make_edge_direct_virtual_call): Likewise.
4012 * tree.c (obj_type_ref_class): New.
4013 * tree.h (obj_type_ref_class): Use it.
4015 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
4017 * sched-vis.c (rtl_slim_pp_initialized): Remove.
4018 (rtl_slim_pp): Likewise.
4019 (init_rtl_slim_pretty_print): Likewise.
4020 (dump_value_slim): Don't call it. Use local pretty printer.
4021 (dump_insn_slim): Likewise.
4022 (dump_rtl_slim): Likewise.
4023 (str_pattern_slim): Likewise.
4024 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
4027 2013-08-16 Jakub Jelinek <jakub@redhat.com>
4029 PR tree-optimization/58164
4030 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
4031 walk gimple_goto_dest of GIMPLE_GOTO.
4033 PR tree-optimization/58165
4034 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
4035 bi_call must be the last stmt in a bb, don't split_block, instead
4036 use fallthru edge from it and give up if there is none.
4037 Release conds vector when returning early.
4039 2013-08-14 Xinliang David Li <davidxl@google.com>
4041 * config/i386/i386.c (ix86_option_override_internal):
4042 Remove unused variable and field.
4044 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4047 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
4048 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
4049 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
4050 and Linux, correct BLKmode alignment when 128-bit alignment is
4051 required and compatibility flag is not set.
4052 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
4053 for zero-size arguments when compatibility flag is not set.
4055 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4057 PR tree-optimization/58145
4058 * tree-sra.c (build_ref_for_offset): If prev_base has
4059 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
4061 2013-08-14 Xinliang David Li <davidxl@google.com>
4063 * config/i386/i386.c (ix86_option_override_internal):
4064 Fix uninitialized variable error.
4066 2013-08-14 Xinliang David Li <davidxl@google.com>
4068 * config/i386/i386.opt: Define two new options.
4069 * config/i386/x86-tune.def: Add arch selector field in macros.
4070 * config/i386/i386.h: Adjust macro definition.
4071 * config/i386/i386.c (ix86_option_override_internal):
4073 (parse_mtune_ctrl_str): New function.
4074 (set_ix86_tune_features): New function.
4075 (ix86_function_specific_restore): Call the new helper function.
4077 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
4079 PR rtl-optimization/57662
4080 * sel-sched.c (code_motion_process_successors): When the current insn
4081 is removed after the recursive traversal, break from the loop.
4082 Add comments and debug printouts.
4084 2013-08-14 Jakub Jelinek <jakub@redhat.com>
4085 Alexandre Oliva <aoliva@redhat.com>
4088 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
4089 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
4090 there also UNSPEC_PLTOFF.
4092 2013-08-14 Marek Polacek <polacek@redhat.com>
4094 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
4095 NUM_CONDITIONS bit positions.
4097 2013-08-13 Cary Coutant <ccoutant@google.com>
4099 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
4100 (attr_checksum): Hash vector contents instead of pointer.
4101 (attr_checksum_ordered): Likewise.
4103 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
4105 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
4106 when Pmode != word_mode. Add length_address attribute.
4107 (sse3_monitor_<mode>): Merge from sse3_monitor and
4108 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
4109 Pmode != word_mode. Update insn length attribute.
4110 * config/i386/i386.c (ix86_option_override_internal): Update
4111 ix86_gen_monitor selection for merged sse3_monitor insn.
4113 2013-08-13 Julian Brown <julian@codesourcery.com>
4115 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
4116 perform invalid legitimization on greater-than-word-size modes for
4119 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
4121 * ira.c (setup_class_translate_array): Use aclass instead of cl
4122 for classes not fully covered by allocno classes.
4124 2013-08-13 Jakub Jelinek <jakub@redhat.com>
4126 PR tree-optimization/57661
4127 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
4128 * tree-inline.c (tree_function_versioning): Initialize it.
4129 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
4130 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
4131 that is not being copied.
4134 * asan.c (instrument_strlen_call): Fix typo in comment.
4135 Use char * type even for the lhs of POINTER_PLUS_EXPR.
4137 2013-08-13 Steve Ellcey <sellcey@mips.com>
4139 * config/mips/mips.md (prefetch): Use lw instead of ld on
4140 loongson in 32bit mode.
4142 2013-08-13 Nick Clifton <nickc@redhat.com>
4144 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
4146 2013-08-13 Jan Hubicka <jh@suse.cz>
4148 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
4149 introduced edge; fix typo in sanity check.
4150 (cgraph_resolve_speculation): Export; improve diagnostic.
4151 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
4152 speculation at type mismatch.
4153 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
4154 (cgraph_resolve_speculation): Declare.
4155 (symtab_can_be_discarded): New function.
4156 * value-prof.c (gimple_ic_transform): Remove actual transform code.
4157 * ipa-inline-transform.c (speculation_removed): New global var.
4158 (clone_inlined_nodes): See if speculation can be removed.
4159 (inline_call): If speculations was removed, we growths may not match.
4160 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
4161 (speculation_useful_p): New function.
4162 (resolve_noninline_speculation): New function.
4163 (inline_small_functions): Resolve useless speculations.
4164 * ipa-inline.h (speculation_useful_p): Declare
4165 * ipa.c (can_replace_by_local_alias): Simplify.
4166 (ipa_profile): Produce speculative calls in non-lto, too;
4167 add simple cost model; produce local aliases.
4169 2013-08-13 David Malcolm <dmalcolm@redhat.com>
4171 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
4174 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
4176 * config/i386/i386.c (ix86_function_versions): Use error + inform.
4178 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
4180 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
4181 iterator instead of X87MODEF.
4183 2013-08-12 Perez Read <netfirewall@gmail.com>
4186 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
4187 operand 0 for intel asm alternative.
4188 (*movabs<mode>_2): Ditto for operand 1.
4190 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
4192 * config/aarch64/arm_none.h
4193 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
4195 2013-08-12 Nick Clifton <nickc@redhat.com>
4197 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
4199 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4201 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
4202 expand for QI/HImode operand to produce more effictive code for
4203 unsigned char(short) --> float(double) conversion.
4205 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
4207 * doc/invoke.texi: Mention that -ftls-model does not force the final
4210 2013-08-12 Marek Polacek <polacek@redhat.com>
4211 Marc Glisse <marc.glisse@inria.fr>
4213 PR tree-optimization/57980
4214 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
4215 when creating -1 constant.
4217 2013-08-10 Jan Hubicka <jh@suse.cz>
4219 Workaround binutils PR14342.
4220 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
4221 (gimple_init_edge_profiler): Likewise.
4222 (gimple_gen_ic_func_profiler): Likewise.
4224 2013-08-09 Jan Hubicka <jh@suse.cz>
4226 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
4228 2013-08-09 Xinliang David Li <davidxl@google.com>
4230 * config/i386/stringop.def: New file.
4231 * config/i386/stringop.opt: New file.
4232 * config/i386/i386-opts.h: Include stringopt.def.
4233 * config/i386/i386.opt: Include stringopt.opt.
4234 * config/i386/i386.c (ix86_option_override_internal):
4235 Override default size based stringop inline strategies with options.
4236 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
4239 2013-08-09 Jan Hubicka <jh@suse.cz>
4241 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
4243 2013-08-09 Jan Hubicka <jh@suse.cz>
4245 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
4247 (dump_cgraph_node): Dump profile-id.
4248 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
4249 and common_target_probability.
4250 * lto-cgraph.c (lto_output_edge): Stream common targets.
4251 (lto_output_node): Stream profile ids.
4252 (input_node): Stream profile ids.
4253 (input_edge): Stream common targets.
4254 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
4255 * ipa.c: Include value-prof.h
4256 (ipa_profile_generate_summary): Turn indirect call statement histograms
4257 into common targets.
4258 (ipa_profile): Turn common targets into speculative edges.
4260 2013-08-09 Jan Hubicka <jh@suse.cz>
4262 * cgraph.h (cgraph_node): Add profile_id.
4263 * value-prof.c (cgraph_node_map): Turn into pointer_map.
4264 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
4265 (del_node_map): Update.
4266 (find_func_by_funcdef_no): Replace by ...
4267 (find_func_by_profile_id): ... this one.
4268 (gimple_ic_transform): Do not remove useful histograms when
4269 speculation is not done; dump info when indirect call removal
4271 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
4272 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
4273 (__gcov_indirect_call_profiler_v2): .. this one.
4274 * profile.h (init_node_map): Update.
4275 * coverage.c (coverage_compute_profile_id): New function.
4276 * coverage.h (coverage_compute_profile_id): Declare.
4277 * tree-profile.c (init_ic_make_global_vars): Make
4278 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
4279 (gimple_init_edge_profiler): Update prototype of
4280 __gcov_indirect_call_profiler.
4281 (gimple_gen_ic_func_profiler): Simplify.
4282 (tree_profiling): Use init_node_map
4284 2013-08-09 Jan Hubicka <jh@suse.cz>
4286 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
4287 non-speculative refs.
4288 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
4289 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
4290 (cgraph_set_call_stmt): Likewise.
4291 (cgraph_create_edge_1): Fix release checking compilatoin;
4293 (cgraph_free_edge): Free indirect info.
4294 (cgraph_turn_edge_to_speculative): New function.
4295 (cgraph_speculative_call_info): New function.
4296 (cgraph_make_edge_direct): Return direct edge; handle speculation.
4297 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
4298 (dump_cgraph_node): Dump speculation.
4299 (verify_edge_count_and_frequency): Accept speculative edges.
4300 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
4301 (verify_cgraph_node): Handle speculation.
4302 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
4303 (cgraph_set_call_stmt): Update prototype.
4304 (cgraph_make_edge_direct): Update prototype.
4305 (cgraph_speculative_call_info): Declare.
4306 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
4307 to change; update call of ipa_find_references.
4308 * ipa-ref.c (ipa_record_reference): Fix return value; clear
4309 lto_stmt_uid and speculative flags.
4310 (ipa_dump_references): Dump speculation.
4311 (ipa_clone_references): Clone speculative flag.
4312 (ipa_clone_referring): Likewise.
4313 (ipa_clone_ref): New function.
4314 (ipa_find_reference): Look into lto_stmt_uids
4315 (ipa_clear_stmts_in_references): Do not clear speculative calls.
4316 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
4317 (ipa_find_reference): Update declaration.
4318 (ipa_clone_ref): Declare.
4319 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
4320 stream speculative flag.
4321 (lto_output_ref): Stream statements uids and speculation.
4322 (input_ref): Likewise.
4323 (input_edge): Stream speuclation.
4324 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
4325 (cgraph_set_call_stmt_including_clones): Handle speculation.
4326 * ipa-inline.c (heap_edge_removal_hook): New function.
4327 (inline_small_functions): Register it.
4328 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
4329 also initialize refs.
4330 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
4332 (try_make_edge_direct_simple_call): Likewise.
4333 (try_make_edge_direct_simple_call): Likewise.
4334 (update_indirect_edges_after_inlining): Likewise.
4335 (remove_described_reference): Look proper lto_stmt_uid.
4336 (propagate_controlled_uses): Likewise.
4337 (propagate_controlled_uses): Liekwise.
4338 * tree-inline.c (copy_bb): Copy speculative edges.
4339 (redirect_all_calls): New function.
4340 (copy_cfg_body): Do redirection after loop info is updated.
4341 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
4343 2013-08-09 Jan Hubicka <jh@suse.cz>
4345 * lto-streamer-out.c (output_function): Renumber PHIs.
4346 * lto-streamer-in.c (input_function): Likewise.
4348 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
4350 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
4351 (get_lane_unsigned): Likewise.
4352 (dup_lane_scalar): Likewise.
4353 (get_lane): enable for VALL.
4354 * config/aarch64/aarch64-simd.md
4355 (aarch64_dup_lane_scalar<mode>): Remove.
4356 (aarch64_get_lane_signed<mode>): Likewise.
4357 (aarch64_get_lane_unsigned<mode>): Likewise.
4358 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
4359 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
4360 (aarch64_get_lane<mode>): Enable for all vector modes.
4361 (aarch64_get_lanedi): Remove misleading constraints.
4362 * config/aarch64/arm_neon.h
4363 (__aarch64_vget_lane_any): Define.
4364 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
4365 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
4366 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
4367 * config/aarch64/iterators.md (VDQQH): New.
4371 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
4373 * configure.ac: Add GAS check for LEON instructions on SPARC.
4374 * configure: Regenerate.
4375 * config.in: Likewise.
4376 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
4378 * config/sparc/sparc.opt (LEON, LEON3): New masks.
4379 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
4381 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
4382 (AS_LEON_FLAG): New macro.
4383 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
4384 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
4385 Deal with LEON and LEON3 for the memory model.
4386 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
4387 (atomic_compare_and_swap<mode>_1): Likewise.
4388 (*atomic_compare_and_swap<mode>_1): Likewise.
4390 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4392 * config/arm/neon.md (vcond): Fix floating-point vector
4393 comparisons against 0.
4395 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
4397 * lra-constraints.c (emit_spill_move): Remove assert.
4398 (process_alt_operands): Add more debugging
4399 output. Increase reject for spilling into memory. Decrease
4400 reject for reloading scratch.
4401 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
4403 2013-08-08 Steve Ellcey <sellcey@mips.com>
4405 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
4406 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
4407 micromips incompatible. Add nan2008.
4408 (MULTILIB_DIRNAMES): Add nan2008.
4409 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4410 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
4411 and micromips incompatible. Add nan2008.
4412 (MULTILIB_DIRNAMES): Add nan2008.
4413 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4415 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
4417 PR rtl-optimization/58079
4418 * combine.c (combine_simplify_rtx): Avoid using SUBST if
4419 simplify_comparison has widened a comparison with an integer.
4421 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4423 * config/arm/neon.md (movmisalign<mode>): Disable when we
4424 don't allow unaligned accesses.
4425 (*movmisalign<mode>_neon_store): Likewise.
4426 (*movmisalign<mode>_neon_load): Likewise.
4427 (*movmisalign<mode>_neon_store): Likewise.
4428 (*movmisalign<mode>_neon_load): Likewise.
4430 2013-08-08 Jan Hubicka <jh@suse.cz>
4432 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
4433 (make_pass_rebuild_cgraph_edges): Also clear references.
4434 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
4435 * ipa-inline-transform.c (inline_transform): Remove all references
4437 * cgraphunit.c (expand_function): Remove all references after
4439 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
4440 (ipa_find_reference): Rewrite to iterator.
4441 (remove_stmt_references): Likewise.
4442 (ipa_clear_stmts_in_references): New function.
4443 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
4444 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
4446 * ipa-split.c (split_function): Remove references in split function.
4448 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4451 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
4452 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
4454 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
4457 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
4458 suggested mode for the assignment isn't compatible with the
4461 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4464 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4466 2013-08-07 Xinliang David Li <davidxl@google.com>
4468 * config/i386/i386.opt: New option -mtune-ctrl=.
4469 * config/i386/x86-tune.def: New file.
4470 * config/i386/i386.h: include x86-tune.def.
4471 * config/i386/i386.c (ix86_option_override_internal):
4472 Parsing -mtune-ctrl= option and set tune features.
4474 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
4477 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
4478 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4481 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4483 * rtl.h (update_alignments): Declare.
4484 * final.c (grow_label_align): New function extracted from...
4485 (shorten_branches): ...here. Call it.
4486 (update_alignments): New function.
4487 * reorg.c (sibling_labels): New variable.
4488 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4489 the new label along with it onto the sibling_labels vector.
4490 (fill_simple_delay_slots): Adjust call to get_label_before.
4491 (fill_slots_from_thread): Likewise.
4492 (relax_delay_slots): Likewise.
4493 (make_return_insns): Likewise.
4494 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4496 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4498 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4499 document its semantics.
4500 (diagnostic_report_diagnostic): Adjust accordingly.
4502 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4504 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
4505 (sparc_option_override): ...and port to new C++ pass API.
4506 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4508 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4510 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4512 2013-08-06 Caroline Tice <cmtice@google.com>
4514 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4515 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4516 * tree-pass.h: Add pass_vtable_verify.
4517 * varasm.c (assemble_variable): Add code to properly set the comdat
4518 section and name for the .vtable_map_vars section.
4519 (assemble_vtyv_preinit_initializer): New function.
4520 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4522 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4523 * vtable-verify.c: New file.
4524 * vtable-verify.h: New file.
4525 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4526 initialiation levels.
4527 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4528 * passes.def: Insert pass_vtable_verify.
4529 * aclocal.m4: Reorder includes.
4530 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4531 -fvtv-counts options.
4532 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4533 as appropriate, if -fvtable-verify=... is used.
4534 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4535 -fvtable-verify=... is used.
4536 * Makefile.in (OBJS): Add vtable-verify.o to list.
4537 (vtable-verify.o): Add new build rule.
4538 (GTFILES): Add vtable-verify.c to list.
4539 * common.opt (fvtable-verify=): New flag.
4540 (vtv_priority): Values for fvtable-verify= flag.
4541 (fvtv-counts): New flag.
4542 (fvtv-debug): New flag.
4543 * tree.h (save_vtable_map_decl): New extern function decl.
4545 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4547 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4548 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4549 (pass_data_rl78_devirt): ...new pass_data instance and...
4550 (make_pass_rl78_devirt): ...new function.
4551 (rl78_asm_file_start): Port pass registration to new C++ API.
4553 2013-08-07 David Malcolm <dmalcolm@redhat.com>
4555 * coretypes.h (rtl_opt_pass): Add.
4556 (gcc::context): Add.
4557 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4558 (epiphany_init): Port to new C++ pass API.
4559 (epiphany_optimize_mode_switching): Likewise.
4560 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4561 (pass_manager::get_pass_mode_switching): New.
4562 (pass_manager::get_pass_peephole2): New.
4563 * mode-switching.c (pass_mode_switching): Add clone method.
4564 * recog.c (pass_peephole2): Add clone method.
4565 (pass_split_all_insns): Add clone method.
4567 2013-08-06 David Malcolm <dmalcolm@redhat.com>
4569 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4570 (mips_option_override): ...here, porting to new C++ API for passes.
4572 2013-08-06 Jan Hubicka <jh@suse.cz>
4574 * cgraph.c (cgraph_get_body): New function based on lto.c
4576 * cgraph.h (cgraph_get_body): Declare.
4577 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4579 * cgraphunit.c (expand_function): Get body prior expanding.
4580 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
4581 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4583 * passes.c (do_per_function_toporder): Get body.
4584 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4585 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4587 2013-08-06 Martin Jambor <mjambor@suse.cz>
4590 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4591 is not re-finalized. Rename second parameter to no_collect.
4593 2013-08-06 Martin Jambor <mjambor@suse.cz>
4596 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4597 MEM_REF has proper alignment information.
4599 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
4602 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4604 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4605 rtx (*) (rtx, ...) with insn_gen_fn.
4606 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4607 insn_gen_fn::stored_funcptr. Add initializer braces.
4609 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4611 Rewrite how instances of passes are cloned to remove assumptions
4612 about their sizes (thus allowing pass subclasses to have
4613 additional data fields, albeit non-GC-managed ones at this point).
4615 * passes.c (make_pass_instance): Now that passes have clone
4616 methods, rewrite this function to eliminate XNEW and memcpy
4617 calls that used hardcoded sizes. Since this function no longer
4618 creates pass instances, rename it to...
4619 (add_pass_instance): ...this. Document the old way that passes were
4620 numbered and flagged, and rework this function to continue using it.
4621 (next_pass_1): Add an initial_pass argument for use by
4623 (position_pass): When adding multiple instances of a pass, use
4624 the pass's clone method, rather than relying on the XNEW/memcpy
4625 within the former make_pass_instance (now add_pass_instance).
4626 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4627 the initial instance of the current pass within the pass manager.
4629 2013-08-05 David Malcolm <dmalcolm@redhat.com>
4631 This is the automated part of the conversion of passes from C
4632 structs to C++ classes.
4634 Patch autogenerated by refactor_passes.py from
4635 https://github.com/davidmalcolm/gcc-refactoring-scripts
4636 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4638 * asan.c (pass_asan): Convert from a global struct to a subclass of
4639 gimple_opt_pass along with...
4640 (pass_data_asan): ...new pass_data instance and...
4641 (make_pass_asan): ...new function.
4642 (pass_asan_O0): Convert from a global struct to a subclass of
4643 gimple_opt_pass along with...
4644 (pass_data_asan_O0): ...new pass_data instance and...
4645 (make_pass_asan_O0): ...new function.
4646 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4647 subclass of rtl_opt_pass along with...
4648 (pass_data_inc_dec): ...new pass_data instance and...
4649 (make_pass_inc_dec): ...new function.
4650 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4651 a subclass of rtl_opt_pass along with...
4652 (pass_data_reorder_blocks): ...new pass_data instance and...
4653 (make_pass_reorder_blocks): ...new function.
4654 (pass_duplicate_computed_gotos): Convert from a global struct to a
4655 subclass of rtl_opt_pass along with...
4656 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4657 (make_pass_duplicate_computed_gotos): ...new function.
4658 (pass_partition_blocks): Convert from a global struct to a subclass of
4659 rtl_opt_pass along with...
4660 (pass_data_partition_blocks): ...new pass_data instance and...
4661 (make_pass_partition_blocks): ...new function.
4662 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4663 struct to a subclass of rtl_opt_pass along with...
4664 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4666 (make_pass_branch_target_load_optimize1): ...new function.
4667 (pass_branch_target_load_optimize2): Convert from a global struct to a
4668 subclass of rtl_opt_pass along with...
4669 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4671 (make_pass_branch_target_load_optimize2): ...new function.
4672 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4673 of rtl_opt_pass along with...
4674 (pass_data_jump): ...new pass_data instance and...
4675 (make_pass_jump): ...new function.
4676 (pass_jump2): Convert from a global struct to a subclass of
4677 rtl_opt_pass along with...
4678 (pass_data_jump2): ...new pass_data instance and...
4679 (make_pass_jump2): ...new function.
4680 * cfgexpand.c (pass_expand): Convert from a global struct to a
4681 subclass of rtl_opt_pass along with...
4682 (pass_data_expand): ...new pass_data instance and...
4683 (make_pass_expand): ...new function.
4684 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4685 of rtl_opt_pass along with...
4686 (pass_data_free_cfg): ...new pass_data instance and...
4687 (make_pass_free_cfg): ...new function.
4688 (pass_into_cfg_layout_mode): Convert from a global struct to a
4689 subclass of rtl_opt_pass along with...
4690 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4691 (make_pass_into_cfg_layout_mode): ...new function.
4692 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4693 subclass of rtl_opt_pass along with...
4694 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4695 (make_pass_outof_cfg_layout_mode): ...new function.
4696 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4697 struct to a subclass of gimple_opt_pass along with...
4698 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4699 (make_pass_build_cgraph_edges): ...new function.
4700 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4701 subclass of gimple_opt_pass along with...
4702 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4703 (make_pass_rebuild_cgraph_edges): ...new function.
4704 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4705 subclass of gimple_opt_pass along with...
4706 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4708 (make_pass_remove_cgraph_callee_edges): ...new function.
4709 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4710 struct to a subclass of rtl_opt_pass along with...
4711 (pass_data_stack_adjustments): ...new pass_data instance and...
4712 (make_pass_stack_adjustments): ...new function.
4713 * combine.c (pass_combine): Convert from a global struct to a subclass
4714 of rtl_opt_pass along with...
4715 (pass_data_combine): ...new pass_data instance and...
4716 (make_pass_combine): ...new function.
4717 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4718 global struct to a subclass of rtl_opt_pass along with...
4719 (pass_data_compare_elim_after_reload): ...new pass_data instance
4721 (make_pass_compare_elim_after_reload): ...new function.
4722 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4723 of rtl_opt_pass along with...
4724 (pass_data_rtl_cprop): ...new pass_data instance and...
4725 (make_pass_rtl_cprop): ...new function.
4726 * cse.c (pass_cse): Convert from a global struct to a subclass of
4727 rtl_opt_pass along with...
4728 (pass_data_cse): ...new pass_data instance and...
4729 (make_pass_cse): ...new function.
4730 (pass_cse2): Convert from a global struct to a subclass of
4731 rtl_opt_pass along with...
4732 (pass_data_cse2): ...new pass_data instance and...
4733 (make_pass_cse2): ...new function.
4734 (pass_cse_after_global_opts): Convert from a global struct to a
4735 subclass of rtl_opt_pass along with...
4736 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4737 (make_pass_cse_after_global_opts): ...new function.
4738 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4739 of rtl_opt_pass along with...
4740 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4741 (make_pass_ud_rtl_dce): ...new function.
4742 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4743 rtl_opt_pass along with...
4744 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4745 (make_pass_fast_rtl_dce): ...new function.
4746 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4747 a subclass of rtl_opt_pass along with...
4748 (pass_data_df_initialize_opt): ...new pass_data instance and...
4749 (make_pass_df_initialize_opt): ...new function.
4750 (pass_df_initialize_no_opt): Convert from a global struct to a
4751 subclass of rtl_opt_pass along with...
4752 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4753 (make_pass_df_initialize_no_opt): ...new function.
4754 (pass_df_finish): Convert from a global struct to a subclass of
4755 rtl_opt_pass along with...
4756 (pass_data_df_finish): ...new pass_data instance and...
4757 (make_pass_df_finish): ...new function.
4758 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4759 rtl_opt_pass along with...
4760 (pass_data_rtl_dse1): ...new pass_data instance and...
4761 (make_pass_rtl_dse1): ...new function.
4762 (pass_rtl_dse2): Convert from a global struct to a subclass of
4763 rtl_opt_pass along with...
4764 (pass_data_rtl_dse2): ...new pass_data instance and...
4765 (make_pass_rtl_dse2): ...new function.
4766 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4767 subclass of rtl_opt_pass along with...
4768 (pass_data_dwarf2_frame): ...new pass_data instance and...
4769 (make_pass_dwarf2_frame): ...new function.
4770 * except.c (pass_set_nothrow_function_flags): Convert from a global
4771 struct to a subclass of rtl_opt_pass along with...
4772 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4774 (make_pass_set_nothrow_function_flags): ...new function.
4775 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4776 subclass of rtl_opt_pass along with...
4777 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4779 (make_pass_convert_to_eh_region_ranges): ...new function.
4780 * final.c (pass_compute_alignments): Convert from a global struct to a
4781 subclass of rtl_opt_pass along with...
4782 (pass_data_compute_alignments): ...new pass_data instance and...
4783 (make_pass_compute_alignments): ...new function.
4784 (pass_final): Convert from a global struct to a subclass of
4785 rtl_opt_pass along with...
4786 (pass_data_final): ...new pass_data instance and...
4787 (make_pass_final): ...new function.
4788 (pass_shorten_branches): Convert from a global struct to a subclass of
4789 rtl_opt_pass along with...
4790 (pass_data_shorten_branches): ...new pass_data instance and...
4791 (make_pass_shorten_branches): ...new function.
4792 (pass_clean_state): Convert from a global struct to a subclass of
4793 rtl_opt_pass along with...
4794 (pass_data_clean_state): ...new pass_data instance and...
4795 (make_pass_clean_state): ...new function.
4796 * function.c (pass_instantiate_virtual_regs): Convert from a global
4797 struct to a subclass of rtl_opt_pass along with...
4798 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4799 (make_pass_instantiate_virtual_regs): ...new function.
4800 (pass_leaf_regs): Convert from a global struct to a subclass of
4801 rtl_opt_pass along with...
4802 (pass_data_leaf_regs): ...new pass_data instance and...
4803 (make_pass_leaf_regs): ...new function.
4804 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4805 subclass of rtl_opt_pass along with...
4806 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4808 (make_pass_thread_prologue_and_epilogue): ...new function.
4809 (pass_match_asm_constraints): Convert from a global struct to a
4810 subclass of rtl_opt_pass along with...
4811 (pass_data_match_asm_constraints): ...new pass_data instance and...
4812 (make_pass_match_asm_constraints): ...new function.
4813 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4814 subclass of rtl_opt_pass along with...
4815 (pass_data_rtl_fwprop): ...new pass_data instance and...
4816 (make_pass_rtl_fwprop): ...new function.
4817 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4818 rtl_opt_pass along with...
4819 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4820 (make_pass_rtl_fwprop_addr): ...new function.
4821 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4822 rtl_opt_pass along with...
4823 (pass_data_rtl_pre): ...new pass_data instance and...
4824 (make_pass_rtl_pre): ...new function.
4825 (pass_rtl_hoist): Convert from a global struct to a subclass of
4826 rtl_opt_pass along with...
4827 (pass_data_rtl_hoist): ...new pass_data instance and...
4828 (make_pass_rtl_hoist): ...new function.
4829 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4830 subclass of gimple_opt_pass along with...
4831 (pass_data_lower_cf): ...new pass_data instance and...
4832 (make_pass_lower_cf): ...new function.
4833 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4834 from a global struct to a subclass of gimple_opt_pass along with...
4835 (pass_data_strength_reduction): ...new pass_data instance and...
4836 (make_pass_strength_reduction): ...new function.
4837 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4838 of rtl_opt_pass along with...
4839 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4840 (make_pass_rtl_ifcvt): ...new function.
4841 (pass_if_after_combine): Convert from a global struct to a subclass of
4842 rtl_opt_pass along with...
4843 (pass_data_if_after_combine): ...new pass_data instance and...
4844 (make_pass_if_after_combine): ...new function.
4845 (pass_if_after_reload): Convert from a global struct to a subclass of
4846 rtl_opt_pass along with...
4847 (pass_data_if_after_reload): ...new pass_data instance and...
4848 (make_pass_if_after_reload): ...new function.
4849 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4850 a subclass of rtl_opt_pass along with...
4851 (pass_data_initialize_regs): ...new pass_data instance and...
4852 (make_pass_initialize_regs): ...new function.
4853 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4854 of ipa_opt_pass_d along with...
4855 (pass_data_ipa_cp): ...new pass_data instance and...
4856 (make_pass_ipa_cp): ...new function.
4857 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4858 global struct to a subclass of gimple_opt_pass along with...
4859 (pass_data_inline_parameters): ...new pass_data instance and...
4860 (make_pass_inline_parameters): ...new function.
4861 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4862 subclass of gimple_opt_pass along with...
4863 (pass_data_early_inline): ...new pass_data instance and...
4864 (make_pass_early_inline): ...new function.
4865 (pass_ipa_inline): Convert from a global struct to a subclass of
4866 ipa_opt_pass_d along with...
4867 (pass_data_ipa_inline): ...new pass_data instance and...
4868 (make_pass_ipa_inline): ...new function.
4869 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4870 struct to a subclass of gimple_opt_pass along with...
4871 (pass_data_local_pure_const): ...new pass_data instance and...
4872 (make_pass_local_pure_const): ...new function.
4873 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4874 ipa_opt_pass_d along with...
4875 (pass_data_ipa_pure_const): ...new pass_data instance and...
4876 (make_pass_ipa_pure_const): ...new function.
4877 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4878 to a subclass of ipa_opt_pass_d along with...
4879 (pass_data_ipa_reference): ...new pass_data instance and...
4880 (make_pass_ipa_reference): ...new function.
4881 * ipa-split.c (pass_split_functions): Convert from a global struct to
4882 a subclass of gimple_opt_pass along with...
4883 (pass_data_split_functions): ...new pass_data instance and...
4884 (make_pass_split_functions): ...new function.
4885 (pass_feedback_split_functions): Convert from a global struct to a
4886 subclass of gimple_opt_pass along with...
4887 (pass_data_feedback_split_functions): ...new pass_data instance and...
4888 (make_pass_feedback_split_functions): ...new function.
4889 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4890 global struct to a subclass of simple_ipa_opt_pass along with...
4891 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4893 (make_pass_ipa_function_and_variable_visibility): ...new function.
4894 (pass_ipa_free_inline_summary): Convert from a global struct to a
4895 subclass of simple_ipa_opt_pass along with...
4896 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4897 (make_pass_ipa_free_inline_summary): ...new function.
4898 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4899 subclass of ipa_opt_pass_d along with...
4900 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4902 (make_pass_ipa_whole_program_visibility): ...new function.
4903 (pass_ipa_profile): Convert from a global struct to a subclass of
4904 ipa_opt_pass_d along with...
4905 (pass_data_ipa_profile): ...new pass_data instance and...
4906 (make_pass_ipa_profile): ...new function.
4907 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4908 ipa_opt_pass_d along with...
4909 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4910 (make_pass_ipa_cdtor_merge): ...new function.
4911 * ira.c (pass_ira): Convert from a global struct to a subclass of
4912 rtl_opt_pass along with...
4913 (pass_data_ira): ...new pass_data instance and...
4914 (make_pass_ira): ...new function.
4915 (pass_reload): Convert from a global struct to a subclass of
4916 rtl_opt_pass along with...
4917 (pass_data_reload): ...new pass_data instance and...
4918 (make_pass_reload): ...new function.
4919 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4920 subclass of rtl_opt_pass along with...
4921 (pass_data_cleanup_barriers): ...new pass_data instance and...
4922 (make_pass_cleanup_barriers): ...new function.
4923 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4924 of rtl_opt_pass along with...
4925 (pass_data_loop2): ...new pass_data instance and...
4926 (make_pass_loop2): ...new function.
4927 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4928 rtl_opt_pass along with...
4929 (pass_data_rtl_loop_init): ...new pass_data instance and...
4930 (make_pass_rtl_loop_init): ...new function.
4931 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4932 rtl_opt_pass along with...
4933 (pass_data_rtl_loop_done): ...new pass_data instance and...
4934 (make_pass_rtl_loop_done): ...new function.
4935 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4936 subclass of rtl_opt_pass along with...
4937 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4938 (make_pass_rtl_move_loop_invariants): ...new function.
4939 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4940 rtl_opt_pass along with...
4941 (pass_data_rtl_unswitch): ...new pass_data instance and...
4942 (make_pass_rtl_unswitch): ...new function.
4943 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4944 subclass of rtl_opt_pass along with...
4945 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4947 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4948 (pass_rtl_doloop): Convert from a global struct to a subclass of
4949 rtl_opt_pass along with...
4950 (pass_data_rtl_doloop): ...new pass_data instance and...
4951 (make_pass_rtl_doloop): ...new function.
4952 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4953 a subclass of rtl_opt_pass along with...
4954 (pass_data_lower_subreg): ...new pass_data instance and...
4955 (make_pass_lower_subreg): ...new function.
4956 (pass_lower_subreg2): Convert from a global struct to a subclass of
4957 rtl_opt_pass along with...
4958 (pass_data_lower_subreg2): ...new pass_data instance and...
4959 (make_pass_lower_subreg2): ...new function.
4960 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4961 struct to a subclass of ipa_opt_pass_d along with...
4962 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4963 (make_pass_ipa_lto_gimple_out): ...new function.
4964 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4965 of ipa_opt_pass_d along with...
4966 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4967 (make_pass_ipa_lto_finish_out): ...new function.
4968 * mode-switching.c (pass_mode_switching): Convert from a global struct
4969 to a subclass of rtl_opt_pass along with...
4970 (pass_data_mode_switching): ...new pass_data instance and...
4971 (make_pass_mode_switching): ...new function.
4972 * modulo-sched.c (pass_sms): Convert from a global struct to a
4973 subclass of rtl_opt_pass along with...
4974 (pass_data_sms): ...new pass_data instance and...
4975 (make_pass_sms): ...new function.
4976 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4977 subclass of gimple_opt_pass along with...
4978 (pass_data_expand_omp): ...new pass_data instance and...
4979 (make_pass_expand_omp): ...new function.
4980 (pass_lower_omp): Convert from a global struct to a subclass of
4981 gimple_opt_pass along with...
4982 (pass_data_lower_omp): ...new pass_data instance and...
4983 (make_pass_lower_omp): ...new function.
4984 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4985 of gimple_opt_pass along with...
4986 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4987 (make_pass_diagnose_omp_blocks): ...new function.
4988 * passes.c (pass_early_local_passes): Convert from a global struct to
4989 a subclass of simple_ipa_opt_pass along with...
4990 (pass_data_early_local_passes): ...new pass_data instance and...
4991 (make_pass_early_local_passes): ...new function.
4992 (pass_all_early_optimizations): Convert from a global struct to a
4993 subclass of gimple_opt_pass along with...
4994 (pass_data_all_early_optimizations): ...new pass_data instance and...
4995 (make_pass_all_early_optimizations): ...new function.
4996 (pass_all_optimizations): Convert from a global struct to a subclass
4997 of gimple_opt_pass along with...
4998 (pass_data_all_optimizations): ...new pass_data instance and...
4999 (make_pass_all_optimizations): ...new function.
5000 (pass_all_optimizations_g): Convert from a global struct to a subclass
5001 of gimple_opt_pass along with...
5002 (pass_data_all_optimizations_g): ...new pass_data instance and...
5003 (make_pass_all_optimizations_g): ...new function.
5004 (pass_rest_of_compilation): Convert from a global struct to a subclass
5005 of rtl_opt_pass along with...
5006 (pass_data_rest_of_compilation): ...new pass_data instance and...
5007 (make_pass_rest_of_compilation): ...new function.
5008 (pass_postreload): Convert from a global struct to a subclass of
5009 rtl_opt_pass along with...
5010 (pass_data_postreload): ...new pass_data instance and...
5011 (make_pass_postreload): ...new function.
5012 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
5013 subclass of rtl_opt_pass along with...
5014 (pass_data_gcse2): ...new pass_data instance and...
5015 (make_pass_gcse2): ...new function.
5016 * postreload.c (pass_postreload_cse): Convert from a global struct to
5017 a subclass of rtl_opt_pass along with...
5018 (pass_data_postreload_cse): ...new pass_data instance and...
5019 (make_pass_postreload_cse): ...new function.
5020 * predict.c (pass_profile): Convert from a global struct to a subclass
5021 of gimple_opt_pass along with...
5022 (pass_data_profile): ...new pass_data instance and...
5023 (make_pass_profile): ...new function.
5024 (pass_strip_predict_hints): Convert from a global struct to a subclass
5025 of gimple_opt_pass along with...
5026 (pass_data_strip_predict_hints): ...new pass_data instance and...
5027 (make_pass_strip_predict_hints): ...new function.
5028 * recog.c (pass_peephole2): Convert from a global struct to a subclass
5029 of rtl_opt_pass along with...
5030 (pass_data_peephole2): ...new pass_data instance and...
5031 (make_pass_peephole2): ...new function.
5032 (pass_split_all_insns): Convert from a global struct to a subclass of
5033 rtl_opt_pass along with...
5034 (pass_data_split_all_insns): ...new pass_data instance and...
5035 (make_pass_split_all_insns): ...new function.
5036 (pass_split_after_reload): Convert from a global struct to a subclass
5037 of rtl_opt_pass along with...
5038 (pass_data_split_after_reload): ...new pass_data instance and...
5039 (make_pass_split_after_reload): ...new function.
5040 (pass_split_before_regstack): Convert from a global struct to a
5041 subclass of rtl_opt_pass along with...
5042 (pass_data_split_before_regstack): ...new pass_data instance and...
5043 (make_pass_split_before_regstack): ...new function.
5044 (pass_split_before_sched2): Convert from a global struct to a subclass
5045 of rtl_opt_pass along with...
5046 (pass_data_split_before_sched2): ...new pass_data instance and...
5047 (make_pass_split_before_sched2): ...new function.
5048 (pass_split_for_shorten_branches): Convert from a global struct to a
5049 subclass of rtl_opt_pass along with...
5050 (pass_data_split_for_shorten_branches): ...new pass_data instance
5052 (make_pass_split_for_shorten_branches): ...new function.
5053 * ree.c (pass_ree): Convert from a global struct to a subclass of
5054 rtl_opt_pass along with...
5055 (pass_data_ree): ...new pass_data instance and...
5056 (make_pass_ree): ...new function.
5057 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
5058 subclass of rtl_opt_pass along with...
5059 (pass_data_stack_regs): ...new pass_data instance and...
5060 (make_pass_stack_regs): ...new function.
5061 (pass_stack_regs_run): Convert from a global struct to a subclass of
5062 rtl_opt_pass along with...
5063 (pass_data_stack_regs_run): ...new pass_data instance and...
5064 (make_pass_stack_regs_run): ...new function.
5065 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
5066 subclass of rtl_opt_pass along with...
5067 (pass_data_cprop_hardreg): ...new pass_data instance and...
5068 (make_pass_cprop_hardreg): ...new function.
5069 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
5070 subclass of rtl_opt_pass along with...
5071 (pass_data_reginfo_init): ...new pass_data instance and...
5072 (make_pass_reginfo_init): ...new function.
5073 * regmove.c (pass_regmove): Convert from a global struct to a subclass
5074 of rtl_opt_pass along with...
5075 (pass_data_regmove): ...new pass_data instance and...
5076 (make_pass_regmove): ...new function.
5077 * regrename.c (pass_regrename): Convert from a global struct to a
5078 subclass of rtl_opt_pass along with...
5079 (pass_data_regrename): ...new pass_data instance and...
5080 (make_pass_regrename): ...new function.
5081 * reorg.c (pass_delay_slots): Convert from a global struct to a
5082 subclass of rtl_opt_pass along with...
5083 (pass_data_delay_slots): ...new pass_data instance and...
5084 (make_pass_delay_slots): ...new function.
5085 (pass_machine_reorg): Convert from a global struct to a subclass of
5086 rtl_opt_pass along with...
5087 (pass_data_machine_reorg): ...new pass_data instance and...
5088 (make_pass_machine_reorg): ...new function.
5089 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
5090 of rtl_opt_pass along with...
5091 (pass_data_sched): ...new pass_data instance and...
5092 (make_pass_sched): ...new function.
5093 (pass_sched2): Convert from a global struct to a subclass of
5094 rtl_opt_pass along with...
5095 (pass_data_sched2): ...new pass_data instance and...
5096 (make_pass_sched2): ...new function.
5097 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
5098 to a subclass of rtl_opt_pass along with...
5099 (pass_data_stack_ptr_mod): ...new pass_data instance and...
5100 (make_pass_stack_ptr_mod): ...new function.
5101 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
5102 to a subclass of rtl_opt_pass along with...
5103 (pass_data_rtl_store_motion): ...new pass_data instance and...
5104 (make_pass_rtl_store_motion): ...new function.
5105 * tracer.c (pass_tracer): Convert from a global struct to a subclass
5106 of gimple_opt_pass along with...
5107 (pass_data_tracer): ...new pass_data instance and...
5108 (make_pass_tracer): ...new function.
5109 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
5110 to a subclass of gimple_opt_pass along with...
5111 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
5112 (make_pass_diagnose_tm_blocks): ...new function.
5113 (pass_lower_tm): Convert from a global struct to a subclass of
5114 gimple_opt_pass along with...
5115 (pass_data_lower_tm): ...new pass_data instance and...
5116 (make_pass_lower_tm): ...new function.
5117 (pass_tm_init): Convert from a global struct to a subclass of
5118 gimple_opt_pass along with...
5119 (pass_data_tm_init): ...new pass_data instance and...
5120 (make_pass_tm_init): ...new function.
5121 (pass_tm_mark): Convert from a global struct to a subclass of
5122 gimple_opt_pass along with...
5123 (pass_data_tm_mark): ...new pass_data instance and...
5124 (make_pass_tm_mark): ...new function.
5125 (pass_tm_edges): Convert from a global struct to a subclass of
5126 gimple_opt_pass along with...
5127 (pass_data_tm_edges): ...new pass_data instance and...
5128 (make_pass_tm_edges): ...new function.
5129 (pass_tm_memopt): Convert from a global struct to a subclass of
5130 gimple_opt_pass along with...
5131 (pass_data_tm_memopt): ...new pass_data instance and...
5132 (make_pass_tm_memopt): ...new function.
5133 (pass_ipa_tm): Convert from a global struct to a subclass of
5134 simple_ipa_opt_pass along with...
5135 (pass_data_ipa_tm): ...new pass_data instance and...
5136 (make_pass_ipa_tm): ...new function.
5137 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
5138 subclass of gimple_opt_pass along with...
5139 (pass_data_call_cdce): ...new pass_data instance and...
5140 (make_pass_call_cdce): ...new function.
5141 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
5142 subclass of gimple_opt_pass along with...
5143 (pass_data_build_cfg): ...new pass_data instance and...
5144 (make_pass_build_cfg): ...new function.
5145 (pass_split_crit_edges): Convert from a global struct to a subclass of
5146 gimple_opt_pass along with...
5147 (pass_data_split_crit_edges): ...new pass_data instance and...
5148 (make_pass_split_crit_edges): ...new function.
5149 (pass_warn_function_return): Convert from a global struct to a
5150 subclass of gimple_opt_pass along with...
5151 (pass_data_warn_function_return): ...new pass_data instance and...
5152 (make_pass_warn_function_return): ...new function.
5153 (pass_warn_function_noreturn): Convert from a global struct to a
5154 subclass of gimple_opt_pass along with...
5155 (pass_data_warn_function_noreturn): ...new pass_data instance and...
5156 (make_pass_warn_function_noreturn): ...new function.
5157 (pass_warn_unused_result): Convert from a global struct to a subclass
5158 of gimple_opt_pass along with...
5159 (pass_data_warn_unused_result): ...new pass_data instance and...
5160 (make_pass_warn_unused_result): ...new function.
5161 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
5162 a subclass of gimple_opt_pass along with...
5163 (pass_data_merge_phi): ...new pass_data instance and...
5164 (make_pass_merge_phi): ...new function.
5165 * tree-complex.c (pass_lower_complex): Convert from a global struct to
5166 a subclass of gimple_opt_pass along with...
5167 (pass_data_lower_complex): ...new pass_data instance and...
5168 (make_pass_lower_complex): ...new function.
5169 (pass_lower_complex_O0): Convert from a global struct to a subclass of
5170 gimple_opt_pass along with...
5171 (pass_data_lower_complex_O0): ...new pass_data instance and...
5172 (make_pass_lower_complex_O0): ...new function.
5173 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
5174 subclass of gimple_opt_pass along with...
5175 (pass_data_lower_eh): ...new pass_data instance and...
5176 (make_pass_lower_eh): ...new function.
5177 (pass_refactor_eh): Convert from a global struct to a subclass of
5178 gimple_opt_pass along with...
5179 (pass_data_refactor_eh): ...new pass_data instance and...
5180 (make_pass_refactor_eh): ...new function.
5181 (pass_lower_resx): Convert from a global struct to a subclass of
5182 gimple_opt_pass along with...
5183 (pass_data_lower_resx): ...new pass_data instance and...
5184 (make_pass_lower_resx): ...new function.
5185 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
5186 of gimple_opt_pass along with...
5187 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
5188 (make_pass_lower_eh_dispatch): ...new function.
5189 (pass_cleanup_eh): Convert from a global struct to a subclass of
5190 gimple_opt_pass along with...
5191 (pass_data_cleanup_eh): ...new pass_data instance and...
5192 (make_pass_cleanup_eh): ...new function.
5193 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
5194 to a subclass of simple_ipa_opt_pass along with...
5195 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
5196 (make_pass_ipa_lower_emutls): ...new function.
5197 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
5198 a subclass of gimple_opt_pass along with...
5199 (pass_data_if_conversion): ...new pass_data instance and...
5200 (make_pass_if_conversion): ...new function.
5201 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
5202 subclass of gimple_opt_pass along with...
5203 (pass_data_build_ssa): ...new pass_data instance and...
5204 (make_pass_build_ssa): ...new function.
5205 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
5206 global struct to a subclass of gimple_opt_pass along with...
5207 (pass_data_loop_distribution): ...new pass_data instance and...
5208 (make_pass_loop_distribution): ...new function.
5209 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
5210 subclass of gimple_opt_pass along with...
5211 (pass_data_mudflap_1): ...new pass_data instance and...
5212 (make_pass_mudflap_1): ...new function.
5213 (pass_mudflap_2): Convert from a global struct to a subclass of
5214 gimple_opt_pass along with...
5215 (pass_data_mudflap_2): ...new pass_data instance and...
5216 (make_pass_mudflap_2): ...new function.
5217 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
5218 subclass of gimple_opt_pass along with...
5219 (pass_data_mudflap_1): ...new pass_data instance and...
5220 (make_pass_mudflap_1): ...new function.
5221 (pass_mudflap_2): Convert from a global struct to a subclass of
5222 gimple_opt_pass along with...
5223 (pass_data_mudflap_2): ...new pass_data instance and...
5224 (make_pass_mudflap_2): ...new function.
5225 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
5226 gimple_opt_pass along with...
5227 (pass_data_nrv): ...new pass_data instance and...
5228 (make_pass_nrv): ...new function.
5229 (pass_return_slot): Convert from a global struct to a subclass of
5230 gimple_opt_pass along with...
5231 (pass_data_return_slot): ...new pass_data instance and...
5232 (make_pass_return_slot): ...new function.
5233 * tree-object-size.c (pass_object_sizes): Convert from a global struct
5234 to a subclass of gimple_opt_pass along with...
5235 (pass_data_object_sizes): ...new pass_data instance and...
5236 (make_pass_object_sizes): ...new function.
5237 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
5238 global struct to a subclass of gimple_opt_pass along with...
5239 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
5241 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5242 (pass_fixup_cfg): Convert from a global struct to a subclass of
5243 gimple_opt_pass along with...
5244 (pass_data_fixup_cfg): ...new pass_data instance and...
5245 (make_pass_fixup_cfg): ...new function.
5246 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
5247 (make_pass_mudflap_1): ...new function.
5248 (pass_mudflap_2): Replace declaration with that of...
5249 (make_pass_mudflap_2): ...new function.
5250 (pass_asan): Replace declaration with that of...
5251 (make_pass_asan): ...new function.
5252 (pass_asan_O0): Replace declaration with that of...
5253 (make_pass_asan_O0): ...new function.
5254 (pass_tsan): Replace declaration with that of...
5255 (make_pass_tsan): ...new function.
5256 (pass_tsan_O0): Replace declaration with that of...
5257 (make_pass_tsan_O0): ...new function.
5258 (pass_lower_cf): Replace declaration with that of...
5259 (make_pass_lower_cf): ...new function.
5260 (pass_refactor_eh): Replace declaration with that of...
5261 (make_pass_refactor_eh): ...new function.
5262 (pass_lower_eh): Replace declaration with that of...
5263 (make_pass_lower_eh): ...new function.
5264 (pass_lower_eh_dispatch): Replace declaration with that of...
5265 (make_pass_lower_eh_dispatch): ...new function.
5266 (pass_lower_resx): Replace declaration with that of...
5267 (make_pass_lower_resx): ...new function.
5268 (pass_build_cfg): Replace declaration with that of...
5269 (make_pass_build_cfg): ...new function.
5270 (pass_early_tree_profile): Replace declaration with that of...
5271 (make_pass_early_tree_profile): ...new function.
5272 (pass_cleanup_eh): Replace declaration with that of...
5273 (make_pass_cleanup_eh): ...new function.
5274 (pass_sra): Replace declaration with that of...
5275 (make_pass_sra): ...new function.
5276 (pass_sra_early): Replace declaration with that of...
5277 (make_pass_sra_early): ...new function.
5278 (pass_early_ipa_sra): Replace declaration with that of...
5279 (make_pass_early_ipa_sra): ...new function.
5280 (pass_tail_recursion): Replace declaration with that of...
5281 (make_pass_tail_recursion): ...new function.
5282 (pass_tail_calls): Replace declaration with that of...
5283 (make_pass_tail_calls): ...new function.
5284 (pass_tree_loop): Replace declaration with that of...
5285 (make_pass_tree_loop): ...new function.
5286 (pass_tree_loop_init): Replace declaration with that of...
5287 (make_pass_tree_loop_init): ...new function.
5288 (pass_lim): Replace declaration with that of...
5289 (make_pass_lim): ...new function.
5290 (pass_tree_unswitch): Replace declaration with that of...
5291 (make_pass_tree_unswitch): ...new function.
5292 (pass_predcom): Replace declaration with that of...
5293 (make_pass_predcom): ...new function.
5294 (pass_iv_canon): Replace declaration with that of...
5295 (make_pass_iv_canon): ...new function.
5296 (pass_scev_cprop): Replace declaration with that of...
5297 (make_pass_scev_cprop): ...new function.
5298 (pass_empty_loop): Replace declaration with that of...
5299 (make_pass_empty_loop): ...new function.
5300 (pass_record_bounds): Replace declaration with that of...
5301 (make_pass_record_bounds): ...new function.
5302 (pass_graphite): Replace declaration with that of...
5303 (make_pass_graphite): ...new function.
5304 (pass_graphite_transforms): Replace declaration with that of...
5305 (make_pass_graphite_transforms): ...new function.
5306 (pass_if_conversion): Replace declaration with that of...
5307 (make_pass_if_conversion): ...new function.
5308 (pass_loop_distribution): Replace declaration with that of...
5309 (make_pass_loop_distribution): ...new function.
5310 (pass_vectorize): Replace declaration with that of...
5311 (make_pass_vectorize): ...new function.
5312 (pass_slp_vectorize): Replace declaration with that of...
5313 (make_pass_slp_vectorize): ...new function.
5314 (pass_complete_unroll): Replace declaration with that of...
5315 (make_pass_complete_unroll): ...new function.
5316 (pass_complete_unrolli): Replace declaration with that of...
5317 (make_pass_complete_unrolli): ...new function.
5318 (pass_parallelize_loops): Replace declaration with that of...
5319 (make_pass_parallelize_loops): ...new function.
5320 (pass_loop_prefetch): Replace declaration with that of...
5321 (make_pass_loop_prefetch): ...new function.
5322 (pass_iv_optimize): Replace declaration with that of...
5323 (make_pass_iv_optimize): ...new function.
5324 (pass_tree_loop_done): Replace declaration with that of...
5325 (make_pass_tree_loop_done): ...new function.
5326 (pass_ch): Replace declaration with that of...
5327 (make_pass_ch): ...new function.
5328 (pass_ccp): Replace declaration with that of...
5329 (make_pass_ccp): ...new function.
5330 (pass_phi_only_cprop): Replace declaration with that of...
5331 (make_pass_phi_only_cprop): ...new function.
5332 (pass_build_ssa): Replace declaration with that of...
5333 (make_pass_build_ssa): ...new function.
5334 (pass_build_alias): Replace declaration with that of...
5335 (make_pass_build_alias): ...new function.
5336 (pass_build_ealias): Replace declaration with that of...
5337 (make_pass_build_ealias): ...new function.
5338 (pass_dominator): Replace declaration with that of...
5339 (make_pass_dominator): ...new function.
5340 (pass_dce): Replace declaration with that of...
5341 (make_pass_dce): ...new function.
5342 (pass_dce_loop): Replace declaration with that of...
5343 (make_pass_dce_loop): ...new function.
5344 (pass_cd_dce): Replace declaration with that of...
5345 (make_pass_cd_dce): ...new function.
5346 (pass_call_cdce): Replace declaration with that of...
5347 (make_pass_call_cdce): ...new function.
5348 (pass_merge_phi): Replace declaration with that of...
5349 (make_pass_merge_phi): ...new function.
5350 (pass_split_crit_edges): Replace declaration with that of...
5351 (make_pass_split_crit_edges): ...new function.
5352 (pass_pre): Replace declaration with that of...
5353 (make_pass_pre): ...new function.
5354 (pass_profile): Replace declaration with that of...
5355 (make_pass_profile): ...new function.
5356 (pass_strip_predict_hints): Replace declaration with that of...
5357 (make_pass_strip_predict_hints): ...new function.
5358 (pass_lower_complex_O0): Replace declaration with that of...
5359 (make_pass_lower_complex_O0): ...new function.
5360 (pass_lower_complex): Replace declaration with that of...
5361 (make_pass_lower_complex): ...new function.
5362 (pass_lower_vector): Replace declaration with that of...
5363 (make_pass_lower_vector): ...new function.
5364 (pass_lower_vector_ssa): Replace declaration with that of...
5365 (make_pass_lower_vector_ssa): ...new function.
5366 (pass_lower_omp): Replace declaration with that of...
5367 (make_pass_lower_omp): ...new function.
5368 (pass_diagnose_omp_blocks): Replace declaration with that of...
5369 (make_pass_diagnose_omp_blocks): ...new function.
5370 (pass_expand_omp): Replace declaration with that of...
5371 (make_pass_expand_omp): ...new function.
5372 (pass_expand_omp_ssa): Replace declaration with that of...
5373 (make_pass_expand_omp_ssa): ...new function.
5374 (pass_object_sizes): Replace declaration with that of...
5375 (make_pass_object_sizes): ...new function.
5376 (pass_strlen): Replace declaration with that of...
5377 (make_pass_strlen): ...new function.
5378 (pass_fold_builtins): Replace declaration with that of...
5379 (make_pass_fold_builtins): ...new function.
5380 (pass_stdarg): Replace declaration with that of...
5381 (make_pass_stdarg): ...new function.
5382 (pass_early_warn_uninitialized): Replace declaration with that of...
5383 (make_pass_early_warn_uninitialized): ...new function.
5384 (pass_late_warn_uninitialized): Replace declaration with that of...
5385 (make_pass_late_warn_uninitialized): ...new function.
5386 (pass_cse_reciprocals): Replace declaration with that of...
5387 (make_pass_cse_reciprocals): ...new function.
5388 (pass_cse_sincos): Replace declaration with that of...
5389 (make_pass_cse_sincos): ...new function.
5390 (pass_optimize_bswap): Replace declaration with that of...
5391 (make_pass_optimize_bswap): ...new function.
5392 (pass_optimize_widening_mul): Replace declaration with that of...
5393 (make_pass_optimize_widening_mul): ...new function.
5394 (pass_warn_function_return): Replace declaration with that of...
5395 (make_pass_warn_function_return): ...new function.
5396 (pass_warn_function_noreturn): Replace declaration with that of...
5397 (make_pass_warn_function_noreturn): ...new function.
5398 (pass_cselim): Replace declaration with that of...
5399 (make_pass_cselim): ...new function.
5400 (pass_phiopt): Replace declaration with that of...
5401 (make_pass_phiopt): ...new function.
5402 (pass_forwprop): Replace declaration with that of...
5403 (make_pass_forwprop): ...new function.
5404 (pass_phiprop): Replace declaration with that of...
5405 (make_pass_phiprop): ...new function.
5406 (pass_tree_ifcombine): Replace declaration with that of...
5407 (make_pass_tree_ifcombine): ...new function.
5408 (pass_dse): Replace declaration with that of...
5409 (make_pass_dse): ...new function.
5410 (pass_nrv): Replace declaration with that of...
5411 (make_pass_nrv): ...new function.
5412 (pass_rename_ssa_copies): Replace declaration with that of...
5413 (make_pass_rename_ssa_copies): ...new function.
5414 (pass_sink_code): Replace declaration with that of...
5415 (make_pass_sink_code): ...new function.
5416 (pass_fre): Replace declaration with that of...
5417 (make_pass_fre): ...new function.
5418 (pass_check_data_deps): Replace declaration with that of...
5419 (make_pass_check_data_deps): ...new function.
5420 (pass_copy_prop): Replace declaration with that of...
5421 (make_pass_copy_prop): ...new function.
5422 (pass_vrp): Replace declaration with that of...
5423 (make_pass_vrp): ...new function.
5424 (pass_uncprop): Replace declaration with that of...
5425 (make_pass_uncprop): ...new function.
5426 (pass_return_slot): Replace declaration with that of...
5427 (make_pass_return_slot): ...new function.
5428 (pass_reassoc): Replace declaration with that of...
5429 (make_pass_reassoc): ...new function.
5430 (pass_rebuild_cgraph_edges): Replace declaration with that of...
5431 (make_pass_rebuild_cgraph_edges): ...new function.
5432 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
5433 (make_pass_remove_cgraph_callee_edges): ...new function.
5434 (pass_build_cgraph_edges): Replace declaration with that of...
5435 (make_pass_build_cgraph_edges): ...new function.
5436 (pass_local_pure_const): Replace declaration with that of...
5437 (make_pass_local_pure_const): ...new function.
5438 (pass_tracer): Replace declaration with that of...
5439 (make_pass_tracer): ...new function.
5440 (pass_warn_unused_result): Replace declaration with that of...
5441 (make_pass_warn_unused_result): ...new function.
5442 (pass_diagnose_tm_blocks): Replace declaration with that of...
5443 (make_pass_diagnose_tm_blocks): ...new function.
5444 (pass_lower_tm): Replace declaration with that of...
5445 (make_pass_lower_tm): ...new function.
5446 (pass_tm_init): Replace declaration with that of...
5447 (make_pass_tm_init): ...new function.
5448 (pass_tm_mark): Replace declaration with that of...
5449 (make_pass_tm_mark): ...new function.
5450 (pass_tm_memopt): Replace declaration with that of...
5451 (make_pass_tm_memopt): ...new function.
5452 (pass_tm_edges): Replace declaration with that of...
5453 (make_pass_tm_edges): ...new function.
5454 (pass_split_functions): Replace declaration with that of...
5455 (make_pass_split_functions): ...new function.
5456 (pass_feedback_split_functions): Replace declaration with that of...
5457 (make_pass_feedback_split_functions): ...new function.
5458 (pass_strength_reduction): Replace declaration with that of...
5459 (make_pass_strength_reduction): ...new function.
5460 (pass_ipa_lower_emutls): Replace declaration with that of...
5461 (make_pass_ipa_lower_emutls): ...new function.
5462 (pass_ipa_function_and_variable_visibility): Replace declaration with
5464 (make_pass_ipa_function_and_variable_visibility): ...new function.
5465 (pass_ipa_tree_profile): Replace declaration with that of...
5466 (make_pass_ipa_tree_profile): ...new function.
5467 (pass_early_local_passes): Replace declaration with that of...
5468 (make_pass_early_local_passes): ...new function.
5469 (pass_ipa_whole_program_visibility): Replace declaration with that
5471 (make_pass_ipa_whole_program_visibility): ...new function.
5472 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5473 (make_pass_ipa_lto_gimple_out): ...new function.
5474 (pass_ipa_increase_alignment): Replace declaration with that of...
5475 (make_pass_ipa_increase_alignment): ...new function.
5476 (pass_ipa_inline): Replace declaration with that of...
5477 (make_pass_ipa_inline): ...new function.
5478 (pass_ipa_free_lang_data): Replace declaration with that of...
5479 (make_pass_ipa_free_lang_data): ...new function.
5480 (pass_ipa_free_inline_summary): Replace declaration with that of...
5481 (make_pass_ipa_free_inline_summary): ...new function.
5482 (pass_ipa_cp): Replace declaration with that of...
5483 (make_pass_ipa_cp): ...new function.
5484 (pass_ipa_reference): Replace declaration with that of...
5485 (make_pass_ipa_reference): ...new function.
5486 (pass_ipa_pure_const): Replace declaration with that of...
5487 (make_pass_ipa_pure_const): ...new function.
5488 (pass_ipa_pta): Replace declaration with that of...
5489 (make_pass_ipa_pta): ...new function.
5490 (pass_ipa_lto_finish_out): Replace declaration with that of...
5491 (make_pass_ipa_lto_finish_out): ...new function.
5492 (pass_ipa_tm): Replace declaration with that of...
5493 (make_pass_ipa_tm): ...new function.
5494 (pass_ipa_profile): Replace declaration with that of...
5495 (make_pass_ipa_profile): ...new function.
5496 (pass_ipa_cdtor_merge): Replace declaration with that of...
5497 (make_pass_ipa_cdtor_merge): ...new function.
5498 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5500 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5501 (pass_init_datastructures): Replace declaration with that of...
5502 (make_pass_init_datastructures): ...new function.
5503 (pass_fixup_cfg): Replace declaration with that of...
5504 (make_pass_fixup_cfg): ...new function.
5505 (pass_expand): Replace declaration with that of...
5506 (make_pass_expand): ...new function.
5507 (pass_instantiate_virtual_regs): Replace declaration with that of...
5508 (make_pass_instantiate_virtual_regs): ...new function.
5509 (pass_rtl_fwprop): Replace declaration with that of...
5510 (make_pass_rtl_fwprop): ...new function.
5511 (pass_rtl_fwprop_addr): Replace declaration with that of...
5512 (make_pass_rtl_fwprop_addr): ...new function.
5513 (pass_jump): Replace declaration with that of...
5514 (make_pass_jump): ...new function.
5515 (pass_jump2): Replace declaration with that of...
5516 (make_pass_jump2): ...new function.
5517 (pass_lower_subreg): Replace declaration with that of...
5518 (make_pass_lower_subreg): ...new function.
5519 (pass_cse): Replace declaration with that of...
5520 (make_pass_cse): ...new function.
5521 (pass_fast_rtl_dce): Replace declaration with that of...
5522 (make_pass_fast_rtl_dce): ...new function.
5523 (pass_ud_rtl_dce): Replace declaration with that of...
5524 (make_pass_ud_rtl_dce): ...new function.
5525 (pass_rtl_dce): Replace declaration with that of...
5526 (make_pass_rtl_dce): ...new function.
5527 (pass_rtl_dse1): Replace declaration with that of...
5528 (make_pass_rtl_dse1): ...new function.
5529 (pass_rtl_dse2): Replace declaration with that of...
5530 (make_pass_rtl_dse2): ...new function.
5531 (pass_rtl_dse3): Replace declaration with that of...
5532 (make_pass_rtl_dse3): ...new function.
5533 (pass_rtl_cprop): Replace declaration with that of...
5534 (make_pass_rtl_cprop): ...new function.
5535 (pass_rtl_pre): Replace declaration with that of...
5536 (make_pass_rtl_pre): ...new function.
5537 (pass_rtl_hoist): Replace declaration with that of...
5538 (make_pass_rtl_hoist): ...new function.
5539 (pass_rtl_store_motion): Replace declaration with that of...
5540 (make_pass_rtl_store_motion): ...new function.
5541 (pass_cse_after_global_opts): Replace declaration with that of...
5542 (make_pass_cse_after_global_opts): ...new function.
5543 (pass_rtl_ifcvt): Replace declaration with that of...
5544 (make_pass_rtl_ifcvt): ...new function.
5545 (pass_into_cfg_layout_mode): Replace declaration with that of...
5546 (make_pass_into_cfg_layout_mode): ...new function.
5547 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5548 (make_pass_outof_cfg_layout_mode): ...new function.
5549 (pass_loop2): Replace declaration with that of...
5550 (make_pass_loop2): ...new function.
5551 (pass_rtl_loop_init): Replace declaration with that of...
5552 (make_pass_rtl_loop_init): ...new function.
5553 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5554 (make_pass_rtl_move_loop_invariants): ...new function.
5555 (pass_rtl_unswitch): Replace declaration with that of...
5556 (make_pass_rtl_unswitch): ...new function.
5557 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5558 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5559 (pass_rtl_doloop): Replace declaration with that of...
5560 (make_pass_rtl_doloop): ...new function.
5561 (pass_rtl_loop_done): Replace declaration with that of...
5562 (make_pass_rtl_loop_done): ...new function.
5563 (pass_web): Replace declaration with that of...
5564 (make_pass_web): ...new function.
5565 (pass_cse2): Replace declaration with that of...
5566 (make_pass_cse2): ...new function.
5567 (pass_df_initialize_opt): Replace declaration with that of...
5568 (make_pass_df_initialize_opt): ...new function.
5569 (pass_df_initialize_no_opt): Replace declaration with that of...
5570 (make_pass_df_initialize_no_opt): ...new function.
5571 (pass_reginfo_init): Replace declaration with that of...
5572 (make_pass_reginfo_init): ...new function.
5573 (pass_inc_dec): Replace declaration with that of...
5574 (make_pass_inc_dec): ...new function.
5575 (pass_stack_ptr_mod): Replace declaration with that of...
5576 (make_pass_stack_ptr_mod): ...new function.
5577 (pass_initialize_regs): Replace declaration with that of...
5578 (make_pass_initialize_regs): ...new function.
5579 (pass_combine): Replace declaration with that of...
5580 (make_pass_combine): ...new function.
5581 (pass_if_after_combine): Replace declaration with that of...
5582 (make_pass_if_after_combine): ...new function.
5583 (pass_ree): Replace declaration with that of...
5584 (make_pass_ree): ...new function.
5585 (pass_partition_blocks): Replace declaration with that of...
5586 (make_pass_partition_blocks): ...new function.
5587 (pass_match_asm_constraints): Replace declaration with that of...
5588 (make_pass_match_asm_constraints): ...new function.
5589 (pass_regmove): Replace declaration with that of...
5590 (make_pass_regmove): ...new function.
5591 (pass_split_all_insns): Replace declaration with that of...
5592 (make_pass_split_all_insns): ...new function.
5593 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5594 (make_pass_fast_rtl_byte_dce): ...new function.
5595 (pass_lower_subreg2): Replace declaration with that of...
5596 (make_pass_lower_subreg2): ...new function.
5597 (pass_mode_switching): Replace declaration with that of...
5598 (make_pass_mode_switching): ...new function.
5599 (pass_sms): Replace declaration with that of...
5600 (make_pass_sms): ...new function.
5601 (pass_sched): Replace declaration with that of...
5602 (make_pass_sched): ...new function.
5603 (pass_ira): Replace declaration with that of...
5604 (make_pass_ira): ...new function.
5605 (pass_reload): Replace declaration with that of...
5606 (make_pass_reload): ...new function.
5607 (pass_clean_state): Replace declaration with that of...
5608 (make_pass_clean_state): ...new function.
5609 (pass_branch_prob): Replace declaration with that of...
5610 (make_pass_branch_prob): ...new function.
5611 (pass_value_profile_transformations): Replace declaration with that
5613 (make_pass_value_profile_transformations): ...new function.
5614 (pass_postreload_cse): Replace declaration with that of...
5615 (make_pass_postreload_cse): ...new function.
5616 (pass_gcse2): Replace declaration with that of...
5617 (make_pass_gcse2): ...new function.
5618 (pass_split_after_reload): Replace declaration with that of...
5619 (make_pass_split_after_reload): ...new function.
5620 (pass_branch_target_load_optimize1): Replace declaration with that
5622 (make_pass_branch_target_load_optimize1): ...new function.
5623 (pass_thread_prologue_and_epilogue): Replace declaration with that
5625 (make_pass_thread_prologue_and_epilogue): ...new function.
5626 (pass_stack_adjustments): Replace declaration with that of...
5627 (make_pass_stack_adjustments): ...new function.
5628 (pass_peephole2): Replace declaration with that of...
5629 (make_pass_peephole2): ...new function.
5630 (pass_if_after_reload): Replace declaration with that of...
5631 (make_pass_if_after_reload): ...new function.
5632 (pass_regrename): Replace declaration with that of...
5633 (make_pass_regrename): ...new function.
5634 (pass_cprop_hardreg): Replace declaration with that of...
5635 (make_pass_cprop_hardreg): ...new function.
5636 (pass_reorder_blocks): Replace declaration with that of...
5637 (make_pass_reorder_blocks): ...new function.
5638 (pass_branch_target_load_optimize2): Replace declaration with that
5640 (make_pass_branch_target_load_optimize2): ...new function.
5641 (pass_leaf_regs): Replace declaration with that of...
5642 (make_pass_leaf_regs): ...new function.
5643 (pass_split_before_sched2): Replace declaration with that of...
5644 (make_pass_split_before_sched2): ...new function.
5645 (pass_compare_elim_after_reload): Replace declaration with that of...
5646 (make_pass_compare_elim_after_reload): ...new function.
5647 (pass_sched2): Replace declaration with that of...
5648 (make_pass_sched2): ...new function.
5649 (pass_stack_regs): Replace declaration with that of...
5650 (make_pass_stack_regs): ...new function.
5651 (pass_stack_regs_run): Replace declaration with that of...
5652 (make_pass_stack_regs_run): ...new function.
5653 (pass_df_finish): Replace declaration with that of...
5654 (make_pass_df_finish): ...new function.
5655 (pass_compute_alignments): Replace declaration with that of...
5656 (make_pass_compute_alignments): ...new function.
5657 (pass_duplicate_computed_gotos): Replace declaration with that of...
5658 (make_pass_duplicate_computed_gotos): ...new function.
5659 (pass_variable_tracking): Replace declaration with that of...
5660 (make_pass_variable_tracking): ...new function.
5661 (pass_free_cfg): Replace declaration with that of...
5662 (make_pass_free_cfg): ...new function.
5663 (pass_machine_reorg): Replace declaration with that of...
5664 (make_pass_machine_reorg): ...new function.
5665 (pass_cleanup_barriers): Replace declaration with that of...
5666 (make_pass_cleanup_barriers): ...new function.
5667 (pass_delay_slots): Replace declaration with that of...
5668 (make_pass_delay_slots): ...new function.
5669 (pass_split_for_shorten_branches): Replace declaration with that of...
5670 (make_pass_split_for_shorten_branches): ...new function.
5671 (pass_split_before_regstack): Replace declaration with that of...
5672 (make_pass_split_before_regstack): ...new function.
5673 (pass_convert_to_eh_region_ranges): Replace declaration with that
5675 (make_pass_convert_to_eh_region_ranges): ...new function.
5676 (pass_shorten_branches): Replace declaration with that of...
5677 (make_pass_shorten_branches): ...new function.
5678 (pass_set_nothrow_function_flags): Replace declaration with that of...
5679 (make_pass_set_nothrow_function_flags): ...new function.
5680 (pass_dwarf2_frame): Replace declaration with that of...
5681 (make_pass_dwarf2_frame): ...new function.
5682 (pass_final): Replace declaration with that of...
5683 (make_pass_final): ...new function.
5684 (pass_rtl_seqabstr): Replace declaration with that of...
5685 (make_pass_rtl_seqabstr): ...new function.
5686 (pass_release_ssa_names): Replace declaration with that of...
5687 (make_pass_release_ssa_names): ...new function.
5688 (pass_early_inline): Replace declaration with that of...
5689 (make_pass_early_inline): ...new function.
5690 (pass_inline_parameters): Replace declaration with that of...
5691 (make_pass_inline_parameters): ...new function.
5692 (pass_update_address_taken): Replace declaration with that of...
5693 (make_pass_update_address_taken): ...new function.
5694 (pass_convert_switch): Replace declaration with that of...
5695 (make_pass_convert_switch): ...new function.
5696 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5697 to a subclass of simple_ipa_opt_pass along with...
5698 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5699 (make_pass_ipa_tree_profile): ...new function.
5700 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5701 subclass of gimple_opt_pass along with...
5702 (pass_data_sra_early): ...new pass_data instance and...
5703 (make_pass_sra_early): ...new function.
5704 (pass_sra): Convert from a global struct to a subclass of
5705 gimple_opt_pass along with...
5706 (pass_data_sra): ...new pass_data instance and...
5707 (make_pass_sra): ...new function.
5708 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5709 gimple_opt_pass along with...
5710 (pass_data_early_ipa_sra): ...new pass_data instance and...
5711 (make_pass_early_ipa_sra): ...new function.
5712 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5713 subclass of gimple_opt_pass along with...
5714 (pass_data_ccp): ...new pass_data instance and...
5715 (make_pass_ccp): ...new function.
5716 (pass_fold_builtins): Convert from a global struct to a subclass of
5717 gimple_opt_pass along with...
5718 (pass_data_fold_builtins): ...new pass_data instance and...
5719 (make_pass_fold_builtins): ...new function.
5720 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5721 subclass of gimple_opt_pass along with...
5722 (pass_data_copy_prop): ...new pass_data instance and...
5723 (make_pass_copy_prop): ...new function.
5724 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5725 global struct to a subclass of gimple_opt_pass along with...
5726 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5727 (make_pass_rename_ssa_copies): ...new function.
5728 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5729 subclass of gimple_opt_pass along with...
5730 (pass_data_dce): ...new pass_data instance and...
5731 (make_pass_dce): ...new function.
5732 (pass_dce_loop): Convert from a global struct to a subclass of
5733 gimple_opt_pass along with...
5734 (pass_data_dce_loop): ...new pass_data instance and...
5735 (make_pass_dce_loop): ...new function.
5736 (pass_cd_dce): Convert from a global struct to a subclass of
5737 gimple_opt_pass along with...
5738 (pass_data_cd_dce): ...new pass_data instance and...
5739 (make_pass_cd_dce): ...new function.
5740 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5741 subclass of gimple_opt_pass along with...
5742 (pass_data_dominator): ...new pass_data instance and...
5743 (make_pass_dominator): ...new function.
5744 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5745 gimple_opt_pass along with...
5746 (pass_data_phi_only_cprop): ...new pass_data instance and...
5747 (make_pass_phi_only_cprop): ...new function.
5748 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5749 subclass of gimple_opt_pass along with...
5750 (pass_data_dse): ...new pass_data instance and...
5751 (make_pass_dse): ...new function.
5752 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5753 a subclass of gimple_opt_pass along with...
5754 (pass_data_forwprop): ...new pass_data instance and...
5755 (make_pass_forwprop): ...new function.
5756 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5757 struct to a subclass of gimple_opt_pass along with...
5758 (pass_data_tree_ifcombine): ...new pass_data instance and...
5759 (make_pass_tree_ifcombine): ...new function.
5760 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5761 subclass of gimple_opt_pass along with...
5762 (pass_data_ch): ...new pass_data instance and...
5763 (make_pass_ch): ...new function.
5764 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5765 subclass of gimple_opt_pass along with...
5766 (pass_data_tree_loop): ...new pass_data instance and...
5767 (make_pass_tree_loop): ...new function.
5768 (pass_tree_loop_init): Convert from a global struct to a subclass of
5769 gimple_opt_pass along with...
5770 (pass_data_tree_loop_init): ...new pass_data instance and...
5771 (make_pass_tree_loop_init): ...new function.
5772 (pass_lim): Convert from a global struct to a subclass of
5773 gimple_opt_pass along with...
5774 (pass_data_lim): ...new pass_data instance and...
5775 (make_pass_lim): ...new function.
5776 (pass_tree_unswitch): Convert from a global struct to a subclass of
5777 gimple_opt_pass along with...
5778 (pass_data_tree_unswitch): ...new pass_data instance and...
5779 (make_pass_tree_unswitch): ...new function.
5780 (pass_predcom): Convert from a global struct to a subclass of
5781 gimple_opt_pass along with...
5782 (pass_data_predcom): ...new pass_data instance and...
5783 (make_pass_predcom): ...new function.
5784 (pass_vectorize): Convert from a global struct to a subclass of
5785 gimple_opt_pass along with...
5786 (pass_data_vectorize): ...new pass_data instance and...
5787 (make_pass_vectorize): ...new function.
5788 (pass_graphite): Convert from a global struct to a subclass of
5789 gimple_opt_pass along with...
5790 (pass_data_graphite): ...new pass_data instance and...
5791 (make_pass_graphite): ...new function.
5792 (pass_graphite_transforms): Convert from a global struct to a subclass
5793 of gimple_opt_pass along with...
5794 (pass_data_graphite_transforms): ...new pass_data instance and...
5795 (make_pass_graphite_transforms): ...new function.
5796 (pass_check_data_deps): Convert from a global struct to a subclass of
5797 gimple_opt_pass along with...
5798 (pass_data_check_data_deps): ...new pass_data instance and...
5799 (make_pass_check_data_deps): ...new function.
5800 (pass_iv_canon): Convert from a global struct to a subclass of
5801 gimple_opt_pass along with...
5802 (pass_data_iv_canon): ...new pass_data instance and...
5803 (make_pass_iv_canon): ...new function.
5804 (pass_scev_cprop): Convert from a global struct to a subclass of
5805 gimple_opt_pass along with...
5806 (pass_data_scev_cprop): ...new pass_data instance and...
5807 (make_pass_scev_cprop): ...new function.
5808 (pass_record_bounds): Convert from a global struct to a subclass of
5809 gimple_opt_pass along with...
5810 (pass_data_record_bounds): ...new pass_data instance and...
5811 (make_pass_record_bounds): ...new function.
5812 (pass_complete_unroll): Convert from a global struct to a subclass of
5813 gimple_opt_pass along with...
5814 (pass_data_complete_unroll): ...new pass_data instance and...
5815 (make_pass_complete_unroll): ...new function.
5816 (pass_complete_unrolli): Convert from a global struct to a subclass of
5817 gimple_opt_pass along with...
5818 (pass_data_complete_unrolli): ...new pass_data instance and...
5819 (make_pass_complete_unrolli): ...new function.
5820 (pass_parallelize_loops): Convert from a global struct to a subclass
5821 of gimple_opt_pass along with...
5822 (pass_data_parallelize_loops): ...new pass_data instance and...
5823 (make_pass_parallelize_loops): ...new function.
5824 (pass_loop_prefetch): Convert from a global struct to a subclass of
5825 gimple_opt_pass along with...
5826 (pass_data_loop_prefetch): ...new pass_data instance and...
5827 (make_pass_loop_prefetch): ...new function.
5828 (pass_iv_optimize): Convert from a global struct to a subclass of
5829 gimple_opt_pass along with...
5830 (pass_data_iv_optimize): ...new pass_data instance and...
5831 (make_pass_iv_optimize): ...new function.
5832 (pass_tree_loop_done): Convert from a global struct to a subclass of
5833 gimple_opt_pass along with...
5834 (pass_data_tree_loop_done): ...new pass_data instance and...
5835 (make_pass_tree_loop_done): ...new function.
5836 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5837 struct to a subclass of gimple_opt_pass along with...
5838 (pass_data_cse_reciprocals): ...new pass_data instance and...
5839 (make_pass_cse_reciprocals): ...new function.
5840 (pass_cse_sincos): Convert from a global struct to a subclass of
5841 gimple_opt_pass along with...
5842 (pass_data_cse_sincos): ...new pass_data instance and...
5843 (make_pass_cse_sincos): ...new function.
5844 (pass_optimize_bswap): Convert from a global struct to a subclass of
5845 gimple_opt_pass along with...
5846 (pass_data_optimize_bswap): ...new pass_data instance and...
5847 (make_pass_optimize_bswap): ...new function.
5848 (pass_optimize_widening_mul): Convert from a global struct to a
5849 subclass of gimple_opt_pass along with...
5850 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5851 (make_pass_optimize_widening_mul): ...new function.
5852 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5853 subclass of gimple_opt_pass along with...
5854 (pass_data_phiopt): ...new pass_data instance and...
5855 (make_pass_phiopt): ...new function.
5856 (pass_cselim): Convert from a global struct to a subclass of
5857 gimple_opt_pass along with...
5858 (pass_data_cselim): ...new pass_data instance and...
5859 (make_pass_cselim): ...new function.
5860 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5861 subclass of gimple_opt_pass along with...
5862 (pass_data_phiprop): ...new pass_data instance and...
5863 (make_pass_phiprop): ...new function.
5864 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5865 subclass of gimple_opt_pass along with...
5866 (pass_data_pre): ...new pass_data instance and...
5867 (make_pass_pre): ...new function.
5868 (pass_fre): Convert from a global struct to a subclass of
5869 gimple_opt_pass along with...
5870 (pass_data_fre): ...new pass_data instance and...
5871 (make_pass_fre): ...new function.
5872 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5873 subclass of gimple_opt_pass along with...
5874 (pass_data_reassoc): ...new pass_data instance and...
5875 (make_pass_reassoc): ...new function.
5876 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5877 subclass of gimple_opt_pass along with...
5878 (pass_data_sink_code): ...new pass_data instance and...
5879 (make_pass_sink_code): ...new function.
5880 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5881 subclass of gimple_opt_pass along with...
5882 (pass_data_strlen): ...new pass_data instance and...
5883 (make_pass_strlen): ...new function.
5884 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5885 struct to a subclass of gimple_opt_pass along with...
5886 (pass_data_build_alias): ...new pass_data instance and...
5887 (make_pass_build_alias): ...new function.
5888 (pass_build_ealias): Convert from a global struct to a subclass of
5889 gimple_opt_pass along with...
5890 (pass_data_build_ealias): ...new pass_data instance and...
5891 (make_pass_build_ealias): ...new function.
5892 (pass_ipa_pta): Convert from a global struct to a subclass of
5893 simple_ipa_opt_pass along with...
5894 (pass_data_ipa_pta): ...new pass_data instance and...
5895 (make_pass_ipa_pta): ...new function.
5896 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5897 subclass of gimple_opt_pass along with...
5898 (pass_data_uncprop): ...new pass_data instance and...
5899 (make_pass_uncprop): ...new function.
5900 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5901 global struct to a subclass of gimple_opt_pass along with...
5902 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5903 (make_pass_late_warn_uninitialized): ...new function.
5904 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5905 to a subclass of gimple_opt_pass along with...
5906 (pass_data_init_datastructures): ...new pass_data instance and...
5907 (make_pass_init_datastructures): ...new function.
5908 (pass_early_warn_uninitialized): Convert from a global struct to a
5909 subclass of gimple_opt_pass along with...
5910 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5911 (make_pass_early_warn_uninitialized): ...new function.
5912 (pass_update_address_taken): Convert from a global struct to a
5913 subclass of gimple_opt_pass along with...
5914 (pass_data_update_address_taken): ...new pass_data instance and...
5915 (make_pass_update_address_taken): ...new function.
5916 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5917 struct to a subclass of gimple_opt_pass along with...
5918 (pass_data_release_ssa_names): ...new pass_data instance and...
5919 (make_pass_release_ssa_names): ...new function.
5920 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5921 subclass of gimple_opt_pass along with...
5922 (pass_data_stdarg): ...new pass_data instance and...
5923 (make_pass_stdarg): ...new function.
5924 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5925 global struct to a subclass of gimple_opt_pass along with...
5926 (pass_data_convert_switch): ...new pass_data instance and...
5927 (make_pass_convert_switch): ...new function.
5928 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5929 to a subclass of gimple_opt_pass along with...
5930 (pass_data_tail_recursion): ...new pass_data instance and...
5931 (make_pass_tail_recursion): ...new function.
5932 (pass_tail_calls): Convert from a global struct to a subclass of
5933 gimple_opt_pass along with...
5934 (pass_data_tail_calls): ...new pass_data instance and...
5935 (make_pass_tail_calls): ...new function.
5936 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5937 struct to a subclass of gimple_opt_pass along with...
5938 (pass_data_lower_vector): ...new pass_data instance and...
5939 (make_pass_lower_vector): ...new function.
5940 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5941 gimple_opt_pass along with...
5942 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5943 (make_pass_lower_vector_ssa): ...new function.
5944 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5945 to a subclass of gimple_opt_pass along with...
5946 (pass_data_slp_vectorize): ...new pass_data instance and...
5947 (make_pass_slp_vectorize): ...new function.
5948 (pass_ipa_increase_alignment): Convert from a global struct to a
5949 subclass of simple_ipa_opt_pass along with...
5950 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5951 (make_pass_ipa_increase_alignment): ...new function.
5952 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5953 gimple_opt_pass along with...
5954 (pass_data_vrp): ...new pass_data instance and...
5955 (make_pass_vrp): ...new function.
5956 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5957 subclass of simple_ipa_opt_pass along with...
5958 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5959 (make_pass_ipa_free_lang_data): ...new function.
5960 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5961 gimple_opt_pass along with...
5962 (pass_data_tsan): ...new pass_data instance and...
5963 (make_pass_tsan): ...new function.
5964 (pass_tsan_O0): Convert from a global struct to a subclass of
5965 gimple_opt_pass along with...
5966 (pass_data_tsan_O0): ...new pass_data instance and...
5967 (make_pass_tsan_O0): ...new function.
5968 * var-tracking.c (pass_variable_tracking): Convert from a global
5969 struct to a subclass of rtl_opt_pass along with...
5970 (pass_data_variable_tracking): ...new pass_data instance and...
5971 (make_pass_variable_tracking): ...new function.
5972 * web.c (pass_web): Convert from a global struct to a subclass of
5973 rtl_opt_pass along with...
5974 (pass_data_web): ...new pass_data instance and...
5975 (make_pass_web): ...new function.
5976 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5977 declaration with that of...
5978 (make_pass_mode_switch_use): ...new function.
5979 (pass_resolve_sw_modes): Replace declaration with that of...
5980 (make_pass_resolve_sw_modes): ...new function.
5981 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5982 from a global struct to a subclass of rtl_opt_pass along with...
5983 (pass_data_mode_switch_use): ...new pass_data instance and...
5984 (make_pass_mode_switch_use): ...new function.
5985 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5986 from a global struct to a subclass of rtl_opt_pass along with...
5987 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5988 (make_pass_resolve_sw_modes): ...new function.
5989 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5990 struct to a subclass of rtl_opt_pass along with...
5991 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5992 (make_pass_insert_vzeroupper): ...new function.
5993 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5994 global struct to a subclass of rtl_opt_pass along with...
5995 (pass_data_work_around_errata): ...new pass_data instance and...
5996 (make_pass_work_around_errata): ...new function.
5997 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5998 struct to a subclass of rtl_opt_pass along with...
5999 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
6000 (make_pass_mips_machine_reorg2): ...new function.
6002 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6004 * passes.c (pass_manager::operator new): New.
6006 2013-08-05 David Malcolm <dmalcolm@redhat.com>
6008 Handwritten part of conversion of passes to C++ classes.
6010 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
6011 (toplev.o): Add dep on PASS_MANAGER_H.
6012 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
6013 of early local pases to reflect this moving from a global to a
6014 member of gcc::pass_manager.
6015 (cgraph_add_new_function): Likewise.
6016 * lto-cgraph.c (lto_output_node): Update for conversion of
6017 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
6018 * passes.c (opt_pass::clone): New.
6019 (opt_pass::gate): New.
6020 (opt_pass::execute): New.
6021 (opt_pass::opt_pass): New.
6022 (pass_manager::execute_early_local_passes): New.
6023 (pass_manager::execute_pass_mode_switching): new.
6024 (finish_optimization_passes): Convert to...
6025 (pass_manager::finish_optimization_passes): ...this.
6026 (finish_optimization_passes): Update for conversion of passes to
6028 (register_dump_files_1): Use has_gate since we cannot portably
6029 check a vtable entry against NULL.
6030 (dump_one_pass): Likewise.
6031 (ipa_write_summaries_2): Likewise.
6032 (ipa_write_optimization_summaries_1): Likewise.
6033 (ipa_read_summaries_1): Likewise.
6034 (ipa_read_optimization_summaries_1): Likewise.
6035 (execute_ipa_stmt_fixups): Likewise.
6036 (pass_manager::pass_manager): Rewrite pass-creation, invoking
6037 pass-creation functions rather than wiring up globals, and
6038 storing the results in fields of pass_manager generated using
6040 (pass_manager::dump_profile_report): Update for conversion of
6041 passes to C++ classes.
6042 (pass_manager::execute_ipa_summary_passes): Likewise.
6043 (execute_one_ipa_transform_pass): Likewise.
6044 (execute_one_pass): Use has_gate and has_execute since we cannot
6045 portably check a vtable entry against NULL.
6046 * pass_manager.h (pass_manager::finish_optimization_passes): New.
6047 (pass_manager): Use pass-instances.def to add fields for the
6048 various pass instances.
6049 * toplev.c (finalize): Update for move of
6050 finish_optimization_passes to a method of gcc::pass_manager.
6051 * toplev.h (finish_optimization_passes): Move to method of class
6053 * tree-pass.h (struct pass_data): New.
6054 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
6055 (opt_pass::gate): Convert to virtual function.
6056 (opt_pass::~opt_pass): New.
6057 (opt_pass::clone): New.
6058 (opt_pass::execute): Convert to virtual function.
6059 (opt_pass::opt_pass): New.
6060 (opt_pass::ctxt_): new.
6061 (gimple_opt_pass): Convert to subclass of opt_pass.
6062 (gimple_opt_pass::gimple_opt_pass): New.
6063 (rtl_opt_pass): Convert to subclass of opt_pass.
6064 (rtl_opt_pass::rtl_opt_pass): New.
6065 (ipa_opt_pass_d): Convert to subclass of opt_pass.
6066 (ipa_opt_pass_d::ipa_opt_pass_d): New.
6067 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
6068 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
6069 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
6070 invocation of pass_mode_switching to reflect this moving from a
6071 global to a member of gcc::pass_manager.
6072 (ix86_option_override): Rework how pass_insert_vzeroupper is
6073 added to the pass_manager to reflect autogenerated changes.
6074 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
6076 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
6078 PR rtl-optimization/57708
6079 * recog.c (peep2_find_free_register): Validate all regs in a
6082 2013-08-05 Jan Hubicka <jh@suse.cz>
6085 * cgraph.c (verify_cgraph_node): Accept local flags from other
6087 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
6088 (function_and_variable_visibility): Likewise.
6089 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
6091 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6093 * graph.c (init_graph_slim_pretty_print): Remove.
6094 (print_graph_cfg): Do not call it. Use local pretty printer.
6095 (start_graph_dump): Likewise.
6097 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6099 * gimple-pretty-print.c (buffer): Remove.
6100 (initialized): Likewise.
6101 (maybe_init_pretty_print): Likewise.
6102 (print_gimple_stmt): Do not call it. Use non-static local
6103 pretty_printer variable.
6104 (print_gimple_expr): Likewise.
6105 (print_gimple_seq): Likewise.
6106 (gimple_dump_bb): Likewise.
6108 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
6110 * asan.c (asan_pp): Remove.
6111 (asan_pp_initialized): Likewise.
6112 (asan_pp_initialize): Likewise.
6113 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
6114 (asan_emit_stack_protection): Tidy. Use local pretty printer.
6115 (asan_add_global): Likewise.
6117 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
6119 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
6120 * diagnostic.h (diagnostic_flush_buffer): Adjust.
6121 * pretty-print.c (pp_formatted_text_data): Likewise.
6122 (pp_indent): Rename from pp_base_indent.
6123 (pp_format): Rename from pp_base_format.
6124 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
6125 (pp_format_verbatim): Rename from pp_base_format_verbatim.
6126 (pp_flush): Rename from pp_base_flush.
6127 (pp_set_line_maximum_length): Rename from
6128 pp_base_set_line_maximum_length.
6129 (pp_clear_output_area): Rename from pp_base_clear_output_area.
6130 (pp_set_prefix): Rename from pp_base_set_prefix.
6131 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
6132 (pp_emit_prefix): Rename from pp_base_emit_prefix.
6133 (pp_append_text): Rename from pp_base_append_text.
6134 (pp_formatted_text): Rename from pp_base_formatted_text.
6135 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
6136 (pp_remaining_character_count_for_line): Rename from
6137 pp_base_remaining_character_count_for_line.
6138 (pp_newline): Rename from pp_base_newline.
6139 (pp_character): Rename from pp_base_character.
6140 (pp_string): Rename from pp_base_string.
6141 (pp_maybe_space): Rename from pp_base_maybe_space.
6142 * asan.c (asan_pp_string): Adjust.
6143 (asan_emit_stack_protection): Likewise.
6144 (asan_add_global): Likewise.
6145 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
6146 * tree-mudflap.c (mf_varname_tree): Likewise.
6147 * tree-pretty-print.c (pp_tree_identifier): Rename from
6148 pp_base_tree_identifier.
6149 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
6150 Declare as function.
6152 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6154 * pretty-print.h (pp_bar_bar): New.
6155 (pp_ampersand_ampersand): Likewise.
6156 (pp_less_equal): Likewise.
6157 (pp_greater_equal): Likewise.
6158 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
6159 printer functions instead of pp_string or operators and punctuators.
6160 (dump_gimple_call): Likewise.
6161 (dump_gimple_omp_for): Likewise.
6162 (dump_gimple_transaction): Likewise.
6163 (dump_gimple_phi): Likewise.
6164 (pp_gimple_stmt_1): Likewise.
6165 * sched-vis.c (print_insn): Likewise.
6166 * tree-mudflap.c (mf_varname_tree): Likewise.
6167 * tree-pretty-print.c (dump_block_node): Likewise.
6168 (dump_generic_node): Likewise.
6170 2013-08-02 Jan Hubicka <jh@suse.cz>
6172 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
6174 * lto-streamer-out.c (tree_is_indexable): Results decls and
6175 parm decls are not indexable.
6176 (DFS_write_tree_body): Do not follow args and results.
6177 (hash_tree): Likewise.
6178 (output_functions): Rearrange so struct function is needed
6179 only when real body is output; be able to also ouptut abstract
6180 functions; output DECL_ARGUMENTS and DECL_RESULT.
6181 (lto_output): When not in WPA, ale store abstract functions.
6182 (write_symbol): Do not care about RESULT_DECL.
6183 (output_symbol_p): Handle correctly sbtract decls.
6184 * lto-streamer-in.c (input_function): Rearrange so struct
6185 function can be NULL at entry; allow streaming of
6186 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
6187 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
6188 sanity check during LTO.
6189 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
6190 RESULT_DECl and DECL_ARGUMENTS.
6191 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
6194 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
6196 * pretty-print.h (pp_underscore): New.
6198 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
6199 printer functions instead of pp_character.
6200 (dump_binary_rhs): Likewise.
6201 (dump_ternary_rhs): Likewise.
6202 (dump_gimple_call_args): Likewise.
6203 (pp_points_to_solution): Likewise.
6204 (dump_gimple_call): Likewise.
6205 (dump_gimple_switch): Likewise.
6206 (dump_gimple_cond): Likewise.
6207 (dump_gimple_bind): Likewise.
6208 (dump_gimple_try): Likewise.
6209 (dump_gimple_omp_for): Likewise.
6210 (dump_gimple_omp_continue): Likewise.
6211 (dump_gimple_omp_single): Likewise.
6212 (dump_gimple_omp_sections): Likewise.
6213 (dump_gimple_omp_block): Likewise.
6214 (dump_gimple_omp_critical): Likewise.
6215 (dump_gimple_transaction): Likewise.
6216 (dump_gimple_asm): Likewise.
6217 (dump_gimple_phi): Likewise.
6218 (dump_gimple_omp_parallel): Likewise.
6219 (dump_gimple_omp_task): Likewise.
6220 (dump_gimple_omp_atomic_load): Likewise.
6221 (dump_gimple_omp_atomic_store): Likewise.
6222 (dump_gimple_mem_ops): Likewise.
6223 (pp_gimple_stmt_1): Likewise.
6224 (pp_cfg_jump): Likewise.
6225 (dump_implicit_edges): Likewise.
6226 (gimple_dump_bb_for_graph): Likewise.
6227 * graph.c (draw_cfg_node): Likewise.
6228 * langhooks.c (lhd_print_error_function): Likewise.
6229 * sched-vis.c (print_exp): Likewise.
6230 (print_value): Likewise.
6231 (print_pattern): Likewise.
6232 (print_insn): Likewise.
6233 (rtl_dump_bb_for_graph): Likewise.
6234 * tree-pretty-print.c (dump_function_declaration): Likewise.
6235 (dump_array_domain): Likewise.
6236 (dump_omp_clause): Likewise.
6237 (dump_location): Likewise.
6238 (dump_generic_node): Likewise.
6239 (print_struct_decl): Likewise.
6240 * diagnostic.c (diagnostic_show_locus): Use pp_space.
6242 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6244 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
6245 candidate table when replacing a candidate statement.
6246 (replace_rhs_if_not_dup): Likewise.
6247 (replace_one_candidate): Likewise.
6249 2013-08-02 Jan Hubicka <jh@suse.cz>
6250 Martin Liska <marxin.liska@gmail.com>
6252 * cgraphunit.c (add_new_function): Fix logic when adding from
6254 (assemble_thunk): Rename to ...
6255 (expand_thunk); .. this one; export; get it working with
6256 general functions; make produced gimple valid.
6257 * cgraph.h (expand_thunk): Declare.
6259 2013-08-02 Jan Hubicka <jh@suse.cz>
6261 * ipa-cp.c (gather_context_independent_values): Use
6262 ipa_get_param_move_cost.
6263 (get_replacement_map): Remove PARAM; move parameter folding
6265 (create_specialized_node): Update.
6266 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
6267 assert that we have gimple body; update move_cost.
6268 (count_formal_params): Assert that we have gimple body.
6269 (ipa_dump_param): New function.
6270 (ipa_alloc_node_params): Break out from ...
6271 (ipa_initialize_node_params): ... here.
6272 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
6273 (ipa_write_node_info): Stream move costs.
6274 (ipa_read_node_info): Read move costs.
6275 (ipa_update_after_lto_read): Do not recompute node params.
6276 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
6277 (ipa_get_param): Check we are not in WPA.
6278 (ipa_get_param_move_cost): New.
6279 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
6280 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
6281 parm numbers to be present.
6283 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6285 PR rtl-optimization/58048
6286 * lra-constraints.c (process_alt_operands): Don't check asm
6287 operand on register.
6289 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
6291 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
6292 the implied StoreLoad barrier for atomic operations if before.
6294 2013-08-02 Jan Hubicka <jh@suse.cz>
6295 Martin Liska <marxin.liska@gmail.com>
6297 * cgraph.c (cgraph_function_body_availability): Do not check
6299 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
6300 symtab_node_availability): Declare.
6301 * ipa.c (can_replace_by_local_alias): New.
6302 (function_and_variable_visibility): Use it.
6303 * symtab.c (symtab_for_node_and_aliases,
6304 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
6306 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6308 PR rtl-optimization/57963
6309 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
6310 (lra_constraints): Use them.
6312 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
6314 * config/arm/types.md (define_attr "type"): Add "load_acq"
6316 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
6318 (cortex_a53_store1): Likewise.
6320 2013-08-01 Jan Hubicka <jh@suse.cz>
6322 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
6323 partitions are not needed.
6325 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
6327 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
6328 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
6329 MAYBE_NON_Q_CLASS_P where appropriate.
6331 2013-08-01 Jan Hubicka <jh@suse.cz>
6333 * cgraph.h (release_function_body): Declare.
6334 * tree.c (free_lang_data_in_decl): Free, parameters and return values
6335 of unused delcarations.
6337 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6339 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
6340 RTL form when subtracting a constant.
6342 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6344 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
6345 Generate canonical plus rtx with negated immediate instead of minus
6347 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
6349 2013-08-01 Jan Hubicka <jh@suse.cz>
6351 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
6352 (cgraph_release_function_body): Likewise.
6353 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
6354 * cgraph.h (cgrpah_node): Rename abstract_and_needed
6355 to used_as_abstract_origin.
6356 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
6357 symbols used as abstract origins.
6358 * cgraphunit.c (analyze_functions): Update.
6359 * ipa.c (symtab_remove_unreachable_nodes): Recompute
6360 used_as_abstract_origin.
6361 * tree-inline.c (tree_function_versioning): Update
6362 used_as_abstract_origin; be ready for DECL_RESULT and
6363 DECL_ARGUMENTS to be NULL.
6365 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
6366 for abstract functions.
6367 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
6370 2013-08-01 Jan Hubicka <jh@suse.cz>
6372 * profile.c (compute_value_histograms): Fix thinko.
6374 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6376 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
6377 aarch-common-protos.h to extra_headers.
6378 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6379 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
6380 * config/aarch64/t-aarch64 (aarch-common.o): Define.
6382 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6384 * config/aarch64/aarch64.md (define_attr "type"): Delete.
6385 Include "../arm/types.md". Define "type" attribute for all patterns.
6386 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
6389 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
6391 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
6392 to support power8 load fusion.
6393 (fusion_gpr_mem_load): Likewise.
6395 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
6397 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
6398 declarations for power8 load fusion.
6399 (emit_fusion_gpr_load): Likewise.
6401 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
6402 tuning for power8, turn on fusion mode by default. Turn on sign
6403 extending fusion mode if normal fusion mode is on, and we are at
6405 (fusion_gpr_load_p): New function, return true if we can fuse an
6406 addis instruction with a dependent load to a GPR.
6407 (emit_fusion_gpr_load): Emit the instructions for power8 load
6410 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
6411 (VSX load fusion peepholes): New peepholes to fuse together an
6412 addi instruction with a VSX load instruction.
6414 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
6415 peepholes to fuse an addis instruction with a load to a GPR base
6416 register. If we are supporting sign extending fusions, convert
6417 sign extending loads to zero extending loads and add an explicit
6420 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6422 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
6423 aarch-common-protos.h to extra_headers.
6424 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6425 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
6426 (arm_early_store_addr_dep): Likewise.
6427 (arm_no_early_alu_shift_dep): Likewise.
6428 (arm_no_early_alu_shift_value_dep): Likewise.
6429 (arm_no_early_mul_dep): Likewise.
6430 (arm_no_early_store_addr_dep): Likewise.
6431 (arm_mac_accumulator_is_mul_result): Likewise.
6432 (arm_mac_accumulator_is_result): Likewise.
6433 * config/arm/aarch-common.c: ... here. New file.
6434 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
6436 (arm_early_store_addr_dep): Likewise.
6437 (arm_no_early_alu_shift_dep): Likewise.
6438 (arm_no_early_alu_shift_value_dep): Likewise.
6439 (arm_no_early_mul_dep): Likewise.
6440 (arm_no_early_store_addr_dep): Likewise.
6441 (arm_mac_accumulator_is_mul_result): Likewise.
6442 (arm_mac_accumulator_is_result): Likewise.
6443 * config/arm/aarch-common-protos.h: ... here. New file.
6444 * config/arm/t-arm (aarch-common.o): Define.
6446 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6448 * config/arm/arm.md: Include new file "types.md".
6449 (define_attr "type"): Move from here to ...
6450 (define_attr "mul32"): Likewise.
6451 (define_attr "mul64"): Likewise.
6452 * config/arm/types.md: ... here. New file.
6454 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
6456 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6457 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6459 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6461 * gen-pass-instances.awk: Fix offset of substr().
6463 2013-07-31 David Malcolm <dmalcolm@redhat.com>
6465 * Makefile.in (pass-instances.def): New.
6466 (passes.o): Replace dependency on passes.def with one on
6469 * gen-pass-instances.awk: New.
6471 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6472 rather than passes.def, updating local definition of NEXT_PASS
6473 macro to add an extra NUM parameter (currently unused).
6475 2013-07-30 David Malcolm <dmalcolm@redhat.com>
6477 * Makefile.in (PASS_MANAGER_H): New.
6478 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6479 (passes.o): Likewise.
6480 (statistics.o): Likewise.
6481 (cgraphunit.o): Likewise.
6482 (context.o): Depend on PASS_MANAGER_H.
6484 * pass_manager.h: New.
6486 * cgraphunit.c (cgraph_add_new_function): Update for moves
6487 of globals to fields of pass_manager.
6488 (analyze_function): Likewise.
6489 (expand_function): Likewise.
6490 (ipa_passes): Likewise.
6491 (compile): Likewise.
6493 * context.c (context::context): New.
6494 * context.h (context::context): New.
6495 (context::get_passes): New.
6496 (context::passes_): New.
6498 * lto-cgraph.c (input_node): Update for moves of globals to
6499 fields of pass_manager.
6501 * passes.c (all_passes): Remove, in favor of a field of the
6502 same name within the new class pass_manager.
6503 (all_small_ipa_passes): Likewise.
6504 (all_lowering_passes): Likewise.
6505 (all_regular_ipa_passes): Likewise.
6506 (all_late_ipa_passes): Likewise.
6507 (all_lto_gen_passes): Likewise.
6508 (passes_by_id): Likewise.
6509 (passes_by_id_size): Likewise.
6510 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6511 the new class pass_manager.
6512 (set_pass_for_id): Convert to...
6513 (pass_manager::set_pass_for_id): ...method.
6514 (get_pass_for_id): Convert to...
6515 (pass_manager::get_pass_for_id): ...method.
6516 (register_one_dump_file): Move body of implementation into...
6517 (pass_manager::register_one_dump_file): ...here.
6518 (register_dump_files_1): Convert to...
6519 (pass_manager::register_dump_files_1): ...method.
6520 (register_dump_files): Convert to...
6521 (pass_manager::register_dump_files): ...method.
6522 (create_pass_tab): Update for moves of globals to fields of
6524 (dump_passes): Move body of implementation into...
6525 (pass_manager::dump_passes): ...here.
6526 (register_pass): Move body of implementation into...
6527 (pass_manager::register_pass): ...here.
6528 (init_optimization_passes): Convert into...
6529 (pass_manager::pass_manager): ...constructor for new
6530 pass_manager class, and initialize the pass_lists array.
6531 (check_profile_consistency): Update for moves of globals to
6532 fields of pass_manager.
6533 (dump_profile_report): Move body of implementation into...
6534 (pass_manager::dump_profile_report): ...here.
6535 (ipa_write_summaries_1): Update for moves of pass lists from
6536 being globals to fields of pass_manager.
6537 (ipa_write_optimization_summaries): Likewise.
6538 (ipa_read_summaries): Likewise.
6539 (ipa_read_optimization_summaries): Likewise.
6540 (execute_all_ipa_stmt_fixups): Likewise.
6542 * statistics.c (statistics_fini): Update for moves of globals to
6543 fields of pass_manager.
6545 * toplev.c (general_init): Replace call to
6546 init_optimization_passes with construction of the pass_manager
6549 * tree-pass.h (all_passes): Remove, in favor of a field of the
6550 same name within the new class pass_manager.
6551 (all_small_ipa_passes): Likewise.
6552 (all_lowering_passes): Likewise.
6553 (all_regular_ipa_passes): Likewise.
6554 (all_lto_gen_passes): Likewise.
6555 (all_late_ipa_passes): Likewise.
6556 (passes_by_id): Likewise.
6557 (passes_by_id_size): Likewise.
6558 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6559 the new class pass_manager.
6560 (get_pass_for_id): Remove.
6562 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6564 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6567 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
6569 * arm.md (mulhi3): New expand pattern.
6571 2013-07-30 Jan Hubicka <jh@suse.cz>
6572 Martin Liska <marxin.liska@gmail.com>
6574 * profile.c (compute_value_histograms): Do not ICE when
6575 there is mismatch only on some counters.
6577 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6579 PR rtl-optimization/57637
6580 * function.c (move_insn_for_shrink_wrap): Also check the
6581 GEN set of the LIVE problem for the liveness analysis
6582 if it exists, otherwise give up.
6584 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6586 PR tree-optimization/57993
6587 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6588 replaced statement in the candidate table.
6589 (phi_add_costs): Return infinite cost when the hidden basis does
6590 not dominate all phis on which the candidate is dependent.
6591 (replace_one_candidate): Record replaced statement in the
6594 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6596 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6597 (ashlv2si3): New expander.
6598 (*ashlv2si3_i): New define_insn_and_split.
6599 * predicates.md (float_operation): Allow patterns with three
6602 PR rtl-optimization/58021
6603 * mode-switching.c (create_pre_exit): Always split off preceding
6604 insns if we are not at the basic block head.
6606 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6608 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6609 (UCLIBC_DYNAMIC_LINKER): New macro.
6610 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6612 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6613 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6615 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6616 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6617 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6618 for SF and DF modes. Use ieee_quad_format for TF mode.
6619 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6620 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6621 (mips_option_override): Handle `-mnan=legacy'.
6622 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6623 `-mabs=2008' and `-mnan=2008'.
6624 (OPTION_DEFAULT_SPECS): Add "nan" default.
6625 (ASM_SPEC): Handle `-mnan='.
6626 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6627 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6628 comment accordingly.
6629 (neg<mode>2): Likewise.
6630 * config/mips/mips.opt (mabs, mnan): New options.
6631 * doc/install.texi (Configuration): Document `--with-nan=' option.
6632 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6634 (MIPS Options): Document them.
6635 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6636 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6637 * configure: Regenerate.
6638 * config.in: Regenerate.
6640 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6642 * config/i386/i386.md (float post-reload splitters): Do not check
6643 for subregs of SSE registers.
6645 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
6646 H.J. Lu <hongjiu.lu@intel.com>
6650 * config/i386/i386.md (post-reload splitter
6651 to avoid partial SSE reg dependency stalls): New pattern.
6653 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6655 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6657 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6659 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6660 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6662 2013-07-26 Cary Coutant <ccoutant@google.com>
6664 * dwarf2out.c (die_checksum_ordered): Don't include template
6665 instantiations in signature.
6666 (is_template_parameter): New function.
6667 (is_template_instantiation): New function.
6668 (generate_skeleton_bottom_up): Don't include template instantiations
6670 (generate_skeleton): Likewise.
6671 (break_out_comdat_types): Move recursive call to break out nested
6673 (prune_unused_types_mark_generic_parms_dies): Call
6674 is_template_parameter.
6676 2013-07-26 Ian Bolton <ian.bolton@arm.com>
6678 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6679 uses vector registers.
6680 * config/aarch64/iterators.md: Add attributes rtn and vas.
6682 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6683 Richard Earnshaw <richard.earnshaw@arm.com>
6685 * combine.c (simplify_comparison): Re-canonicalize operands
6687 * config/arm/arm.md (movcond_addsi): New splitter.
6689 2013-07-25 Sterling Augustine <saugustine@google.com>
6691 * dwarf2out.c (size_of_pubnames): Move code to...
6692 (include_pubname_in_output): ...here. New.
6693 (want_pubnames): Rearrange.
6694 (output_pubnames): Call include_pubname_in_output. Move assertion.
6696 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6698 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6700 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6703 * doc/extend.texi: Remove obsolete builtins. Fix
6704 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6706 2013-07-25 Jan Hubicka <jh@suse.cz>
6708 * cgraph.c (release_function_body): Break out from ...
6709 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6711 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6712 old_tree in the map.
6713 (create_specialized_node): Update.
6714 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6716 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6717 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
6718 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6719 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6722 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6724 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6725 addsi3_carryin_alt2_<optab>): Correct output template.
6727 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6729 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6730 Adjust for arm_restrict_it.
6731 Remove trailing whitespace.
6733 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6735 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6736 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6738 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6740 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6742 PR rtl-optimization/57960
6743 * lra-constraints.c (process_alt_operands): Use the right mode
6744 when checking strict_low.
6746 2013-07-25 Jan Hubicka <jh@suse.cz>
6748 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6749 * cgraph.c (cgraph_remove_node): Do not release function body
6750 when in cgraph streaming.
6751 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
6752 in other partitions are not considered reachable; fix handling of
6755 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6757 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6759 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6764 * config/arm/arm.md ("*sibcall_insn): Replace use of
6765 Ss with US. Adjust output for v5 and v4t.
6766 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
6768 * config/arm/constraints.md ("Ss"): Rename to US.
6770 2013-07-25 Terry Guo <terry.guo@arm.com>
6772 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6773 shift_add/shift_sub0/shift_sub1 RTXs.
6775 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
6776 Anton Blanchard <anton@au1.ibm.com>
6778 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6779 (altivec_vpks<VI_char>ss): Likewise.
6780 (altivec_vpks<VI_char>us): Likewise.
6781 (altivec_vpku<VI_char>us): Likewise.
6782 (altivec_vpku<VI_char>um): Likewise.
6784 2013-07-24 David Malcolm <dmalcolm@redhat.com>
6786 Introduce context class.
6788 * Makefile.in (CONTEXT_H): New.
6789 (OBJS): Add context.o.
6790 (toplev.o): Add CONTEXT_H to dependencies.
6793 * toplev.c (general_init): Create the singleton gcc::context instance.
6799 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6801 PR rtl-optimization/57968
6802 * mode-switching.c (create_pre_exit): Allow instructions that
6803 don't set a return register to need a non-exit mode.
6805 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6806 Anton Blanchard <anton@au1.ibm.com>
6808 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6809 operands to vperm for little endian.
6810 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6811 of lvsl to create the control mask for a vperm for little endian.
6813 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6814 Anton Blanchard <anton@au1.ibm.com>
6816 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6817 two operands for little-endian.
6819 2013-07-23 Steve Ellcey <sellcey@mips.com>
6821 * config/mips/mips.c (mips_case_values_threshold): New.
6822 (TARGET_CASE_VALUES_THRESHOLD): Define.
6824 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6825 Anton Blanchard <anton@au1.ibm.com>
6827 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6828 selection of field for vector splat in little endian mode.
6830 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6832 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6833 expanders to rs6000.md.
6834 (ior<mode>3): Likewise.
6835 (and<mode>3): Likewise.
6836 (one_cmpl<mode>2): Likewise.
6837 (nor<mode>3): Likewise.
6838 (andc<mode>3): Likewise.
6839 (eqv<mode>3): Likewise.
6840 (nand<mode>3): Likewise.
6841 (orc<mode>3): Likewise.
6843 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6846 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6847 to split multi-word logical operations.
6848 (rs6000_split_logical_di): Likewise.
6849 (rs6000_split_logical): Likewise.
6851 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6852 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6853 and allow TImode operations in 32-bit.
6854 (vsx_and<mode>3_64bit): Likewise.
6855 (vsx_ior<mode>3_32bit): Likewise.
6856 (vsx_ior<mode>3_64bit): Likewise.
6857 (vsx_xor<mode>3_32bit): Likewise.
6858 (vsx_xor<mode>3_64bit): Likewise.
6859 (vsx_one_cmpl<mode>2_32bit): Likewise.
6860 (vsx_one_cmpl<mode>2_64bit): Likewise.
6861 (vsx_nor<mode>3_32bit): Likewise.
6862 (vsx_nor<mode>3_64bit): Likewise.
6863 (vsx_andc<mode>3_32bit): Likewise.
6864 (vsx_andc<mode>3_64bit): Likewise.
6865 (vsx_eqv<mode>3_32bit): Likewise.
6866 (vsx_eqv<mode>3_64bit): Likewise.
6867 (vsx_nand<mode>3_32bit): Likewise.
6868 (vsx_nand<mode>3_64bit): Likewise.
6869 (vsx_orc<mode>3_32bit): Likewise.
6870 (vsx_orc<mode>3_64bit): Likewise.
6872 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6873 logical types in GPRs.
6875 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6876 logical insns to rs6000.md, and allow TImode operations in
6878 (altivec_ior<mode>3): Likewise.
6879 (altivec_xor<mode>3): Likewise.
6880 (altivec_one_cmpl<mode>2): Likewise.
6881 (altivec_nor<mode>3): Likewise.
6882 (altivec_andc<mode>3): Likewise.
6884 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6885 attributes for moving the 128-bit logical operations into
6887 (BOOL_REGS_OUTPUT): Likewise.
6888 (BOOL_REGS_OP1): Likewise.
6889 (BOOL_REGS_OP2): Likewise.
6890 (BOOL_REGS_UNARY): Likewise.
6891 (BOOL_REGS_AND_CR0): Likewise.
6892 (one_cmpl<mode>2): Add support for DI logical operations on
6893 32-bit, splitting the operations to 32-bit.
6897 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6898 changes to combine the 32/64-bit code, allow logical operations on
6899 TI mode in 32-bit, and to use similar match_operator patterns like
6900 scalar mode uses. Combine the Altivec and VSX code for logical
6901 operations, and move it here.
6902 (ior<mode>3, 128-bit types): Likewise.
6903 (xor<mode>3, 128-bit types): Likewise.
6904 (one_cmpl<mode>3, 128-bit types): Likewise.
6905 (nor<mode>3, 128-bit types): Likewise.
6906 (andc<mode>3, 128-bit types): Likewise.
6907 (eqv<mode>3, 128-bit types): Likewise.
6908 (nand<mode>3, 128-bit types): Likewise.
6909 (orc<mode>3, 128-bit types): Likewise.
6910 (and<mode>3_internal): Likewise.
6911 (bool<mode>3_internal): Likewise.
6912 (boolc<mode>3_internal1): Likewise.
6913 (boolc<mode>3_internal2): Likewise.
6914 (boolcc<mode>3_internal1): Likewise.
6915 (boolcc<mode>3_internal2): Likewise.
6916 (eqv<mode>3_internal1): Likewise.
6917 (eqv<mode>3_internal2): Likewise.
6918 (one_cmpl1<mode>3_internal): Likewise.
6920 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6922 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6923 Rename flag_stack_usage to flag_stack_usage_info.
6925 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
6927 * config/microblaze/sync.md: New file.
6928 * config/microblaze/microblaze.md: Include sync.md
6929 * config/microblaze/microblaze.c: Add print_operand 'y'.
6930 * config/microblaze/constraints.md: Add memory_contraint
6931 'Q' which is a single register.
6933 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6935 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6937 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6939 * reload.c (find_reloads): Exit loop once we find this operand
6940 cannot be reloaded somehow for this alternative.
6942 * reload.c (find_reloads): Exit loop once we find a hard register.
6944 * rtlanal.c (computed_jump_p): Exit loop once we find label
6947 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6949 * cfgloopmanip.c (remove_path): Exit loop after setting
6952 * gensupport.c (subst_dup): Avoid loop if code is not
6953 MATCH_DUP nor MATCH_OP_DUP.
6955 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6957 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6959 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6961 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6962 true for SP_REGNUM if mode == ptr_mode.
6963 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6964 with value R0_REGNUM + 31.
6966 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6968 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6969 pad pointer-typed argument downward.
6971 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6973 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6974 and __ILP32__ when the ILP32 model is in use.
6976 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6978 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6979 (aarch64_load_symref_appropriately): In the case of
6980 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6981 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6982 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6983 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6984 if 'mode' doesn't equal to 'ptr_mode'.
6985 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6986 'vcall_offset'; change to call aarch64_emit_move differently depending
6987 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6988 to calculate the upper bound of 'vcall_offset'.
6989 (aarch64_cannot_force_const_mem): Change to also return true if
6991 (aarch64_legitimize_reload_address): In the case of large
6992 displacements, add new local variable 'xmode' and an assertion
6993 based on it; change to use 'xmode' to generate the new rtx and
6995 (aarch64_asm_trampoline_template): Change to generate the template
6996 differently depending on TARGET_ILP32 or not; change to use
6997 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6998 (aarch64_trampoline_size): Removed.
6999 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
7000 and replace immediate literals with it. Change to use 'ptr_mode'
7001 instead of 'DImode' and call convert_memory_address if the mode
7002 of 'fnaddr' doesn't equal to 'ptr_mode'.
7003 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
7005 (aarch64_elf_asm_destructor): Likewise.
7006 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
7007 on TARGET_ILP32 instead of aarch64_trampoline_size.
7008 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
7009 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
7010 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
7011 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
7012 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
7013 (storewb_pair<GPI:mode>_<P:mode>): ... this.
7014 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
7015 depending on the value of 'mode'.
7016 (add_losym_<mode>): New.
7017 (ldr_got_small_<mode>): New, based on ldr_got_small.
7018 (ldr_got_small): Remove.
7019 (ldr_got_small_sidi): New.
7020 * config/aarch64/iterators.md (P): New.
7021 (PTR): Change to 'ptr_mode' in the condition.
7023 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
7025 * config.gcc (aarch64*-*-*): Support --with-abi.
7026 (aarch64*-*-elf): Support --with-multilib-list.
7027 (aarch64*-*-linux*): Likewise.
7028 (supported_defaults): Add abi to aarch64*-*-*.
7029 * configure.ac: Mention AArch64 for --with-multilib-list.
7030 * configure: Re-generated.
7031 * config/aarch64/biarchilp32.h: New file.
7032 * config/aarch64/biarchlp64.h: New file.
7033 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
7035 (MULTILIB_DEFAULTS): Ditto.
7036 (DRIVER_SELF_SPECS): Ditto.
7037 (ASM_SPEC): Update to also substitute -mabi.
7038 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
7039 file whose name depends on -mabi= and -mbig-endian.
7040 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
7042 (POINTER_SIZE): New define.
7043 (POINTERS_EXTEND_UNSIGNED): Ditto.
7044 (enum aarch64_abi_type): New enumeration tag.
7045 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
7046 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
7047 (TARGET_ILP32): New define.
7048 * config/aarch64/aarch64.opt (mabi): New.
7050 (ilp32, lp64): New values for -mabi.
7051 * config/aarch64/t-aarch64 (comma): New define.
7052 (MULTILIB_OPTIONS): Ditto.
7053 (MULTILIB_DIRNAMES): Ditto.
7054 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
7055 * doc/invoke.texi: Document -mabi for AArch64.
7057 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
7059 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
7061 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7062 Anton Blanchard <anton@au1.ibm.com>
7064 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
7065 endianness when selecting field to splat.
7067 2013-07-22 Eric Christopher <echristo@gmail.com>
7069 * dwarf2out.c (die_odr_checksum): New function to use
7070 CHECKSUM_ macros and ULEB128 for DIE tag.
7071 (generate_type_signature): Use.
7073 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
7075 * config.gcc (sparc*-*-*): Accept leon3 processor.
7076 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
7077 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
7078 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
7079 * config/sparc/sparc.opt (enum processor_type): Add leon3.
7080 (mfix-ut699): Adjust comment.
7081 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
7082 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
7083 (CPP_CPU_SPEC): Likewise.
7084 (ASM_CPU_SPEC): Likewise.
7085 * config/sparc/sparc.c (leon3_cost): New constant.
7086 (sparc_option_override): Add leon3 support.
7087 (mem_ref): New function.
7088 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
7089 (sparc_do_work_around_errata): Look into the instruction in the delay
7090 slot and adjust accordingly. Add fix for the data cache nullify issues
7091 of the UT699. Change insertion position for the NOP.
7092 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
7093 (leon3_load): New reservation.
7094 (leon_store): Bump latency to 2.
7095 (grfpu): New automaton.
7096 (grfpu_alu): New unit.
7097 (grfpu_ds): Likewise.
7098 (leon_fp_alu): Adjust.
7099 (leon_fp_mult): Delete.
7100 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
7101 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
7102 * config/sparc/sparc.md (cpu): Add leon3.
7103 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
7105 (atomic_test_and_set): Likewise.
7108 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
7110 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
7111 default for R5900 targets.
7112 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
7113 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
7114 * config/mips/mips.c (mips_option_override): Report an error for
7115 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
7116 for -march=r5900 -mhard-float.
7118 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
7120 * df-problems.c (can_move_insns_across): Exit loop once we
7121 find a non-fixed, non-global register.
7123 * ipa-pure-const.c (propagate_nothrow): Exit loop after
7126 * omega.c (omega_eliminate_red): Break after setting red_found.
7127 (omega_problem_has_red_equations): Similarly after setting found.
7128 (omega_query_variable): Similarly after setting coupled.
7130 2013-07-22 Marek Polacek <polacek@redhat.com>
7132 * gimplify.c: Don't include gimple.h twice.
7134 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7136 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
7137 instead of TARGET_THUMB1.
7138 (Pz): New constraint.
7139 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
7141 (compare_negsi_si): Likewise.
7142 (compare_addsi2_op0): Likewise.
7143 (compare_addsi2_op1): Likewise.
7144 (addsi3_carryin_<optab>): Likewise.
7145 (addsi3_carryin_alt2_<optab>): Likewise.
7146 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
7147 for arm_restrict_it.
7148 (subsi3_carryin): Likewise.
7149 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
7150 (minmax_arithsi): Disable for arm_restrict_it.
7151 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
7152 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
7153 (satsi_<SAT:code>_shift): Likewise.
7154 (arm_shiftsi3): Add alternative for 16-bit encoding.
7155 (arm32_movhf): Disable for arm_restrict_it.
7156 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
7157 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
7159 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
7161 * config/arm/arm.md (attribute "insn"): Delete.
7162 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
7163 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
7164 (not_shiftsi): Update for attribute change.
7165 (not_shiftsi_compare0): Likewise.
7166 (not_shiftsi_compare0_scratch): Likewise.
7167 (arm_one_cmplsi2): Likewise.
7168 (thumb1_one_cmplsi2): Likewise.
7169 (notsi_compare0): Likewise.
7170 (notsi_compare0_scratch): Likewise.
7171 (thumb1_movdi_insn): Likewise.
7172 (arm_movsi_insn): Likewise.
7173 (movhi_insn_arch4): Likewise.
7174 (movhi_bytes): Likewise.
7175 (arm_movqi_insn): Likewise.
7176 (thumb1_movqi_insn): Likewise.
7177 (arm32_movhf): Likewise.
7178 (thumb1_movhf): Likewise.
7179 (arm_movsf_soft_insn): Likewise.
7180 (thumb1_movsf_insn): Likewise.
7181 (thumb_movdf_insn): Likewise.
7182 (movsicc_insn): Likewise.
7183 (movsfcc_soft_insn): Likewise.
7184 (and_scc): Likewise.
7185 (cond_move): Likewise.
7186 (if_move_not): Likewise.
7187 (if_not_move): Likewise.
7188 (if_shift_move): Likewise.
7189 (if_move_shift): Likewise.
7190 (if_shift_shift): Likewise.
7191 (if_not_arith): Likewise.
7192 (if_arith_not): Likewise.
7193 (cond_move_not): Likewise.
7194 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
7195 (neon_mov<mode>): Likewise.
7196 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
7197 (thumb2_movsi_vfp): Likewise.
7198 (movsf_vfp): Likewise.
7199 (thumb2_movsf_vfp): Likewise.
7200 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7202 (cortexa7_older_only): Likewise.
7203 (cortexa7_younger): Likewise.
7204 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
7205 (1020alu_shift_op): Likewise.
7206 (1020alu_shift_reg_op): Likewise.
7207 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
7208 (alu_shift_op): Likewise.
7209 (alu_shift_reg_op): Likewise.
7210 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
7211 (11_alu_shift_op): Likewise.
7212 (11_alu_shift_reg_op): Likewise.
7213 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
7214 (9_alu_shift_reg_op): Likewise.
7215 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7217 (cortex_a15_alu_shift): Likewise.
7218 (cortex_a15_alu_shift_reg): Likewise.
7219 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
7220 (cortex_a5_alu_shift): Likewise.
7221 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
7223 (cortex_a53_alu_shift): Likewise.
7224 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7226 (cortex_a7_alu_reg): Likewise.
7227 (cortex_a7_alu_shift): Likewise.
7228 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7229 (cortex_a8_alu_shift): Likewise.
7230 (cortex_a8_alu_shift_reg): Likewise.
7231 (cortex_a8_mov): Likewise.
7232 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7233 (cortex_a9_dp_shift): Likewise.
7234 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
7235 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
7236 (cortex_r4_mov): Likewise.
7237 (cortex_r4_alu_shift): Likewise.
7238 (cortex_r4_alu_shift_reg): Likewise.
7239 * config/arm/fa526.md (526_alu_op): Update for attribute change.
7240 (526_alu_shift_op): Likewise.
7241 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
7242 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
7243 (626te_alu_shift_op): Likewise.
7244 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
7245 (726te_alu_op): Likewise.
7246 (726te_alu_shift_op): Likewise.
7247 (726te_alu_shift_reg_op): Likewise.
7248 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
7249 (mp626_alu_shift_op): Likewise.
7250 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
7251 (pj4_alu_e1_conds): Likewise.
7252 (pj4_alu): Likewise.
7253 (pj4_alu_conds): Likewise.
7254 (pj4_shift): Likewise.
7255 (pj4_shift_conds): Likewise.
7256 (pj4_alu_shift): Likewise.
7257 (pj4_alu_shift_conds): Likewise.
7259 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7261 * config/arm/predicates.md (shiftable_operator_strict_it):
7263 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
7264 Disable cond_exec version for arm_restrict_it.
7265 (thumb2_smaxsi3): Convert to generate cond_exec.
7266 (thumb2_sminsi3): Likewise.
7267 (thumb32_umaxsi3): Likewise.
7268 (thumb2_uminsi3): Likewise.
7269 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
7270 (thumb2_neg_abssi2): Likewise.
7271 (thumb2_mov_scc): Add alternative for 16-bit encoding.
7272 (thumb2_movsicc_insn): Adjust alternatives.
7273 (thumb2_mov_negscc): Disable for arm_restrict_it.
7274 (thumb2_mov_negscc_strict_it): New pattern.
7275 (thumb2_mov_notscc_strict_it): New pattern.
7276 (thumb2_mov_notscc): Disable for arm_restrict_it.
7277 (thumb2_ior_scc): Likewise.
7278 (thumb2_ior_scc_strict_it): New pattern.
7279 (thumb2_cond_move): Adjust for arm_restrict_it.
7280 (thumb2_cond_arith): Disable for arm_restrict_it.
7281 (thumb2_cond_arith_strict_it): New pattern.
7282 (thumb2_cond_sub): Adjust for arm_restrict_it.
7283 (thumb2_movcond): Likewise.
7284 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
7285 (thumb2_zero_extendhisi2_v6): Likewise.
7286 (thumb2_zero_extendqisi2_v6): Likewise.
7287 (orsi_notsi_si): Likewise.
7288 (orsi_not_shiftsi_si): Likewise.
7290 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
7292 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
7293 instruction sequence is 1 byte shorter.
7295 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
7297 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
7298 it is not needed after split.
7300 2013-07-20 Iain Sandoe <iain@codesourcery.com>
7303 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
7304 second label for nonlocal goto receivers. Don't output pic base labels
7305 unless we're producing PIC; mark that action unreachable().
7306 (ix86_save_reg): If the function contains a nonlocal label, save the
7308 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
7309 * config/darwin.c (emitted_pic_label_num): New GTY.
7310 (update_pic_label_number_if_needed): New.
7311 (machopic_output_function_base_name): Adjust for nonlocal receiver
7313 (machopic_should_output_picbase_label): New.
7314 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
7315 (nonlocal_goto_receiver): New insn and split.
7317 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
7319 * config/aarch64/aarch64-builtins.c
7320 (aarch64_fold_builtin): Fold abs in all modes.
7321 * config/aarch64/aarch64-simd-builtins.def
7322 (abs): Enable for all modes.
7323 * config/aarch64/arm_neon.h
7324 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
7325 (vabs_f64): Add missing intrinsic.
7327 2013-07-19 Ian Bolton <ian.bolton@arm.com>
7329 * config/aarch64/arm_neon.h (vabs_s64): New function
7331 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
7334 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
7335 * config/avr/avr.md (adjust_len): Add `round'.
7336 * config/avr/avr-protos.h (avr_out_round): New prototype.
7337 (avr_out_plus): Add `out_label' argument.
7338 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
7339 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
7340 Handle the case where `insn' is just a pattern.
7341 (avr_out_bitop): Handle the case where `insn' is just a pattern.
7342 (avr_out_round): New function.
7343 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
7345 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
7347 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7348 Add check for flag_stack_usage to handle -fstack-usage support
7350 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
7352 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
7353 interaction for new Power8 flags and VSX.
7355 2013-07-18 Sriraman Tallam <tmsriram@google.com>
7358 * tree-inline.c (expand_call_inline): Emit errors during
7359 early_inlining only if optimization is not turned on.
7361 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7365 * passes.c (init_optimization_passes): Move the construction of
7366 the pass hierarchy into a new passes.def file.
7368 * Makefile.in (passes.o): Add dependency on passes.def.
7370 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7372 * passes.c (init_optimization_passes): Introduce macros for
7373 constructing the tree of passes (INSERT_PASSES_AFTER,
7374 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
7375 TERMINATE_PASS_LIST).
7377 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
7378 Wei Mi <wmi@google.com>
7380 PR rtl-optimization/57878
7381 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
7383 (reload_pseudo_compare_func): Check nregs first for reload
7386 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7388 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
7390 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
7392 * read-rtl.c (validate_const_int): Once an invalid character is
7393 seen, quit the loop.
7395 * gengtype.c (write_roots): Similarly once we find the "deletable"
7396 or "if_marked" option.
7398 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7400 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
7401 "xtab" and "sat". Move value "clz" from here to ...
7402 (attriubte "type"): ... here.
7403 (satsi_<SAT:code>): Delete "insn" attribute.
7404 (satsi_<SAT:code>_shift): Likewise.
7405 (arm_zero_extendqisi2addsi): Likewise.
7406 (arm_extendqisi2addsi): Likewise.
7407 (clzsi2): Update for attribute changes.
7408 (rbitsi2): Likewise.
7409 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
7411 (arm_usatsihi): Likewise.
7412 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7414 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7416 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
7417 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
7418 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
7419 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
7420 in alphabetical order.
7421 (attribute "core_cycles"): Update for attribute changes.
7422 (arm_addsi3): Likewise.
7423 (addsi3_compare0): Likewise.
7424 (addsi3_compare0_scratch): Likewise.
7425 (addsi3_compare_op1): Likewise.
7426 (addsi3_compare_op2): Likewise.
7427 (compare_addsi2_op0): Likewise.
7428 (compare_addsi2_op1): Likewise.
7429 (addsi3_carryin_shift_<optab>): Likewise.
7430 (subsi3_carryin_shift): Likewise.
7431 (rsbsi3_carryin_shift): Likewise.
7432 (arm_subsi3_insn): Likewise.
7433 (subsi3_compare0): Likewise.
7434 (subsi3_compare): Likewise.
7435 (arm_andsi3_insn): Likewise.
7436 (thumb1_andsi3_insn): Likewise.
7437 (andsi3_compare0): Likewise.
7438 (andsi3_compare0_scratch): Likewise.
7439 (zeroextractsi_compare0_scratch
7440 (andsi_not_shiftsi_si): Likewise.
7441 (iorsi3_insn): Likewise.
7442 (iorsi3_compare0): Likewise.
7443 (iorsi3_compare0_scratch): Likewise.
7444 (arm_xorsi3): Likewise.
7445 (thumb1_xorsi3_insn): Likewise.
7446 (xorsi3_compare0): Likewise.
7447 (xorsi3_compare0_scratch): Likewise.
7448 (satsi_<SAT:code>_shift): Likewise.
7450 (arm_shiftsi3): Likewise.
7451 (shiftsi3_compare0): Likewise.
7452 (not_shiftsi): Likewise.
7453 (not_shiftsi_compare0): Likewise.
7454 (not_shiftsi_compare0_scratch): Likewise.
7455 (arm_one_cmplsi2): Likewise.
7456 (thumb_one_complsi2): Likewise.
7457 (notsi_compare0): Likewise.
7458 (notsi_compare0_scratch): Likewise.
7459 (thumb1_zero_extendhisi2): Likewise.
7460 (arm_zero_extendhisi2): Likewise.
7461 (arm_zero_extendhisi2_v6): Likewise.
7462 (arm_zero_extendhisi2addsi): Likewise.
7463 (thumb1_zero_extendqisi2): Likewise.
7464 (thumb1_zero_extendqisi2_v6): Likewise.
7465 (arm_zero_extendqisi2): Likewise.
7466 (arm_zero_extendqisi2_v6): Likewise.
7467 (arm_zero_extendqisi2addsi): Likewise.
7468 (thumb1_extendhisi2): Likewise.
7469 (arm_extendhisi2): Likewise.
7470 (arm_extendhisi2_v6): Likewise.
7471 (arm_extendqisi): Likewise.
7472 (arm_extendqisi_v6): Likewise.
7473 (arm_extendqisi2addsi): Likewise.
7474 (thumb1_extendqisi2): Likewise.
7475 (thumb1_movdi_insn): Likewise.
7476 (arm_movsi_insn): Likewise.
7477 (movsi_compare0): Likewise.
7478 (movhi_insn_arch4): Likewise.
7479 (movhi_bytes): Likewise.
7480 (arm_movqi_insn): Likewise.
7481 (thumb1_movqi_insn): Likewise.
7482 (arm32_movhf): Likewise.
7483 (thumb1_movhf): Likewise.
7484 (arm_movsf_soft_insn): Likewise.
7485 (thumb1_movsf_insn): Likewise.
7486 (movdf_soft_insn): Likewise.
7487 (thumb_movdf_insn): Likewise.
7488 (arm_cmpsi_insn): Likewise.
7489 (cmpsi_shiftsi): Likewise.
7490 (cmpsi_shiftsi_swp): Likewise.
7491 (arm_cmpsi_negshiftsi_si): Likewise.
7492 (movsicc_insn): Likewise.
7493 (movsfcc_soft_insn): Likewise.
7494 (arith_shiftsi): Likewise.
7495 (arith_shiftsi_compare0
7496 (arith_shiftsi_compare0_scratch
7497 (sub_shiftsi): Likewise.
7498 (sub_shiftsi_compare0
7499 (sub_shiftsi_compare0_scratch
7500 (and_scc): Likewise.
7501 (cond_move): Likewise.
7502 (if_plus_move): Likewise.
7503 (if_move_plus): Likewise.
7504 (if_move_not): Likewise.
7505 (if_not_move): Likewise.
7506 (if_shift_move): Likewise.
7507 (if_move_shift): Likewise.
7508 (if_shift_shift): Likewise.
7509 (if_not_arith): Likewise.
7510 (if_arith_not): Likewise.
7511 (cond_move_not): Likewise.
7512 (thumb1_ashlsi3): Set type attribute.
7513 (thumb1_ashrsi3): Likewise.
7514 (thumb1_lshrsi3): Likewise.
7515 (thumb1_rotrsi3): Likewise.
7516 (shiftsi3_compare0_scratch): Likewise.
7517 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7518 (neon_mov<mode>): Likewise.
7519 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7521 (thumb2_movsi_insn): Likewise.
7522 (thumb2_cmpsi_neg_shiftsi): Likewise.
7523 (thumb2_extendqisi_v6): Likewise.
7524 (thumb2_zero_extendhisi2_v6): Likewise.
7525 (thumb2_zero_extendqisi2_v6): Likewise.
7526 (thumb2_shiftsi3_short): Likewise.
7527 (thumb2_addsi3_compare0_scratch): Likewise.
7528 (orsi_not_shiftsi_si): Likewise.
7529 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7530 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7532 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7533 (1020alu_shift_op): Likewise.
7534 (1020alu_shift_reg_op): Likewise.
7535 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7536 (alu_shift_op): Likewise.
7537 (alu_shift_reg_op): Likewise.
7538 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7539 (11_alu_shift_op): Likewise.
7540 (11_alu_shift_reg_op): Likewise.
7541 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7542 (9_alu_shift_reg_op): Likewise.
7543 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7545 (cortex_a15_alu_shift): Likewise.
7546 (cortex_a15_alu_shift_reg): Likewise.
7547 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7549 (cortex_a5_alu_shift): Likewise.
7550 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7552 (cortex_a53_alu_shift): Likewise.
7553 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7555 (cortex_a7_alu_reg): Likewise.
7556 (cortex_a7_alu_shift): Likewise.
7557 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7559 (cortex_a8_alu_shift): Likewise.
7560 (cortex_a8_alu_shift_reg): Likewise.
7561 (cortex_a8_mov): Likewise.
7562 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7563 (cortex_a9_dp_shift): Likewise.
7564 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7566 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7568 (cortex_r4_mov): Likewise.
7569 (cortex_r4_alu_shift): Likewise.
7570 (cortex_r4_alu_shift_reg): Likewise.
7571 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7572 (526_alu_shift_op): Likewise.
7573 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7574 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7575 (626te_alu_shift_op): Likewise.
7576 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7577 (726te_alu_op): Likewise.
7578 (726te_alu_shift_op): Likewise.
7579 (726te_alu_shift_reg_op): Likewise.
7580 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7581 (mp626_alu_shift_op): Likewise.
7582 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7583 (pj4_alu_e1_conds): Likewise.
7584 (pj4_alu): Likewise.
7585 (pj4_alu_conds): Likewise.
7586 (pj4_shift): Likewise.
7587 (pj4_shift_conds): Likewise.
7588 (pj4_alu_shift): Likewise.
7589 (pj4_alu_shift_conds): Likewise.
7590 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7592 (cortexa7_older_only): Likewise.
7593 (cortexa7_younger): Likewise.
7595 2013-07-18 David Malcolm <dmalcolm@redhat.com>
7597 * ipa-pure-const.c (generate_summary): Rename to...
7598 (pure_const_generate_summary): ... this.
7600 2013-07-17 Iain Sandoe <iain@codesourcery.com>
7602 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7604 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
7607 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7610 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7612 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7613 enabled without -march=zEC12.
7614 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7617 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7619 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7620 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7621 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7622 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7623 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7624 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7625 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7627 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7628 and ISA_HAS_NMADD3_NMSUB3.
7629 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7630 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7631 (nmsub4<mode>, nmsub3<mode>): Likewise.
7632 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7634 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7636 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7637 TARGET_MIPS5400 checking.
7639 2013-07-16 Jakub Jelinek <jakub@redhat.com>
7640 Peter Bergner <bergner@vnet.ibm.com>
7642 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7643 registers in the comment.
7644 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7645 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7646 rather than FIRST_PSEUDO_REGISTERS.
7648 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
7650 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7651 enable extra ISA flags with TARGET_HTM.
7653 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7655 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7658 2013-07-15 Cong Hou <congh@google.com>
7660 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7661 in compare function for sorting.
7663 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
7665 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7666 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7667 * config/rs6000/rs6000.opt: Add -mhtm option.
7668 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7669 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7670 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7671 __HTM__ if the HTM instructions are available.
7672 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7673 htm_spr_reg_operand): New define_predicates.
7674 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7675 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7677 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7678 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7679 HTM builtin functions.
7680 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7681 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7682 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7683 (rs6000_builtin_mask_calculate): Likewise.
7684 (rs6000_option_override_internal): Likewise.
7685 (bdesc_htm): Add new HTM builtin support.
7686 (htm_spr_num): New function.
7687 (htm_spr_regno): Likewise.
7688 (rs6000_htm_spr_icode): Likewise.
7689 (htm_expand_builtin): Likewise.
7690 (htm_init_builtins): Likewise.
7691 (rs6000_expand_builtin): Add support for HTM builtin functions.
7692 (rs6000_init_builtins): Likewise.
7693 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7695 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7696 (TARGET_HTM, MASK_HTM): Define macros.
7697 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7698 (FIXED_REGISTERS): Likewise.
7699 (CALL_USED_REGISTERS): Likewise.
7700 (CALL_REALLY_USED_REGISTERS): Likewise.
7701 (REG_ALLOC_ORDER): Likewise.
7702 (enum reg_class): Likewise.
7703 (REG_CLASS_NAMES): Likewise.
7704 (REG_CLASS_CONTENTS): Likewise.
7705 (REGISTER_NAMES): Likewise.
7706 (ADDITIONAL_REGISTER_NAMES): Likewise.
7707 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7708 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7709 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7710 * config/rs6000/htm.md: New file.
7711 * config/rs6000/htmintrin.h: New file.
7712 * config/rs6000/htmxlintrin.h: New file.
7714 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7716 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7717 Define SYMBOL_TINY_GOT, update comment.
7718 * config/aarch64/aarch64.c
7719 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7720 (aarch64_expand_mov_immediate): Likewise.
7721 (aarch64_print_operand): Likewise.
7722 (aarch64_classify_symbol): Likewise.
7723 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7724 (ldr_got_tiny): Define.
7726 2013-07-13 Tobias Grosser <tobias@grosser.es>
7728 PR tree-optimization/54094
7729 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7730 scheduling dimension for the parallelism check from the polyhedral
7731 information in the AST.
7732 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7735 2013-07-13 Jason Merrill <jason@redhat.com>
7737 * print-tree.c (debug_vec_tree): Use debug_raw.
7738 (debug_raw (vec<tree, va_gc> &)): New.
7739 (debug_raw (vec<tree, va_gc> *)): New.
7740 * tree.h: Declare them.
7742 2013-07-13 Bin Cheng <bin.cheng@arm.com>
7744 * ifcvt.c (ifcvt_after_combine): New static variable.
7745 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7747 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7748 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7749 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7751 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7753 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7755 2013-07-12 Michael Matz <matz@suse.de>
7758 * convert.c (convert_to_real): Reject non-float inner types.
7760 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7762 * config/aarch64/aarch64-protos.h
7763 (aarch64_simd_immediate_valid_for_move): Remove.
7764 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7765 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7767 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7769 2013-07-11 Steve Ellcey <sellcey@mips.com>
7771 * config/mips/mips.c (mips_conditional_register_usage): Do not
7772 use t[0-7] registers in MIPS16 mode when optimizing for size.
7774 2013-07-11 Sriraman Tallam <tmsriram@google.com>
7776 * config/i386/i386.c (dispatch_function_versions): Fix array
7777 indexing of function_version_info to match actual_versions.
7779 2013-07-11 Teresa Johnson <tejohnson@google.com>
7781 * vec.h (struct va_gc): Move release out-of-line.
7782 (va_gc::release): Call ggc_free on released vec.
7784 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7786 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7787 Require GOT register as additional operand in UNSPEC.
7788 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7789 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7790 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7791 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7792 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7793 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7794 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7796 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
7799 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7800 name seen by assembler/linker rather if available.
7802 2013-07-11 Andreas Schwab <schwab@suse.de>
7804 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7806 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7808 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
7810 2013-07-10 Joseph Myers <joseph@codesourcery.com>
7812 * doc/tm.texi.in: Move hook documentation to ....
7813 * target.def: ... here.
7815 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7817 * doc/tm.texi: Regenerate.
7819 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7822 * doc/invoke.texi: Document -Wconditionally-supported.
7824 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7827 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7830 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7833 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7834 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7835 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7836 Remove duplicate devices.
7837 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7838 * config/avr/t-multilib: Regenerate.
7839 * config/avr/avr-tables.opt: Regenerate.
7840 * doc/avr-mmcu.texi: Regenerate.
7842 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
7845 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7847 2013-07-10 Graham Stott <graham.stott@btinternet.com>
7849 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7850 the cost of MULT when optimizing for size.
7852 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7854 * config/cr16/cr16-protos.h: Don't include target.h.
7856 2013-07-09 Joseph Myers <joseph@codesourcery.com>
7858 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7859 adjust register size for TDmode and TFmode for VSX registers.
7861 2013-07-08 Kai Tietz <ktietz@redhat.com>
7864 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7865 hook_bool_const_tree_true.
7867 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7869 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7870 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7871 * config/s390/s390.md: Define FPR*_REGNUM constants.
7872 Fix FPR2_REGNUM constant (18 -> 17).
7873 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7874 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7875 ("trunc<BFP:mode><DFP_ALL:mode>2")
7876 ("trunc<DFP_ALL:mode><BFP:mode>2")
7877 ("*extend<BFP:mode><DFP_ALL:mode>2")
7878 ("*extend<DFP_ALL:mode><BFP:mode>2")
7879 ("extend<BFP:mode><DFP_ALL:mode>2")
7880 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7883 2013-07-08 Graham Stott <graham.stott@btinternet.com>
7885 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
7887 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7889 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7890 and cfun_fpr_bit_p to cfun_fpr_save_p.
7891 (s390_frame_area, s390_register_info, s390_frame_info)
7892 (s390_emit_prologue, s390_emit_epilogue)
7893 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7895 * config/s390/s390.h: Define *_REGNUM macros for floating point
7898 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7900 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7902 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7904 PR rtl-optimization/57786
7905 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7906 and break out of the loop when it is set to false.
7908 2013-07-08 Jakub Jelinek <jakub@redhat.com>
7911 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7912 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7913 (const_int 63)) 0)).
7914 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7915 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7916 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7918 PR rtl-optimization/57829
7919 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7920 mask bits outside of mode are just sign-extension from mode to HWI.
7922 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7924 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7925 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7926 adjust_address instead of change_address to keep info about alignment.
7927 (emit_strmov): Remove.
7928 (emit_memmov): New function.
7929 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7930 (expand_movmem_epilogue): Likewise and return updated rtx for
7932 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7933 destination and source.
7934 (decide_alignment): Refactor, handle vector_loop.
7935 (ix86_expand_movmem): Likewise.
7936 (ix86_expand_setmem): Likewise.
7937 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7939 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
7941 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7942 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7944 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7946 * config/i386/sse.md (sse_movlhps): Change alternative 3
7947 of operand 2 to "m".
7949 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
7952 * config/i386/sse.md (iptr): New mode attribute.
7953 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7954 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7955 (<sse>_vmmul<mode>3): Ditto.
7956 (<sse>_vmdiv<mode>3): Ditto.
7957 (sse_vmrcpv4sf2): Ditto.
7958 (<sse>_vmsqrt<mode>2): Ditto.
7959 (sse_vmrsqrtv4sf2): Ditto.
7960 (<sse>_vm<code><mode>3): Ditto.
7961 (avx_vmcmp<mode>3): Ditto.
7962 (<sse>_vmmaskcmp<mode>3): Ditto.
7963 (<sse>_comi): Ditto.
7964 (<sse>_ucomi): Ditto.
7965 (*xop_vmfrcz_<mode>): Ditto.
7966 (*fmai_fmadd_<mode>): Ditto.
7967 (*fmai_fmsub_<mode>): Ditto.
7968 (*fmai_fnmadd_<mode>): Ditto.
7969 (*fmai_fnmsub_<mode>): Ditto.
7970 (*fma4i_vmfmadd_<mode>): Ditto.
7971 (*fma4i_vmfmsub_<mode>): Ditto.
7972 (*fma4i_vmfnmadd_<mode>): Ditto.
7973 (*fma4i_vmfnmsub_<mode>): Ditto.
7974 (*xop_vmfrcz_<mode>): Ditto.
7975 (sse_cvtps2pi): Ditto.
7976 (sse_cvttps2pi): Ditto.
7977 (sse_cvtss2si): Ditto.
7978 (sse_cvtss2si_2): Ditto.
7979 (sse_cvtss2siq_2): Ditto.
7980 (sse_cvttss2si): Ditto.
7981 (sse_cvttss2siq): Ditto.
7982 (sse_cvtsd2si): Ditto.
7983 (sse_cvtsd2si_2): Ditto.
7984 (sse_cvtsd2siq_2): Ditto.
7985 (sse_cvttsd2si): Ditto.
7986 (sse_cvttsd2siq): Ditto.
7987 (sse_cvtsd2ss): Ditto.
7988 (sse_cvtss2sd): Ditto.
7989 (avx2_pbroadcast<mode>): Ditto.
7990 (avx2_pbroadcast<mode>_1): Ditto.
7991 (*avx_vperm_broadcast_v4sf): Ditto.
7993 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7994 (sse_movlhps): Ditto.
7995 (sse_storehps): Ditto.
7996 (sse_loadhps): Ditto.
7997 (sse_storelps): Ditto.
7998 (sse_loadlps): Ditto.
7999 (*vec_concatv4sf): Ditto.
8000 (*vec_interleave_highv2df): Ditto.
8001 (*vec_interleave_lowv2df): Ditto.
8002 (*vec_extractv2df_1_sse): Ditto.
8003 (*vec_extractv2df_0_sse): Ditto.
8004 (sse2_storelpd): Ditto.
8005 (sse2_loadlpd): Ditto.
8006 (sse2_movsd): Ditto.
8007 (*vec_concatv4si): Ditto.
8008 (vec_concatv2di): Ditto.
8010 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
8011 for Intel asm dialect.
8012 (mmx_punpcklwd): Ditto.
8013 (mmx_punpckldq): Ditto.
8015 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
8016 for intel assembler dialect.
8018 2013-07-06 Jakub Jelinek <jakub@redhat.com>
8021 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
8022 for BUILT_IN_C{LZ,LRSB}*.
8023 * tree.h (CASE_INT_FN): Add FN##IMAX case.
8024 * tree-vrp.c (extract_range_basic): Handle
8025 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
8026 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
8027 fall thru to code calling set_value*.
8028 * builtins.c (expand_builtin): Remove *IMAX cases.
8029 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
8030 if width is bigger than 2*HWI.
8032 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
8034 PR rtl-optimization/55342
8035 * lra-int.h (lra_subreg_reload_pseudos): New.
8036 * lra.c: Add undoing optional reloads to the block diagram.
8037 (lra_subreg_reload_pseudos): New.
8038 (lra_optional_reload_pseudos): Change comments.
8039 (lra): Init and clear lra_subreg_reload_pseudos. Clear
8040 lra_optional_reload_pseudos after undo transformations.
8041 * lra-assigns.c (pseudo_prefix_title): New.
8042 (lra_setup_reg_renumber): Use it.
8043 (spill_for): Ditto. Check subreg reload pseudos too.
8044 (assign_by_spills): Consider subreg reload pseudos too.
8045 * lra-constraints.c (simplify_operand_subreg): Use
8046 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
8047 (curr_insn_transform): Recognize and do optional reloads.
8048 (undo_optional_reloads): New.
8049 (lra_undo_inheritance): Call undo_optional_reloads.
8051 2013-07-05 Thomas Quinot <quinot@adacore.com>
8053 * tree-complex.c (expand_complex_operations_1): Fix typo.
8055 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
8057 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
8058 (tune_params): New member 'const vec_costs'.
8059 * config/aarch64/aarch64.c (generic_vector_cost): New.
8060 (generic_tunings): New member 'generic_vector_cost'.
8061 (aarch64_builtin_vectorization_cost): New.
8062 (aarch64_add_stmt_cost): New.
8063 (TARGET_VECTORIZE_ADD_STMT_COST): New.
8064 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
8066 2013-07-03 Jakub Jelinek <jakub@redhat.com>
8069 * config/i386/predicates.md (vsib_address_operand): Disallow
8070 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
8072 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
8075 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
8076 expand_builtin_setjmp_receiver.
8077 (expand_label): Adjust, call expand_builtin_setjmp_receiver
8078 with NULL for the label parameter.
8079 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
8080 the frame-pointer. Adjust comments.
8081 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
8082 only if LABEL is non-NULL.
8084 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
8086 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
8087 (ARM_ABI_AAPCS64): Ditto.
8089 (ARM_DEFAULT_ABI): Ditto.
8091 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
8093 * config/aarch64/aarch64-builtins.c
8094 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
8095 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
8097 * config/aarch64/aarch64-simd.md
8098 (aarch64_ld1<VALL:mode>): New.
8099 (aarch64_st1<VALL:mode>): Likewise.
8100 * config/aarch64/arm_neon.h
8101 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
8103 2013-07-02 Sriraman Tallam <tmsriram@google.com>
8105 * config/i386/i386.c (gate_insert_vzeroupper): Check if
8107 (ix86_option_override_internal):Turn on all -mavx target flags by
8108 default as they are dependent on AVX anyway.
8110 2013-07-02 Cary Coutant <ccoutant@google.com>
8112 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
8114 (loc_checksum_ordered): Likewise.
8115 (hash_loc_operands): Remove inline keyword.
8117 2013-07-02 Jakub Jelinek <jakub@redhat.com>
8119 PR tree-optimization/57741
8120 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
8121 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
8122 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
8123 Allow REAL_CST step_exprs if flag_associative_math.
8124 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
8126 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8128 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
8130 2013-07-02 Ian Bolton <ian.bolton@arm.com>
8132 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
8134 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8136 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
8138 (iorsi3_insn): Likewise.
8139 (arm_xorsi3): Likewise.
8141 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
8143 * arm.md (attribute "wtype"): Delete. Move attribute values from here
8145 (attribute "type"): ... here, and prefix with "wmmx_".
8146 (attribute "core_cycles"): Update for attribute changes.
8147 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
8148 (tbcstv4hi): Likewise.
8149 (tbcstv2si): Likewise.
8150 (iwmmxt_iordi3): Likewise.
8151 (iwmmxt_xordi3): Likewise.
8152 (iwmmxt_anddi3): Likewise.
8153 (iwmmxt_nanddi3): Likewise.
8154 (iwmmxt_arm_movdi): Likewise.
8155 (iwmmxt_movsi_insn): Likewise.
8156 (mov<mode>_internal): Likewise.
8157 (and<mode>3_iwmmxt): Likewise.
8158 (ior<mode>3_iwmmxt): Likewise.
8159 (xor<mode>3_iwmmxt): Likewise.
8160 (add<mode>3_iwmmxt): Likewise.
8161 (ssaddv8qi3): Likewise.
8162 (ssaddv4hi3): Likewise.
8163 (ssaddv2si3): Likewise.
8164 (usaddv8qi3): Likewise.
8165 (usaddv4hi3): Likewise.
8166 (usaddv2si3): Likewise.
8167 (sub<mode>3_iwmmxt): Likewise.
8168 (sssubv8qi3): Likewise.
8169 (sssubv4hi3): Likewise.
8170 (sssubv2si3): Likewise.
8171 (ussubv8qi3): Likewise.
8172 (ussubv4hi3): Likewise.
8173 (ussubv2si3): Likewise.
8174 (mulv4hi3_iwmmxt): Likewise.
8175 (smulv4hi3_highpart): Likewise.
8176 (umulv4hi3_highpart): Likewise.
8177 (iwmmxt_wmacs): Likewise.
8178 (iwmmxt_wmacsz): Likewise.
8179 (iwmmxt_wmacu): Likewise.
8180 (iwmmxt_wmacuz): Likewise.
8181 (iwmmxt_clrdi): Likewise.
8182 (iwmmxt_clrv8qi): Likewise.
8183 (iwmmxt_clr4hi): Likewise.
8184 (iwmmxt_clr2si): Likewise.
8185 (iwmmxt_uavgrndv8qi3): Likewise.
8186 (iwmmxt_uavgrndv4hi3): Likewise.
8187 (iwmmxt_uavgv8qi3): Likewise.
8188 (iwmmxt_uavgv4hi3): Likewise.
8189 (iwmmxt_tinsrb): Likewise.
8190 (iwmmxt_tinsrh): Likewise.
8191 (iwmmxt_tinsrw): Likewise.
8192 (iwmmxt_textrmub): Likewise.
8193 (iwmmxt_textrmsb): Likewise.
8194 (iwmmxt_textrmuh): Likewise.
8195 (iwmmxt_textrmsh): Likewise.
8196 (iwmmxt_textrmw): Likewise.
8197 (iwmxxt_wshufh): Likewise.
8198 (eqv8qi3): Likewise.
8199 (eqv4hi3): Likewise.
8200 (eqv2si3): Likewise.
8201 (gtuv8qi3): Likewise.
8202 (gtuv4hi3): Likewise.
8203 (gtuv2si3): Likewise.
8204 (gtv8qi3): Likewise.
8205 (gtv4hi3): Likewise.
8206 (gtv2si3): Likewise.
8207 (smax<mode>3_iwmmxt): Likewise.
8208 (umax<mode>3_iwmmxt): Likewise.
8209 (smin<mode>3_iwmmxt): Likewise.
8210 (umin<mode>3_iwmmxt): Likewise.
8211 (iwmmxt_wpackhss): Likewise.
8212 (iwmmxt_wpackwss): Likewise.
8213 (iwmmxt_wpackdss): Likewise.
8214 (iwmmxt_wpackhus): Likewise.
8215 (iwmmxt_wpackwus): Likewise.
8216 (iwmmxt_wpackdus): Likewise.
8217 (iwmmxt_wunpckihb): Likewise.
8218 (iwmmxt_wunpckihh): Likewise.
8219 (iwmmxt_wunpckihw): Likewise.
8220 (iwmmxt_wunpckilb): Likewise.
8221 (iwmmxt_wunpckilh): Likewise.
8222 (iwmmxt_wunpckilw): Likewise.
8223 (iwmmxt_wunpckehub): Likewise.
8224 (iwmmxt_wunpckehuh): Likewise.
8225 (iwmmxt_wunpckehuw): Likewise.
8226 (iwmmxt_wunpckehsb): Likewise.
8227 (iwmmxt_wunpckehsh): Likewise.
8228 (iwmmxt_wunpckehsw): Likewise.
8229 (iwmmxt_wunpckelub): Likewise.
8230 (iwmmxt_wunpckeluh): Likewise.
8231 (iwmmxt_wunpckeluw): Likewise.
8232 (iwmmxt_wunpckelsb): Likewise.
8233 (iwmmxt_wunpckelsh): Likewise.
8234 (iwmmxt_wunpckelsw): Likewise.
8235 (ror<mode>3): Likewise.
8236 (ashr<mode>3_iwmmxt): Likewise.
8237 (lshr<mode>3_iwmmxt): Likewise.
8238 (ashl<mode>3_iwmmxt): Likewise.
8239 (ror<mode>3_di): Likewise.
8240 (ashr<mode>3_di): Likewise.
8241 (lshr<mode>3_di): Likewise.
8242 (ashl<mode>3_di): Likewise.
8243 (iwmmxt_wmadds): Likewise.
8244 (iwmmxt_wmaddu): Likewise.
8245 (iwmmxt_tmia): Likewise.
8246 (iwmmxt_tmiaph): Likewise.
8247 (iwmmxt_tmiabb): Likewise.
8248 (iwmmxt_tmiatb): Likewise.
8249 (iwmmxt_tmiabt): Likewise.
8250 (iwmmxt_tmiatt): Likewise.
8251 (iwmmxt_tmovmskb): Likewise.
8252 (iwmmxt_tmovmskh): Likewise.
8253 (iwmmxt_tmovmskw): Likewise.
8254 (iwmmxt_waccb): Likewise.
8255 (iwmmxt_wacch): Likewise.
8256 (iwmmxt_waccw): Likewise.
8257 (iwmmxt_waligni): Likewise.
8258 (iwmmxt_walignr): Likewise.
8259 (iwmmxt_walignr0): Likewise.
8260 (iwmmxt_walignr1): Likewise.
8261 (iwmmxt_walignr2): Likewise.
8262 (iwmmxt_walignr3): Likewise.
8263 (iwmmxt_wsadb): Likewise.
8264 (iwmmxt_wsadh): Likewise.
8265 (iwmmxt_wsadbz): Likewise.
8266 (iwmmxt_wsadhz): Likewise.
8267 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
8268 (iwmmxt_wabsdiffb): Likewise.
8269 (iwmmxt_wabsdiffh): Likewise.
8270 (iwmmxt_wabsdiffw): Likewise.
8271 (iwmmxt_waddsubhx): Likewise
8272 (iwmmxt_wsubaddhx): Likewise.
8273 (addc<mode>3): Likewise.
8274 (iwmmxt_avg4): Likewise.
8275 (iwmmxt_avg4r): Likewise.
8276 (iwmmxt_wmaddsx): Likewise.
8277 (iwmmxt_wmaddux): Likewise.
8278 (iwmmxt_wmaddsn): Likewise.
8279 (iwmmxt_wmaddun): Likewise.
8280 (iwmmxt_wmulwsm): Likewise.
8281 (iwmmxt_wmulwum): Likewise.
8282 (iwmmxt_wmulsmr): Likewise.
8283 (iwmmxt_wmulumr): Likewise.
8284 (iwmmxt_wmulwsmr): Likewise.
8285 (iwmmxt_wmulwumr): Likewise.
8286 (iwmmxt_wmulwl): Likewise.
8287 (iwmmxt_wqmulm): Likewise.
8288 (iwmmxt_wqmulwm): Likewise.
8289 (iwmmxt_wqmulmr): Likewise.
8290 (iwmmxt_wqmulwmr): Likewise.
8291 (iwmmxt_waddbhusm): Likewise.
8292 (iwmmxt_waddbhusl): Likewise.
8293 (iwmmxt_wqmiabb): Likewise.
8294 (iwmmxt_wqmiabt): Likewise.
8295 (iwmmxt_wqmiatb): Likewise.
8296 (iwmmxt_wqmiatt): Likewise.
8297 (iwmmxt_wqmiabbn): Likewise.
8298 (iwmmxt_wqmiabtn): Likewise.
8299 (iwmmxt_wqmiatbn): Likewise.
8300 (iwmmxt_wqmiattn): Likewise.
8301 (iwmmxt_wmiabb): Likewise.
8302 (iwmmxt_wmiabt): Likewise.
8303 (iwmmxt_wmiatb): Likewise.
8304 (iwmmxt_wmiatt): Likewise.
8305 (iwmmxt_wmiabbn): Likewise.
8306 (iwmmxt_wmiabtn): Likewise.
8307 (iwmmxt_wmiatbn): Likewise.
8308 (iwmmxt_wmiattn): Likewise.
8309 (iwmmxt_wmiawbb): Likewise.
8310 (iwmmxt_wmiawbt): Likewise.
8311 (iwmmxt_wmiawtb): Likewise.
8312 (iwmmxt_wmiawtt): Likewise.
8313 (iwmmxt_wmiawbbn): Likewise.
8314 (iwmmxt_wmiawbtn): Likewise.
8315 (iwmmxt_wmiawtbn): Likewise.
8316 (iwmmxt_wmiawttn): Likewise.
8317 (iwmmxt_wmerge): Likewise.
8318 (iwmmxt_tandc<mode>3): Likewise.
8319 (iwmmxt_torc<mode>3): Likewise.
8320 (iwmmxt_torvsc<mode>3): Likewise.
8321 (iwmmxt_textrc<mode>3): Likewise.
8322 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
8323 (wmmxt_pack): Likewise.
8324 (wmmxt_mult_c1): Likewise.
8325 (wmmxt_mult_c2): Likewise.
8326 (wmmxt_alu_c1): Likewise.
8327 (wmmxt_alu_c2): Likewise.
8328 (wmmxt_alu_c3): Likewise.
8329 (wmmxt_transfer_c1): Likewise.
8330 (wmmxt_transfer_c2): Likewise.
8331 (wmmxt_transfer_c3): Likewise.
8332 (marvell_f_iwmmxt_wstr): Likewise.
8333 (marvell_f_iwmmxt_wldr): Likewise.
8335 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
8337 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
8339 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8342 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8343 * lra-constraints.c (need_for_split_p): Check call used hard regs
8344 living through calls.
8346 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
8347 call used regs for call insn.
8349 2013-06-28 Jakub Jelinek <jakub@redhat.com>
8352 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
8353 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
8355 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8357 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
8359 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8361 * lra-constraints.c (need_for_split_p): Check call used hard regs
8362 living through calls.
8364 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
8367 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
8368 to tie with any other modes. Eliminate Altivec vector mode tests,
8369 since these are a subset of ALTIVEC or VSX vector modes. Simplify
8370 code, to return 0 if testing MODE2 for a condition, if we've
8371 already tested MODE1 for the same condition.
8373 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8375 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
8378 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8380 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8381 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
8383 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8385 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
8387 (aarch64_symbolic_constant_p): Remove.
8388 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
8389 static. Fix line length and white space.
8390 (aarch64_symbolic_constant_p): Remove.
8391 * config/aarch64/predicates.md (aarch64_valid_symref):
8392 Use aarch64_classify_symbol_expression.
8394 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8396 * config/arm/constraints.md (Ts): New constraint.
8397 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
8399 (compare_scc): Use "Ts" constraint for operand 0.
8400 (ior_scc_scc): Likewise.
8401 (and_scc_scc): Likewise.
8402 (and_scc_scc_nodom): Likewise.
8403 (ior_scc_scc_cmp): Likewise for operand 7.
8404 (and_scc_scc_cmp): Likewise.
8405 * config/arm/thumb2.md (thumb2_movsi_insn):
8406 Add alternatives for 16-bit encodings.
8407 (thumb2_movhi_insn): Likewise.
8408 (thumb2_movsicc_insn): Likewise.
8409 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
8410 (thumb2_negscc): Use "Ts" constraint.
8411 Move mvn instruction outside cond_exec block.
8412 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
8413 for 16-bit encodings.
8415 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8417 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
8419 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
8420 (mulsi3subsi): Likewise.
8421 (mulsidi3adddi): Likewise.
8422 (mulsidi3_v6): Likewise.
8423 (umulsidi3_v6): Likewise.
8424 (umulsidi3adddi_v6): Likewise.
8425 (smulsi3_highpart_v6): Likewise.
8426 (umulsi3_highpart_v6): Likewise.
8427 (mulhisi3tb): Likewise.
8428 (mulhisi3bt): Likewise.
8429 (mulhisi3tt): Likewise.
8430 (maddhisi4): Likewise.
8431 (maddhisi4tb): Likewise.
8432 (maddhisi4tt): Likewise.
8433 (maddhidi4): Likewise.
8434 (maddhidi4tb): Likewise.
8435 (maddhidi4tt): Likewise.
8436 (zeroextractsi_compare0_scratch): Likewise.
8437 (insv_zero): Likewise.
8438 (insv_t2): Likewise.
8439 (anddi_notzesidi_di): Likewise.
8440 (anddi_notsesidi_di): Likewise.
8441 (andsi_notsi_si): Likewise.
8442 (iordi_zesidi_di): Likewise.
8443 (xordi_zesidi_di): Likewise.
8444 (andsi_iorsi3_notsi): Likewise.
8446 (smax_m1): Likewise.
8448 (not_shiftsi): Likewise.
8449 (unaligned_loadsi): Likewise.
8450 (unaligned_loadhis): Likewise.
8451 (unaligned_loadhiu): Likewise.
8452 (unaligned_storesi): Likewise.
8453 (unaligned_storehi): Likewise.
8454 (extv_reg): Likewise.
8455 (extzv_t2): Likewise.
8457 (udivsi3): Likewise.
8458 (arm_zero_extendhisi2addsi): Likewise.
8459 (arm_zero_extendqisi2addsi): Likewise.
8460 (compareqi_eq0): Likewise.
8461 (arm_extendhisi2_v6): Likewise.
8462 (arm_extendqisi2addsi): Likewise.
8463 (arm_movt): Likewise.
8464 (thumb2_ldrd): Likewise.
8465 (thumb2_ldrd_base): Likewise.
8466 (thumb2_ldrd_base_neg): Likewise.
8467 (thumb2_strd): Likewise.
8468 (thumb2_strd_base): Likewise.
8469 (thumb2_strd_base_neg): Likewise.
8470 (arm_negsi2): Add alternative for 16-bit encoding.
8471 (arm_one_cmplsi2): Likewise.
8473 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8475 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8476 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8477 (movdfcc): Likewise.
8478 * config/arm/vfp.md (*thumb2_movsf_vfp):
8479 Disable predication for arm_restrict_it.
8480 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8481 (*thumb2_movdfcc_vfp): Likewise.
8482 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8483 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8484 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8485 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8486 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8487 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8488 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8489 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8490 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8491 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8492 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8493 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8494 Disable predication for arm_restrict_it.
8496 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8498 * config/i386/bmiintrin.h (_bextr_u32): New.
8499 (_bextr_u64): Ditto.
8501 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8503 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8504 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8505 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8506 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8507 * config/mips/n32-elf.h: ...this new file.
8509 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
8512 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8513 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8515 2013-06-27 Catherine Moore <clm@codesourcery.com>
8517 * config/mips/mips-tables.opt: Regenerate.
8518 * config/mips/mips-cpus.def: Add m14ke and m14kec.
8519 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8520 * doc/invoke.texi: Add -m14kc.
8522 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8525 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8526 constraints of operand 1 and 2.
8529 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8530 to match RTL canonicalization. Swap predicates and
8531 constraints of operand 1 and 2.
8533 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8535 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8536 Process OP_INOUT regs for splitting too.
8538 2013-06-27 Jakub Jelinek <jakub@redhat.com>
8540 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8541 decl before the loop, initialize to NULL.
8542 (vectorizable_load): Initialize ptr_incr to NULL.
8544 2013-06-27 Martin Jambor <mjambor@suse.cz>
8547 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8548 * ipa-ref.c (ipa_maybe_record_reference): New function.
8549 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8550 * ipa-cp.c (create_specialized_node): Record potential references from
8552 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8554 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8556 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8557 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8558 'mode' to force_reg.
8559 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8560 (aarch64_expand_mov_immediate): Likewise.
8562 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8564 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8565 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8567 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8569 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8570 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8571 (struct machine_function): Add tbegin_p.
8572 (s390_canonicalize_comparison): Fold CC mode compares to
8573 conditional jump if possible.
8574 (s390_emit_jump): Return the emitted jump.
8575 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8576 Handle CCRAWmode compares.
8577 (s390_option_override): Default to -mhtm if available.
8578 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8579 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8580 FPRs instead of df_regs_ever_live_p.
8581 (s390_optimize_nonescaping_tx): New function.
8582 (s390_init_frame_layout): Extend clobbered_regs array to cover
8584 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8585 (s390_expand_tbegin): New function.
8586 (enum s390_builtin): New enum definition.
8587 (code_for_builtin): New array definition.
8588 (s390_init_builtins): New function.
8589 (s390_expand_builtin): New function.
8590 (TARGET_INIT_BUILTINS): Define.
8591 (TARGET_EXPAND_BUILTIN): Define.
8592 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
8593 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8594 (s390_alc_comparison): Likewise.
8595 * config/s390/s390-modes.def: Add CCRAWmode.
8596 * config/s390/s390.h (processor_flags): Add PF_TX.
8597 (TARGET_CPU_HTM): Define macro.
8598 (TARGET_HTM): Define macro.
8599 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8600 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8601 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8602 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
8603 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
8604 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8605 ("*cc_to_int"): Move up.
8606 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8607 constants other than 0.
8608 ("*ccraw_to_int"): New insn and splitter definition.
8609 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8610 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8611 ("tx_assist"): New expander.
8612 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8613 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8614 * config/s390/s390.opt: Add -mhtm option.
8615 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8616 * config/s390/htmxlintrin.h: New file.
8617 * config/s390/htmintrin.h: New file.
8618 * config/s390/s390intrin.h: New file.
8619 * doc/extend.texi: Document htm builtins.
8620 * config.gcc: Add the new header files to extra_headers.
8622 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8624 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
8625 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
8627 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
8628 Pat Haugen <pthaugen@us.ibm.com>
8629 Peter Bergner <bergner@vnet.ibm.com>
8631 * config/rs6000/power8.md: New.
8632 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8633 setting for power8 entry.
8634 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8635 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8636 test for Power4/Power5 only.
8637 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8639 (force_new_group): Adjust comment.
8640 * config/rs6000/rs6000.md: Include power8.md.
8642 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8644 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8645 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8647 (tune_params): Update comment.
8648 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8649 (arm_max_conditional_execute): New function.
8650 (thumb2_final_prescan_insn): Use max_insn_skipped and
8651 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8653 2013-06-25 Jakub Jelinek <jakub@redhat.com>
8655 PR tree-optimization/57705
8656 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8657 SSA_NAME step, provided that it is not defined inside the loop.
8658 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
8659 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8661 2013-06-25 Martin Jambor <mjambor@suse.cz>
8664 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8665 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8667 (ipa_note_param_call): Initialize member_ptr flag.
8668 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8669 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8670 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8671 (ipa_read_indirect_edge_info): Likewise.
8673 2013-06-25 Richard Biener <rguenther@suse.de>
8676 * passes.c (init_optimization_passes): Move pass_fold_builtins
8677 and pass_dce earlier with -Og.
8679 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8681 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8682 <BIT_FIELD_REF>: Remove trailing TAB.
8683 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8686 2013-06-24 Martin Jambor <mjambor@suse.cz>
8688 PR tree-optimization/57358
8689 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8690 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8691 (ipa_analyze_params_uses): Generate pessimistic info when true.
8693 2013-06-24 Martin Jambor <mjambor@suse.cz>
8695 PR tree-optimization/57539
8696 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8697 global.inlined_to of the new node to it. All callers changed.
8698 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8699 inlining_into, pass it to cgraph_clone_node.
8700 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8701 ipa_free_edge_args_substructures.
8702 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8703 rdesc linked list. Do not assert rdesc edges have inlined caller.
8704 Assert we have found an rdesc in the rdesc list.
8706 2013-06-24 Richard Biener <rguenther@suse.de>
8708 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
8709 (pointer_set_lookup): Declare.
8710 (class pointer_map): New template class implementing a
8711 generic pointer to T map.
8712 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8713 pointer_map<T>::contains, pointer_map<T>::insert,
8714 pointer_map<T>::traverse): New functions.
8715 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8716 (pointer_set_lookup): New function.
8717 (pointer_set_contains): Use pointer_set_lookup.
8718 (pointer_set_insert): Likewise.
8719 (insert_aux): Remove.
8720 (struct pointer_map_t): Embed a pointer_set_t.
8721 (pointer_map_create): Adjust.
8722 (pointer_map_destroy): Likewise.
8723 (pointer_map_contains): Likewise.
8724 (pointer_map_insert): Likewise.
8725 (pointer_map_traverse): Likewise.
8726 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8727 pointer_map<unsigned> instead of a pointer_map_t.
8728 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8729 (streamer_tree_cache_lookup): Likewise.
8730 (streamer_tree_cache_create): Likewise.
8731 (streamer_tree_cache_delete): Likewise.
8732 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8733 pointer_map<unsigned> instead of a pointer_map_t.
8734 (lto_init_tree_ref_encoder): Adjust.
8735 (lto_destroy_tree_ref_encoder): Likewise.
8736 * lto-section-out.c (lto_output_decl_index): Likewise.
8737 (lto_record_function_out_decl_state): Likewise.
8738 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8740 2013-06-24 Richard Biener <rguenther@suse.de>
8742 PR tree-optimization/57488
8743 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8745 2013-06-24 Alan Modra <amodra@gmail.com>
8747 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8748 (gen_easy_altivec_constant): Likewise.
8749 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8750 easy_vector_constant_msb): Likewise.
8752 2013-06-23 Jakub Jelinek <jakub@redhat.com>
8755 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8756 add missing return true.
8758 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
8761 * config/sh/predicates.md (general_extend_operand): Invoke
8762 general_movsrc_operand for memory operands.
8763 (general_movsrc_operand): Allow reg+reg addressing, do not use
8764 general_operand for memory operands.
8766 2013-06-23 Sriraman Tallam <tmsriram@google.com>
8768 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8769 when current target options does not apply.
8770 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8771 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
8772 * config/i386/bmiintrin.h: Pass appropriate target
8773 attributes to header.
8774 * config/i386/mmintrin.h: Ditto.
8775 * config/i386/nmmintrin.h: Ditto.
8776 * config/i386/avx2intrin.h: Ditto.
8777 * config/i386/fxsrintrin.h: Ditto.
8778 * config/i386/tbmintrin.h: Ditto.
8779 * config/i386/xsaveintrin.h: Ditto.
8780 * config/i386/f16cintrin.h: Ditto.
8781 * config/i386/xtestintrin.h: Ditto.
8782 * config/i386/xsaveoptintrin.h: Ditto.
8783 * config/i386/bmi2intrin.h: Ditto.
8784 * config/i386/lzcntintrin.h: Ditto.
8785 * config/i386/smmintrin.h: Ditto.
8786 * config/i386/wmmintrin.h: Ditto.
8787 * config/i386/x86intrin.h: Remove all header include guards.
8788 * config/i386/prfchwintrin.h: Ditto.
8789 * config/i386/pmmintrin.h: Ditto.
8790 * config/i386/tmmintrin.h: Ditto.
8791 * config/i386/xmmintrin.h: Ditto.
8792 * config/i386/popcntintrin.h: Ditto.
8793 * config/i386/rdseedintrin.h: Ditto.
8794 * config/i386/ammintrin.h: Ditto.
8795 * config/i386/emmintrin.h: Ditto.
8796 * config/i386/immintrin.h: Remove all header include guards.
8797 * config/i386/fma4intrin.h: Ditto.
8798 * config/i386/lwpintrin.h: Ditto.
8799 * config/i386/xopintrin.h: Ditto.
8800 * config/i386/ia32intrin.h: Ditto.
8801 * config/i386/avxintrin.h: Ditto.
8802 * config/i386/rtmintrin.h: Ditto.
8803 * config/i386/fmaintrin.h: Ditto.
8804 * config/i386/mm3dnow.h: Ditto.
8806 2013-06-22 Sriraman Tallam <tmsriram@google.com>
8808 * common/config/i386/i386-common.c: Handle LZCNT.
8810 2013-06-22 Andi Kleen <ak@linux.intel.com>
8812 * doc/extend.texi: Use __atomic_store_n instead of
8813 __atomic_store in HLE example.
8815 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
8817 * config/sh/sh.c: Remove <cstdlib> workaround.
8819 2013-06-21 Andi Kleen <ak@linux.intel.com>
8821 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
8823 2013-06-21 Andi Kleen <ak@linux.intel.com>
8825 * doc/extend.texi: Document that __atomic_clear and
8826 __atomic_test_and_set should only be used with bool.
8828 2013-06-20 Jan Hubicka <jh@suse.cz>
8830 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8832 * tree.c (decls_same_for_odr): New function.
8833 (same_for_edr): New function.
8834 (types_same_for_odr): New function.
8835 (get_binfo_at_offset): Use it.
8836 * tree.h (types_same_for_odr): Declare.
8838 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
8839 Jason Merrill <jason@redhat.com>
8841 * system.h: Include <cstdlib> as well as <stdlib.h>.
8843 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
8846 * config/i386/i386.c (construct_container): Report error if
8847 long double is used with disabled x87 float returns.
8849 2013-06-20 Jan Hubicka <jh@suse.cz>
8851 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8853 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
8855 PR rtl-optimization/57425
8856 PR rtl-optimization/57569
8857 * alias.c (write_dependence_p): Remove parameters mem_mode and
8858 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8859 Changed all callers.
8860 (canon_anti_dependence): Get comments and semantics in sync.
8861 Add parameter mem_canonicalized. Changed all callers.
8862 * rtl.h (canon_anti_dependence): Update prototype.
8864 2013-06-20 Richard Biener <rguenther@suse.de>
8866 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8867 byte case, inline streamer_read_uchar and defer section
8870 2013-06-20 Richard Biener <rguenther@suse.de>
8872 PR tree-optimization/57584
8873 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8874 SSA names into the expanded expression that take part in
8875 abnormal coalescing.
8877 2013-06-19 Sharad Singhai <singhai@google.com>
8879 * gcov.c (print_usage): Handle new option.
8880 (process_args): Ditto.
8881 (get_gcov_intermediate_filename): New function.
8882 (output_intermediate_file): New function.
8883 (output_gcov_file): New function
8884 (generate_results): Handle new option.
8885 (release_function): Relase demangled name.
8886 (read_graph_file): Handle demangled name.
8887 (output_lines): Ditto.
8888 * doc/gcov.texi: Document gcov intermediate format.
8890 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8893 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8894 (lra_emit_add): Use the functions. Add comment about Y as an
8897 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
8900 * collect2.c (collect_atexit): New.
8901 (collect_exit): Delete.
8902 (main): Register collect_atexit with atexit.
8903 (collect_wait): Change collect_exit to exit.
8905 * collect2.h (collect_exit): Delete.
8906 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8908 2013-06-19 Wei Mi <wmi@google.com>
8910 PR rtl-optimization/57518
8911 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8912 if regno is used in paradoxical subreg.
8913 (update_equiv_regs): Check pdx_subregs[regno] before
8914 set a reg to be equivalent with a mem.
8916 2013-06-19 Matthias Klose <doko@ubuntu.com>
8919 * file-find.h (find_a_file): Add a mode parameter.
8920 * file-find.c (find_a_file): Likewise.
8921 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8922 with X_OK for the executables.
8923 * collect2.c (main): Call find_a_file with X_OK.
8925 2013-06-19 Steve Ellcey <sellcey@mips.com>
8928 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8929 Use NEXT_INSN instead of next_real_insn.
8931 2013-06-19 Jan Hubicka <jh@suse.cz>
8933 * cgraph.h (const_value_known_p): Replace by ...
8934 (ctor_for_folding): .. this one.
8935 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8936 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8937 * expr.c (expand_expr_real_1): Likewise.
8938 (string_constant): Likewise.
8939 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8940 * ipa.c (process_references): Likewise.
8941 (symtab_remove_unreachable_nodes): Likewise.
8942 * ipa-inline-analysis.c (param_change_prob): Likewise.
8943 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8944 (get_base_constructor): Likwise.
8945 * varpool.c (varpool_remove_node): Likewise.
8946 (varpool_remove_initializer): LIkewise.
8947 (dump_varpool_node): LIkwise.
8948 (const_value_known_p): Rewrite to ...
8949 (ctor_for_folding): ... this one.
8951 2013-06-19 Jakub Jelinek <jakub@redhat.com>
8954 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8955 PERSONALITY in $PATH derived prefixes.
8957 2013-06-19 Jeff Law <law@redhat.com>
8959 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8962 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8963 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8966 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8968 * config/arm/vfp.md: Move VFP instruction classification documentation
8970 * config/arm/arm.md: ... here. Update instruction classification
8973 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
8975 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8976 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8977 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8979 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8982 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8983 with NEXT_INSN. Use tablejump_p to check for jump table data
8986 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8989 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8990 that now in C++ the value is correct per the C++ standards.
8992 2013-06-19 Richard Biener <rguenther@suse.de>
8994 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8997 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9000 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9003 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9004 with next_active_insn.
9006 2013-06-18 Sriraman Tallam <tmsriram@google.com>
9008 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
9009 functions are inlined during failures to flag an error.
9010 * tree-inline.c (expand_call_inline): Allow the error to be flagged
9011 in early inline pass.
9013 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
9015 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
9018 2013-06-18 Julian Brown <julian@codesourcery.com>
9020 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
9021 Permit virtual register pre-reload if !strict.
9022 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
9024 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
9026 * config/arm/neon.md (movmisalign<mode>): Use
9027 neon_perm_struct_or_reg_operand instead of
9028 neon_struct_or_register_operand.
9029 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
9030 neon_permissive_struct_operand instead of neon_struct_operand.
9031 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
9032 neon_vector_mem_operand.
9033 * config/arm/predicates.md (neon_struct_operand): Adjust call to
9034 neon_vector_mem_operand.
9035 (neon_permissive_struct_operand): New.
9036 (neon_struct_or_register_operand): Rename to...
9037 (neon_perm_struct_or_reg_operand): This. Adjust call to
9038 neon_vector_mem_operand.
9040 2013-06-18 Richard Biener <rguenther@suse.de>
9042 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
9043 * lto-streamer.h: Include pointer-set.h.
9044 (struct lto_decl_slot): Remove.
9045 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
9046 Remove next_index entry.
9047 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9048 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9049 (lto_init_tree_ref_encoder): Adjust.
9050 (lto_destroy_tree_ref_encoder): Likewise.
9051 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
9052 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
9053 (lto_output_decl_index): Adjust.
9054 (lto_new_out_decl_state): Likewise.
9055 (lto_record_function_out_decl_state): Likewise.
9056 * lto-streamer-out.c (copy_function): Likewise.
9058 2013-06-18 Richard Biener <rguenther@suse.de>
9060 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
9061 * cgraphunit.c: Include cfgloop.h.
9062 (init_lowered_empty_function): Initialize the loop tree.
9063 (assemble_thunk): Insert new BBs into loops.
9065 2013-06-18 Richard Biener <rguenther@suse.de>
9067 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
9068 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
9069 the map from cache entry to cache index optional.
9070 (streamer_tree_cache_replace_tree): Adjust accordingly.
9071 (streamer_tree_cache_append): Likewise.
9072 (streamer_tree_cache_delete): Likewise.
9073 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
9074 streamer cache map from cache entry to cache index.
9075 * lto-streamer-out.c (create_output_block): Adjust.
9077 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
9079 * config/arm/arm.md (attribute "insn"): Move multiplication and
9080 division attributes to...
9081 (attribute "type"): ... here. Remove mult.
9082 (attribute "mul32"): New attribute.
9083 (attribute "mul64"): Add umaal.
9084 (*arm_mulsi3): Update attributes.
9085 (*arm_mulsi3_v6): Likewise.
9086 (*thumb_mulsi3): Likewise.
9087 (*thumb_mulsi3_v6): Likewise.
9088 (*mulsi3_compare0): Likewise.
9089 (*mulsi3_compare0_v6): Likewise.
9090 (*mulsi_compare0_scratch): Likewise.
9091 (*mulsi_compare0_scratch_v6): Likewise.
9092 (*mulsi3addsi): Likewise.
9093 (*mulsi3addsi_v6): Likewise.
9094 (*mulsi3addsi_compare0): Likewise.
9095 (*mulsi3addsi_compare0_v6): Likewise.
9096 (*mulsi3addsi_compare0_scratch): Likewise.
9097 (*mulsi3addsi_compare0_scratch_v6): Likewise.
9098 (*mulsi3subsi): Likewise.
9099 (*mulsidi3adddi): Likewise.
9100 (*mulsi3addsi_v6): Likewise.
9101 (*mulsidi3adddi_v6): Likewise.
9102 (*mulsidi3_nov6): Likewise.
9103 (*mulsidi3_v6): Likewise.
9104 (*umulsidi3_nov6): Likewise.
9105 (*umulsidi3_v6): Likewise.
9106 (*umulsidi3adddi): Likewise.
9107 (*umulsidi3adddi_v6): Likewise.
9108 (*smulsi3_highpart_nov6): Likewise.
9109 (*smulsi3_highpart_v6): Likewise.
9110 (*umulsi3_highpart_nov6): Likewise.
9111 (*umulsi3_highpart_v6): Likewise.
9112 (mulhisi3): Likewise.
9113 (*mulhisi3tb): Likewise.
9114 (*mulhisi3bt): Likewise.
9115 (*mulhisi3tt): Likewise.
9116 (maddhisi4): Likewise.
9117 (*maddhisi4tb): Likewise.
9118 (*maddhisi4tt): Likewise.
9119 (maddhidi4): Likewise.
9120 (*maddhidi4tb): Likewise.
9121 (*maddhidi4tt): Likewise.
9123 (udivsi3): Likewise.
9124 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
9125 (thumb2_mulsi_short_compare0): Likewise.
9126 (thumb2_mulsi_short_compare0_scratch): Likewise.
9127 * config/arm/arm1020e.md (1020mult1): Update attribute change.
9128 (1020mult2): Likewise.
9129 (1020mult3): Likewise.
9130 (1020mult4): Likewise.
9131 (1020mult5): Likewise.
9132 (1020mult6): Likewise.
9133 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
9135 (cortex_a15_mult64): Likewise.
9136 (cortex_a15_sdiv): Likewise.
9137 (cortex_a15_udiv): Likewise.
9138 * config/arm/arm1026ejs.md (mult1): Update attribute change.
9144 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
9145 (pj4_ir_div): Likewise.
9146 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
9147 (11_mult2): Likewise.
9148 (11_mult3): Likewise.
9149 (11_mult4): Likewise.
9150 (11_mult5): Likewise.
9151 (11_mult6): Likewise.
9152 (11_mult7): Likewise.
9153 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
9154 (cortex_a8_mla): Likewise.
9155 (cortex_a8_mull): Likewise.
9156 (cortex_a8_smulwy): Likewise.
9157 (cortex_a8_smlald): Likewise.
9158 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
9159 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
9160 (cortex_r4_mul_3): Likewise.
9161 (cortex_r4_mla_4): Likewise.
9162 (cortex_r4_mla_3): Likewise.
9163 (cortex_r4_smlald): Likewise.
9164 (cortex_r4_mull): Likewise.
9165 (cortex_r4_sdiv): Likewise.
9166 (cortex_r4_udiv): Likewise.
9167 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
9168 (cortex_a7_idiv): Likewise.
9169 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
9170 (9_mult2): Likewise.
9171 (9_mult3): Likewise.
9172 (9_mult4): Likewise.
9173 (9_mult5): Likewise.
9174 (9_mult6): Likewise.
9175 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
9176 (cortex_a53_sdiv): Likewise.
9177 (cortex_a53_udiv): Likewise.
9178 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
9179 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
9180 (mp626_mult2): Likewise.
9181 (mp626_mult3): Likewise.
9182 (mp626_mult4): Likewise.
9183 * config/arm/fa526.md (526_mult1): Update attribute change.
9184 (526_mult2): Likewise.
9185 * config/arm/arm-generic.md (mult): Update attribute change.
9186 (mult_ldsched_strongarm): Likewise.
9187 (mult_ldsched): Likewise.
9188 (multi_cycle): Likewise.
9189 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
9190 * config/arm/fa606te.md (606te_mult1): Update attribute change.
9191 (606te_mult2): Likewise.
9192 (606te_mult3): Likewise.
9193 (606te_mult4): Likewise.
9194 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
9195 (cortex_a9_mac16): Likewise.
9196 (cortex_a9_multiply): Likewise.
9197 (cortex_a9_mac): Likewise.
9198 (cortex_a9_multiply_long): Likewise.
9199 * config/arm/fa626te.md (626te_mult1): Update attribute change.
9200 (626te_mult2): Likewise.
9201 (626te_mult3): Likewise.
9202 (626te_mult4): Likewise.
9204 2013-06-18 Richard Biener <rguenther@suse.de>
9207 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
9209 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9212 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
9213 with next_active_insn.
9215 2013-06-18 Alan Modra <amodra@gmail.com>
9217 * config/rs6000/rs6000.h (enum data_align): New.
9218 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
9219 (DATA_ABI_ALIGNMENT): Define.
9220 (CONSTANT_ALIGNMENT): Correct comment.
9221 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
9222 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
9224 2013-06-17 David Malcolm <dmalcolm@redhat.com>
9226 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
9227 ATTRIBUTE_UNUSED marking.
9229 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
9231 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
9232 alternative and update.
9233 (aarch64_dup_lanedi): Delete.
9234 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
9235 * config/aarch64/aarch64-simd-builtins.def: Update.
9237 2013-06-17 Richard Biener <rguenther@suse.de>
9239 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
9240 (lto_input_scc): Declare.
9241 (lto_input_tree_1): Likewise.
9242 (struct lto_stats_d): Add num_tree_bodies_output and
9243 num_pickle_refs_output.
9244 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
9245 (lto_read_tree_1): Split out from ...
9246 (lto_read_tree): ... this.
9247 (lto_input_scc): New function.
9248 (lto_input_tree_1): Split out from ...
9249 (lto_input_tree): ... this. Handle LTO_tree_scc.
9250 (lto_data_in_create): Create the streamer cache without hashes.
9251 * lto-streamer-out.c (create_output_block): Create the streamer
9252 cache with hashes when not doing WPA.
9253 (lto_write_tree_1): Split out from ...
9254 (lto_write_tree): ... this.
9255 (get_symbol_initial_value): New function.
9256 (lto_output_tree_1): Split out from ...
9257 (lto_output_tree): ... this. Write trees as series of SCCs
9258 using a DFS walk via DFS_write_tree.
9259 (struct sccs, struct scc_entry): New types.
9260 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
9261 (DFS_write_tree_body): New function.
9262 (DFS_write_tree): Likewise.
9263 (hash_tree): Likewise.
9264 (scc_entry_compare): Likewise.
9265 (hash_scc): Likewise.
9266 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
9267 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
9268 TREE_CHAIN as regular reference.
9269 (streamer_read_integer_cst): Remove.
9270 (streamer_get_pickled_tree): Adjust.
9271 * tree-streamer-out.c (streamer_write_chain): Disable streaming
9272 of DECL_EXTERNALs in BLOCK_VARS for now.
9273 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
9275 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
9276 Add hash value argument and record that if hashes are recorded
9278 (streamer_tree_cache_insert_1): Adjust.
9279 (streamer_tree_cache_insert): Likewise.
9280 (streamer_tree_cache_insert_at): Rename to ...
9281 (streamer_tree_cache_replace_tree): ... this and adjust.
9282 (streamer_tree_cache_append): Adjust.
9283 (record_common_node): Likewise.
9284 (streamer_tree_cache_create): Add argument whether to
9285 record hash values together with trees.
9286 (streamer_tree_cache_delete): Adjust.
9287 * tree-streamer.h (struct streamer_tree_cache_d): Add
9289 (streamer_read_integer_cst): Remove.
9290 (streamer_tree_cache_insert): Adjust.
9291 (streamer_tree_cache_append): Likewise.
9292 (streamer_tree_cache_insert_at): Rename to ...
9293 (streamer_tree_cache_replace_tree): ... this and adjust.
9294 (streamer_tree_cache_create): Add argument whether to record hashes.
9295 (streamer_tree_cache_get): Rename to ...
9296 (streamer_tree_cache_get_tree): ... this.
9297 (streamer_tree_cache_get_hash): New function.
9298 * tree.c (cache_integer_cst): New function.
9299 * tree.h (cache_integer_cst): Declare.
9300 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
9301 * lto-symtab.c (lto_varpool_replace_node): Only release
9302 DECL_INITIAL of non-prevailing decls.
9303 * varpool.c (varpool_remove_initializer): Do not release
9304 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
9306 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9308 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
9309 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
9310 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
9311 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
9312 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
9313 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
9314 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
9315 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
9316 instead of TARGET_64BIT.
9317 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
9318 Require ISA_HAS_<D>DIV.
9320 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
9322 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
9323 (mips*-*-linux*): Move default with_llsc setting to where other
9325 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
9327 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
9328 Likewise. Remove default with_tune setting. Move default float
9329 setting to its own block. Handle with_llsc in the same block as above.
9331 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
9333 PR rtl-optimization/57425
9334 PR rtl-optimization/57569
9335 * alias.c (write_dependence_p): Add new parameters mem_mode,
9336 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
9337 Changed all callers.
9338 (canon_anti_dependence): New function.
9339 * cse.c (check_dependence): Use canon_anti_dependence.
9340 * cselib.c (cselib_invalidate_mem): Likewise.
9341 * rtl.h (canon_anti_dependence): Declare.
9343 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9345 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
9346 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
9347 ".set mips3" for 64-bit targets.
9349 2013-06-15 Dehao Chen <dehao@google.com>
9351 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
9352 * gimple-low.c (gimple_check_call_matching_types): Likewise.
9353 (gimple_check_call_args): Likewise.
9354 * value-prof.c (check_ic_target): Likewise.
9355 * ipa-inline.c (early_inliner): Likewise.
9356 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
9357 * cgraph.c (cgraph_create_edge_1): Likewise.
9358 (cgraph_make_edge_direct): Likewise.
9360 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
9363 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
9364 rs6000_output_move_128bit to handle emitting quad memory
9365 operations. Set attribute length to 8 bytes.
9367 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
9369 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
9371 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
9372 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
9373 (aarch64_<su>mlsl<mode>): Likewise.
9375 2013-06-14 Mike Stump <mikestump@comcast.net>
9377 * Makefile.in (TARGET_H): Add insn-codes.h.
9379 2013-06-14 Alan Modra <amodra@gmail.com>
9383 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
9384 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
9385 bitfield expansion when EXPAND_MEMORY.
9386 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
9388 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
9390 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
9391 test for clearing quad memory on 32-bit later.
9393 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9395 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
9396 (fold_negate_expr): Likewise.
9397 (fold_real_zero_addition_p): Handle vectors.
9398 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
9400 2013-06-14 Alan Modra <amodra@gmail.com>
9402 * varasm.c (force_const_mem): Revert 2013-06-07 change.
9404 2013-06-13 Jan Hubicka <jh@suse.cz>
9406 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
9407 Local comdats are not externally visible.
9408 * symtab.c (dump_symtab_base): Dump externally visible.
9409 (verify_symtab_base): Verify back links in the symtab hash.
9411 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9413 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
9414 CONVERT_EXPR as equal nodes.
9416 2013-06-13 Bin Cheng <bin.cheng@arm.com>
9418 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
9420 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
9422 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
9423 Generalize to complex and vector.
9424 * tree.c (build_all_ones_cst): New function.
9425 * tree.h (build_all_ones_cst): Declare it.
9427 2013-06-13 Alan Modra <amodra@gmail.com>
9429 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
9430 * config/rs6000/rs6000.md (signbittf2): New insn.
9431 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
9432 (abstf2_internal, cmptf_internal2): Likewise.
9433 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
9435 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
9436 Pat Haugen <pthaugen@us.ibm.com>
9437 Peter Bergner <bergner@vnet.ibm.com>
9439 * config/rs6000/rs6000.c (emit_load_locked): Add support for
9440 power8 byte, half-word, and quad-word atomic instructions.
9441 (emit_store_conditional): Likewise.
9442 (rs6000_expand_atomic_compare_and_swap): Likewise.
9443 (rs6000_expand_atomic_op): Likewise.
9445 * config/rs6000/sync.md (larx): Add new modes for power8.
9447 (AINT): New mode iterator to include TImode as well as normal
9448 integer modes on power8.
9449 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
9450 that VSX registers are not considered. Use AINT mode iterator
9451 instead of INT1 to allow inclusion of quad word atomic operations
9453 (load_locked<mode>): Likewise.
9454 (store_conditional<mode>): Likewise.
9455 (atomic_compare_and_swap<mode>): Likewise.
9456 (atomic_exchange<mode>): Likewise.
9457 (atomic_nand<mode>): Likewise.
9458 (atomic_fetch_<fetchop_name><mode>): Likewise.
9459 (atomic_nand_fetch<mode>): Likewise.
9460 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9462 (ATOMIC): On power8, add QImode, HImode modes.
9463 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9464 modes that promote to SImode.
9465 (load_lockedti): Convert TImode arguments to PTImode, so that we
9466 get a guaranteed even/odd register pair.
9467 (load_lockedpti): Likewise.
9468 (store_conditionalti): Likewise.
9469 (store_conditionalpti): Likewise.
9471 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9472 atomic load/store instructions.
9475 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9477 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9479 (insn_count): New attribute, with most cases extracted from...
9480 (length): ...here. Redefine most cases in terms of insn_count.
9481 (single_insn): Delete.
9482 (can_delay): Use insn_count to check for single instructions.
9483 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9484 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9485 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9486 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9487 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9488 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9489 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9490 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9491 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9492 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9493 rather than "length".
9494 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9495 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9496 Use "insn_count" rather than "length".
9497 * config/mips/mips-dsp.md
9498 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9499 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9502 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
9504 PR tree-optimization/57361
9505 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9507 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9509 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9511 (aarch64_simd_combine<mode>): New instruction expansion.
9512 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9514 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9515 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9517 2013-06-12 Jan Hubicka <jh@suse.cz>
9519 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9520 decl has when in streaming stage.
9521 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9522 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9524 2013-06-12 Roland Stigge <stigge@antcom.de>
9527 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9529 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9531 PR tree-optimization/57537
9532 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9533 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9535 2013-06-12 Richard Biener <rguenther@suse.de>
9537 * data-streamer.h (streamer_write_char_stream): CSE
9538 obs->current_pointer.
9539 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9540 streamer_write_char_stream manually and optimize the resulting loop.
9541 (streamer_write_hwi_stream): Likewise.
9543 2013-06-12 Jan Hubicka <jh@suse.cz>
9545 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9546 * cgraph.h (varpool_create_empty_node): Declare.
9547 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9549 * symtab.c (symtab_unregister_node): Be lax about missin entries
9551 (symtab_get_node): Update comment.
9552 * varpool.c (varpool_create_empty_node): Break out from ...
9553 (varpool_node_for_decl): ... here.
9554 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9556 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9558 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9559 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9560 part. Use straight-line flow at the end.
9561 <COMPONENT_REF>: Remove superfluous else.
9562 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9564 2013-06-12 Jakub Jelinek <jakub@redhat.com>
9567 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9568 target hook even for !TREE_PUBLIC decls. If no resolution info
9569 is available, return false for common and external decls.
9571 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9573 * config/rl78/constraints.md (U): New constraint.
9574 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9577 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9580 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9581 to allow returning address to AT_PLATFORM name.
9583 2013-06-11 Jan Hubicka <jh@suse.cz>
9585 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9586 * cgraph.h (symtab_node_base): Add weakref flag.
9587 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9588 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9589 (output_weakrefs): Use weakref flag.
9590 * fold-const.c (simple_operand_p): Handle WEAK.
9591 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9592 * ipa.c (varpool_externally_visible_p): Drop weakref.
9593 (function_and_variable_visibility): Update comment; fix weakref
9594 sanity checks; do not clear DECL_WEAK on them.
9595 * lto-cgraph.c (lto_output_node): update.
9596 (lto_output_varpool_node): Update.
9597 (input_overwrite_node): Update.
9598 (input_node): Update.
9599 (input_varpool_node): Update.
9600 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9601 (lto_symtab_merge_symbols): Add sanity check.
9602 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9603 * passes.c (rest_of_decl_compilation): Set static flag, too.
9604 * symtab.c (dump_symtab_base): Dump weakref.
9605 (verify_symtab_base): Sanity check weakrefs.
9606 (symtab_make_decl_local): Remove duplicated code.
9607 (symtab_alias_ultimate_target): Simplify.
9608 * varpool.c (varpool_create_variable_alias): Set weakref flag.
9610 2013-06-11 Tom de Vries <tom@codesourcery.com>
9612 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9613 sequence_vect == NULL.
9615 2013-06-11 DJ Delorie <dj@redhat.com>
9617 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9618 (rl78_unwind_word_mode): New.
9620 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9622 * final.c (debug_prefix_maps): Make static.
9624 2013-06-11 David Malcolm <dmalcolm@redhat.com>
9626 * function.c (initial_trampoline): Remove stray copy.
9628 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9630 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9632 2013-06-11 Martin Jambor <mjambor@suse.cz>
9634 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9635 within bounds at the beginning of the function.
9637 2013-06-11 Alan Modra <amodra@gmail.com>
9639 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9641 (get_named_section): Don't NULL !DECL_P decl.
9643 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9645 * doc/invoke.texi (core-avx2): Document.
9647 (atom): Updated with MOVBE.
9649 2013-06-11 Richard Biener <rguenther@suse.de>
9651 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
9653 2013-06-11 Anton Blanchard <anton@samba.org>
9655 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9656 correct shift value in little-endian mode.
9658 2013-06-11 Jakub Jelinek <jakub@redhat.com>
9661 * varasm.c (get_variable_align): Move #endif to the right place.
9663 2013-06-10 Cary Coutant <ccoutant@google.com>
9665 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9666 for hash so that hash table traversal order is deterministic.
9668 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
9669 Pat Haugen <pthaugen@us.ibm.com>
9670 Peter Bergner <bergner@vnet.ibm.com>
9672 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9673 of vectors in GPRS if they are direct moves or quad word load or
9676 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9678 (direct_move_p): Likewise.
9679 (quad_load_store_p): Likewise.
9681 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9682 classes into bins based on the physical register type.
9683 (reg_class_to_reg_type): Likewise.
9684 (IS_STD_REG_TYPE): Likewise.
9685 (IS_FP_VECT_REG_TYPE): Likewise.
9686 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9687 use direct move instructions.
9688 (reload_gpr_vsx): Likewise.
9689 (reload_vsx_gpr): Likewise.
9690 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9691 information that is a simplification of register classes. Also
9692 precalculate direct move reload helpers.
9693 (direct_move_p): New function to return true if the operation can
9694 be done as a direct move instruciton.
9695 (quad_load_store_p): New function to return true if the operation
9696 is a quad memory operation.
9697 (rs6000_legitimize_address): If quad memory, only allow register
9698 indirect for TImode addresses.
9699 (rs6000_legitimate_address_p): Likewise.
9700 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9701 (rs6000_reload_register_type): Likewise.
9702 (register_to_reg_type): Return register type.
9703 (rs6000_secondary_reload_simple_move): New helper function for
9704 secondary reload and secondary memory needed to identify anything
9705 that is a simple move, and does not need reloading.
9706 (rs6000_secondary_reload_direct_move): New helper function for
9707 secondary reload to identify cases that can be done with several
9708 instructions via the direct move instructions.
9709 (rs6000_secondary_reload_move): New helper function for secondary
9710 reload to identify moves between register types that can be done.
9711 (rs6000_secondary_reload): Add support for quad memory operations
9712 and for direct move.
9713 (rs6000_secondary_memory_needed): Likewise.
9714 (rs6000_debug_secondary_memory_needed): Change argument names.
9715 (rs6000_output_move_128bit): New function to return the move to
9716 use for 128-bit moves, including knowing about the various
9717 limitations of quad memory operations.
9719 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9720 memory operations. call rs6000_output_move_128bit for the actual
9721 instruciton(s) to generate.
9722 (vsx_movti_64bit): Likewise.
9724 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9725 (UNSPEC_P8V_MTVSRWZ): Likewise.
9726 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9727 (UNSPEC_P8V_MTVSRD): Likewise.
9728 (UNSPEC_P8V_XXPERMDI): Likewise.
9729 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9730 (UNSPEC_FUSION_GPR): Likewise.
9731 (FMOVE128_GPR): New iterator for direct move.
9732 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
9735 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9736 loads and direct move instructions.
9737 (zero_extendsidi2_lfiwzx): Likewise.
9738 (extendsidi2_lfiwax): Likewise.
9739 (extendsidi2_nocell): Likewise.
9740 (floatsi<mode>2_lfiwax): Likewise.
9742 (floatunssi<mode>2_lfiwzx): Likewise.
9744 (fix_trunc<mode>_stfiwx): Likewise.
9745 (fixuns_trunc<mode>_stfiwx): Likewise.
9746 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9747 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9748 (parity<mode>2_cmpb): Set length/type attr.
9749 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9750 for 'mr.' to fast_compare.
9751 (bpermd_<mode>): Change type attr to popcnt.
9752 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9753 (p8_mtvsrwz_1): Likewise.
9754 (p8_mtvsrwz_2): Likewise.
9755 (reload_fpr_from_gpr<mode>): Likewise.
9756 (p8_mtvsrd_1): Likewise.
9757 (p8_mtvsrd_2): Likewise.
9758 (p8_xxpermdi_<mode>): Likewise.
9759 (reload_vsx_from_gpr<mode>): Likewise.
9760 (reload_vsx_from_gprsf): Likewise.
9761 (p8_mfvsrd_3_<mode>): LIkewise.
9762 (reload_gpr_from_vsx<mode>): Likewise.
9763 (reload_gpr_from_vsxsf): Likewise.
9764 (p8_mfvsrd_4_disf): Likewise.
9765 (multi-word GPR splits): Do not split direct moves or quad memory
9768 2013-06-10 David Malcolm <dmalcolm@redhat.com>
9770 * tree-into-ssa.c (interesting_blocks): Make static.
9772 2013-06-10 Jakub Jelinek <jakub@redhat.com>
9775 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9776 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9777 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9778 (get_variable_align): New function.
9779 (get_variable_section, emit_bss, emit_common,
9780 assemble_variable_contents, place_block_symbol): Use
9781 get_variable_align instead of DECL_ALIGN.
9782 (assemble_noswitch_variable): Add align argument, use it
9783 instead of DECL_ALIGN.
9784 (assemble_variable): Adjust caller. Use get_variable_align
9785 instead of DECL_ALIGN.
9786 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9788 (DATA_ABI_ALIGNMENT): Define.
9789 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9790 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9791 opt is false, only return the psABI mandated alignment increase.
9792 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9793 (DATA_ABI_ALIGNMENT): ... this.
9794 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9795 (DATA_ABI_ALIGNMENT): ... this.
9796 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9797 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9798 (DATA_ABI_ALIGNMENT): ... this.
9799 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9800 * doc/tm.texi: Regenerated.
9802 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
9804 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9805 cmp_code to construct REG_EQUAL note.
9807 2013-06-09 Jakub Jelinek <jakub@redhat.com>
9810 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9811 that operands[2] doesn't overlap with operands[0].
9813 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
9814 Jan Hubicka <jh@suse.cz>
9816 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9817 hack to mark symbols as used.
9819 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9821 PR rtl-optimization/57559
9822 * lra-constraints.c (process_alt_operands): Don't discourage
9823 memory with known offset for offsetable memory constraint.
9824 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9826 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9828 * varasm.c (struct oc_local_state): Reorder fields.
9829 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9830 and adjust accordingly.
9831 (output_constructor): Reorder initialization code and adjust call to
9832 output_constructor_bitfield.
9834 2013-06-07 Jan Hubicka <jh@suse.cz>
9836 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9838 2013-06-07 David Malcolm <dmalcolm@redhat.com>
9840 * tree-object-size.c (unknown): Make const.
9842 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9844 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9845 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9846 for last alternative in the cpu_facility attribute.
9848 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9851 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9852 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9853 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9854 * config/arm/constraints.md (Dg): New constraint.
9855 * config/arm/neon.md (xordi3_neon): Remove.
9856 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9857 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9859 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9861 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9862 Clean up alternatives.
9864 2013-06-07 Alan Modra <amodra@gmail.com>
9866 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9869 2013-06-07 Alan Modra <amodra@gmail.com>
9871 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9873 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9875 * config/arm/constraints.md (Df): New constraint.
9876 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9877 Correct length attribute for last two alternatives.
9879 2013-06-07 Alan Modra <amodra@gmail.com>
9881 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9882 override user -mfp-in-toc.
9883 (offsettable_ok_by_alignment): Consider just the current access
9884 rather than the whole object, unless BLKmode. Handle
9885 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9886 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9887 for -mcmodel=medium.
9888 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9889 override user -mfp-in-toc or -msum-in-toc. Default to
9890 -mno-fp-in-toc for -mcmodel=medium.
9892 2013-06-06 DJ Delorie <dj@redhat.com>
9894 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9895 TARGET_VALID_POINTER_MODE.
9897 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
9898 Pat Haugen <pthaugen@us.ibm.com>
9899 Peter Bergner <bergner@vnet.ibm.com>
9901 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9902 Document new power8 builtins.
9904 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9905 condition code register, to allow 128-bit logical operations to be
9906 done in the VSX or GPR registers.
9907 (nor<mode>3): Use the canonical form for nor.
9908 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9909 vclz*, and vpopcnt* vector instructions.
9910 (nand<mode>3): Likewise.
9911 (orc<mode>3): Likewise.
9912 (clz<mode>2): LIkewise.
9913 (popcount<mode>2): Likewise.
9915 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9916 that only the GPRs are recognized.
9918 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9919 support for new power8 builtins.
9921 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9923 (xscvdpspn): Likewise.
9929 (vpopcnt): Likewise.
9930 (vpopcntb): Likewise.
9931 (vpopcnth): Likewise.
9932 (vpopcntw): Likewise.
9933 (vpopcntd): Likewise.
9938 (eqv_v16qi3): Likewise.
9939 (eqv_v8hi3): Likewise.
9940 (eqv_v4si3): Likewise.
9941 (eqv_v2di3): Likewise.
9942 (eqv_v4sf3): Likewise.
9943 (eqv_v2df3): Likewise.
9945 (nand_v16qi3): Likewise.
9946 (nand_v8hi3): Likewise.
9947 (nand_v4si3): Likewise.
9948 (nand_v2di3): Likewise.
9949 (nand_v4sf3): Likewise.
9950 (nand_v2df3): Likewise.
9952 (orc_v16qi3): Likewise.
9953 (orc_v8hi3): Likewise.
9954 (orc_v4si3): Likewise.
9955 (orc_v2di3): Likewise.
9956 (orc_v4sf3): Likewise.
9957 (orc_v2df3): Likewise.
9959 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9960 allow power8 quad mode in 64-bit.
9961 (rs6000_builtin_vectorized_function): Add support to vectorize
9962 ISA 2.07 count leading zeros, population count builtins.
9963 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9964 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9965 (builtin_function_type): Add vgbbd builtin function which takes an
9967 (altivec_expand_vec_perm_const): Add support for new power8 merge
9970 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9971 that does not include TImdoe for use with 32-bit.
9972 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9974 (UNSPEC_VSX_CVDPSPN): Likewise.
9975 (vsx_xscvdpspn): Likewise.
9976 (vsx_xscvspdpn): Likewise.
9977 (vsx_xscvdpspn_scalar): Likewise.
9978 (vsx_xscvspdpn_directmove): Likewise.
9979 (vsx_and<mode>3): Split logical operations into 32-bit and
9980 64-bit. Add support to do logical operations on TImode as well as
9981 VSX vector types. Allow logical operations to be done in either
9982 VSX registers or in general purpose registers in 64-bit mode. Add
9983 splitters if GPRs were used. For AND, add clobber of CCmode to
9984 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9986 (vsx_and<mode>3_32bit): Likewise.
9987 (vsx_and<mode>3_64bit): Likewise.
9988 (vsx_ior<mode>3): Likewise.
9989 (vsx_ior<mode>3_32bit): Likewise.
9990 (vsx_ior<mode>3_64bit): Likewise.
9991 (vsx_xor<mode>3): Likewise.
9992 (vsx_xor<mode>3_32bit): Likewise.
9993 (vsx_xor<mode>3_64bit): Likewise.
9994 (vsx_one_cmpl<mode>2): Likewise.
9995 (vsx_one_cmpl<mode>2_32bit): Likewise.
9996 (vsx_one_cmpl<mode>2_64bit): Likewise.
9997 (vsx_nor<mode>3): Likewise.
9998 (vsx_nor<mode>3_32bit): Likewise.
9999 (vsx_nor<mode>3_64bit): Likewise.
10000 (vsx_andc<mode>3): Likewise.
10001 (vsx_andc<mode>3_32bit): Likewise.
10002 (vsx_andc<mode>3_64bit): Likewise.
10003 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
10004 and xxlorc instructions.
10005 (vsx_eqv<mode>3_64bit): Likewise.
10006 (vsx_nand<mode>3_32bit): Likewise.
10007 (vsx_nand<mode>3_64bit): Likewise.
10008 (vsx_orc<mode>3_32bit): Likewise.
10009 (vsx_orc<mode>3_64bit): Likewise.
10011 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
10013 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
10015 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
10016 (p8_vmrgow): Likewise.
10017 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
10018 GPRs to be split under VSX.
10019 (p8v_clz<mode>2): Add power8 count leading zero support.
10020 (p8v_popcount<mode>2): Add power8 population count support.
10021 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
10024 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
10027 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
10029 (vec_nand): Likewise.
10030 (vec_vclz): Likewise.
10031 (vec_vclzb): Likewise.
10032 (vec_vclzd): Likewise.
10033 (vec_vclzh): Likewise.
10034 (vec_vclzw): Likewise.
10035 (vec_vgbbd): Likewise.
10036 (vec_vmrgew): Likewise.
10037 (vec_vmrgow): Likewise.
10038 (vec_vpopcnt): Likewise.
10039 (vec_vpopcntb): Likewise.
10040 (vec_vpopcntd): Likewise.
10041 (vec_vpopcnth): Likewise.
10042 (vec_vpopcntw): Likewise.
10044 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10046 PR rtl-optimization/57468
10047 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
10050 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10052 PR rtl-optimization/57459
10053 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
10054 type when setting live regs.
10056 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
10058 * config/s390/s390.opt (mlra): New option.
10059 * config/s390/s390.c (s390_decompose_address): Check displacement
10060 for all registers for LRA.
10061 (s390_secondary_reload): Don't used secondary reloads for LRA.
10062 (s390_lra_p): New function.
10063 (TARGET_LRA_P): Define.
10064 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
10065 of attribute cpu_facility to zarch for the last alternative.
10066 (*cmpmem_short): Ditto.
10068 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
10070 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
10071 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
10072 (arm_expand_prologue): Likewise.
10074 2013-06-06 Teresa Johnson <tejohnson@google.com>
10077 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
10078 as this is now done by redirect_edge_and_branch_force.
10079 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
10080 barriers, and fix interaction with splitting.
10081 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
10082 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
10083 reflect changes made in the routine.
10084 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
10085 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
10086 since this is called in cfglayout mode, and replace partition fixup
10087 with assert as that is now done by force_nonfallthru_and_redirect.
10088 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
10089 already be marked with region crossing note.
10090 (insert_section_boundary_note): Make non-static, gate on flag
10091 has_bb_partition, rewrite to also check for multiple partitions.
10092 (rest_of_handle_reorder_blocks): Remove call to
10093 insert_section_boundary_note, now done later during free_cfg.
10094 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
10095 * bb-reorder.h (insert_section_boundary_note): Declare.
10096 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
10097 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
10098 invoke insert_section_boundary_note.
10099 (try_redirect_by_replacing_jump): Remove unnecessary
10100 check for region crossing note.
10101 (fixup_partition_crossing): New function.
10102 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
10103 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
10104 in non-cfglayout mode.
10105 (force_nonfallthru_and_redirect): Fixup partition boundaries,
10106 remove old code that tried to do this. Emit barrier correctly
10107 when we are in cfglayout mode.
10108 (last_bb_in_partition): New function.
10109 (rtl_split_edge): Correctly fixup partition boundaries.
10110 (commit_one_edge_insertion): Remove old code that tried to
10111 fixup region crossing edge since this is now handled in
10112 split_block, and set up insertion point correctly since
10113 block may now end in a jump.
10114 (verify_hot_cold_block_grouping): Guard against checking when not in
10115 linearized RTL mode.
10116 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
10118 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
10119 rtl_verify_flow_info, so not called in cfglayout mode.
10120 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
10121 (fixup_reorder_chain): Remove old code that attempted to fixup region
10122 crossing note as this is now handled in force_nonfallthru_and_redirect.
10123 (duplicate_insn_chain): Don't duplicate switch section notes.
10124 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
10126 * basic-block.h (emit_barrier_after_bb): Declare.
10128 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10130 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
10131 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
10132 arm_usatsihi): Adjust alternatives for arm_restrict_it.
10134 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10136 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
10138 * config/arm/ldmstm.md: Regenerate.
10140 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10142 * config/arm/sync.md (atomic_loaddi_1):
10143 Disable predication for arm_restrict_it.
10144 (arm_load_exclusive<mode>): Likewise.
10145 (arm_load_exclusivesi): Likewise.
10146 (arm_load_exclusivedi): Likewise.
10147 (arm_load_acquire_exclusive<mode>): Likewise.
10148 (arm_load_acquire_exclusivesi): Likewise.
10149 (arm_load_acquire_exclusivedi): Likewise.
10150 (arm_store_exclusive<mode>): Likewise.
10151 (arm_store_exclusive<mode>): Likewise.
10152 (arm_store_release_exclusivedi): Likewise.
10153 (arm_store_release_exclusive<mode>): Likewise.
10155 2013-06-06 Richard Biener <rguenther@suse.de>
10157 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
10159 (lto_tag_is_tree_code_p): Adjust.
10160 (lto_tag_is_gimple_code_p): Likewise.
10161 (lto_gimple_code_to_tag): Likewise.
10162 (lto_tag_to_gimple_code): Likewise.
10163 (lto_tree_code_to_tag): Likewise.
10164 (lto_tag_to_tree_code): Likewise.
10165 * data-streamer.h (streamer_write_hwi_in_range): Use
10166 uhwi streaming to stream the normalized range.
10167 (streamer_read_hwi_in_range): Likewise.
10169 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10171 * config/arm/arm.md (enabled_for_depr_it): New attribute.
10172 (predicable_short_it): Likewise.
10173 (predicated): Likewise.
10174 (enabled): Handle above.
10175 (define_cond_exec): Set predicated attribute to yes.
10177 2013-06-05 Mike Stump <mikestump@comcast.net>
10179 * gdbinit.in (__FUNCTION__): Add.
10181 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
10183 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
10184 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
10186 2013-06-05 Jan Hubicka <jh@suse.cz>
10188 * varasm.c (mark_decl_referenced): Revert the removal until targets
10191 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
10193 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
10194 instead of mark_decl_referenced.
10196 2013-06-05 Jan Hubicka <jh@suse.cz>
10198 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
10199 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
10200 and symtab_used_from_object_file_p.
10201 (cgraph_make_node_local_1): Clear forced_by_abi.
10202 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
10203 * cgraph.h (symtab_node_base): Add forced_by_abi.
10204 (decide_is_variable_needed): Remove.
10205 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
10206 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
10207 (decide_is_symbol_needed): ... this one; handle symbols in general;
10208 always analyze virtuals; honnor forced_by_abi.
10209 (cgraph_finalize_function): Update.
10210 (varpool_finalize_decl): Update.
10211 (symbol_defined_and_needed): Remove.
10212 (analyze_functions): Update.
10213 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
10214 output_refs, input_overwrite_node): Handle forced_by_abi.
10215 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
10216 (address_taken_from_non_vtable_p): ... this one.
10217 (comdat_can_be_unshared_p_1): New function.
10218 (cgraph_comdat_can_be_unshared_p): Rename to ...
10219 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
10220 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
10221 (function_and_variable_visibility): Clear forced_by_abi as needed.
10222 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
10223 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
10224 * symtab.c (dump_symtab_base): Dump forced_by_abi.
10225 * varpool.c (decide_is_variable_needed): Remove.
10227 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10229 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
10230 (arm_option_override): Override arm_restrict_it where appropriate.
10231 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
10232 * config/arm/arm.opt (mrestrict-it): New command-line option.
10233 * doc/invoke.texi: Document -mrestrict-it.
10235 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10237 * tsan.c (tsan_atomic_table): Make const.
10239 2013-06-05 Richard Biener <rguenther@suse.de>
10241 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
10242 index associated with the tree we are supposed to replace.
10243 * tree-streamer-out.c (pack_ts_base_value_fields): Output
10244 TREE_ASM_WRITTEN as zero for everything but SSA names.
10246 2013-06-05 David Malcolm <dmalcolm@redhat.com>
10248 * tree-ssa-structalias.c (call_stmt_vars): Make static.
10250 2013-06-04 Jan Hubicka <jh@suse.cz>
10252 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
10253 (input_node, input_varpool_node): Handle correctly external same
10255 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
10256 nodes at ltrans stage.
10258 2013-06-04 Jan Hubicka <jh@suse.cz>
10260 * ipa-inline.c (update_caller_keys): Fix availability test.
10261 (update_callee_keys): Likewise.
10262 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
10263 to follow ELF standard.
10265 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
10267 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
10268 (mips64r5900el-*-elf*): New configurations.
10269 * config/mips/mips-cpus.def (r5900): New processor.
10270 * config/mips/mips-tables.opt: Regenerate.
10271 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
10272 (mips_issue_rate): Handle PROCESSOR_R5900.
10273 (mips_reorg_process_insns): Force reorder mode for the R5900.
10274 * config/mips/mips.h (TARGET_MIPS5900): Define.
10275 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
10277 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
10279 * config/mips/mips.md (processor): Add r5900.
10280 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
10282 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10284 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
10285 into function to generate MOVI instruction.
10286 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
10287 (aarch64_preferred_simd_mode): Turn into wrapper.
10288 (aarch64_output_scalar_simd_mov_immediate): New function.
10289 * config/aarch64/aarch64-protos.h: Add prototype for above.
10291 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10293 * config/aarch64/aarch64.c (simd_immediate_info): Remove
10294 element_char member.
10295 (sizetochar): Return signed char.
10296 (aarch64_simd_valid_immediate): Remove elchar and other
10297 unnecessary variables.
10298 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
10299 Calculate element_char as required.
10300 * config/aarch64/aarch64-protos.h: Update and move prototype
10301 for aarch64_output_simd_mov_immediate.
10302 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
10305 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10307 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
10308 information completed by aarch64_simd_valid_immediate.
10309 (aarch64_legitimate_constant_p): Update arguments.
10310 (aarch64_simd_valid_immediate): Work with struct rather than many
10312 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
10313 (aarch64_simd_make_constant): Update arguments.
10314 (aarch64_output_simd_mov_immediate): Work with struct rather than
10315 many pointers. Output immediate directly rather than as operand.
10316 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10318 * config/aarch64/constraints.md (Dn): Update arguments.
10320 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10322 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
10324 (aarch64_simd_immediate_valid_for_move): Remove.
10325 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
10326 (aarch64_simd_make_constant): Update call.
10327 (aarch64_output_simd_mov_immediate): Update call.
10328 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10330 * config/aarch64/constraints.md (Dn): Update call.
10332 2013-06-04 Ian Bolton <ian.bolton@arm.com>
10334 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
10335 return type to bool for prototype.
10336 (aarch64_legitimate_constant_p): Check for true instead of not -1.
10337 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
10338 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
10340 2013-06-04 Catherine Moore <clm@codesourcery.com>
10342 * config/mips/mips.opt (meva): New.
10343 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
10344 (ASM_SPEC): Handle -meva.
10345 * doc/invoke.texi (meva): Document.
10347 2013-06-04 Alan Modra <amodra@gmail.com>
10349 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
10352 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10354 * rtl.def: Add extra fourth optional field to define_cond_exec.
10355 * gensupport.c (process_one_cond_exec): Process attributes from
10357 * doc/md.texi: Document fourth field in define_cond_exec.
10359 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
10361 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
10362 out the processing order as in store_bit_field_1.
10364 2013-06-04 Jan Hubicka <jh@suse.cz>
10366 PR middle-end/57500
10367 * cgraphunit.c (cgraph_process_same_body_aliases): Create
10368 non-VAR_DECL node if it does not exist yet.
10370 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
10372 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
10373 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
10374 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
10375 target_cpu_default setting.
10377 2013-06-03 Teresa Johnson <tejohnson@google.com>
10379 * dumpfile.c (opt_info_switch_p): Change -fopt-info
10380 default to -fopt-info=optimized instead of all.
10381 * doc/invoke.texi: Ditto.
10382 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
10383 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
10384 (execute_vect_slp): Emit BB vectorization success under
10385 MSG_OPTIMIZED_LOCATIONS.
10386 * tree-vect-slp.c (vect_slp_transform_bb): Change
10387 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
10388 * tree-vect-loop.c (vect_transform_loop): Ditto.
10390 2013-06-03 Jason Merrill <jason@redhat.com>
10393 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10394 Use TARGET_EXPR for C++.
10396 2013-06-03 Jakub Jelinek <jakub@redhat.com>
10398 PR rtl-optimization/57268
10399 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
10400 if DEBUG_INSN_P (insn).
10403 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10405 PR rtl-optimization/57268
10406 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10407 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10409 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10411 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
10412 (ix86_avoid_lea_for_addr): Likewise.
10413 (exact_dependency_1): Likewise.
10414 (ix86_adjust_cost): Likewise.
10415 (swap_top_of_ready_list): Fix formatting and !reload_completed check
10417 (do_reorder_for_imul): Fix typo, formatting and
10418 !reload_completed check removed.
10419 (ix86_sched_reorder): Fix typo and formatting.
10420 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
10423 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
10425 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
10427 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
10429 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
10430 <STRING_CST>: Likewise.
10431 <VECTOR_CST>: Likewise.
10433 2013-06-01 Janus Weil <janus@gcc.gnu.org>
10434 Mikael Morin <mikael@gcc.gnu.org>
10436 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
10437 * config.in: Regenerated.
10438 * configure: Regenerated.
10440 2013-06-01 Jan Hubicka <jh@suse.cz>
10442 PR middle-end/57366
10443 * cgraphunit.c (compile): When weakref is not supported,
10444 set up transparent aliases before final output pass.
10445 * varasm.c (assemble_alias): Do not try to do it here.
10447 2013-06-01 Jan Hubicka <jh@suse.cz>
10449 PR middle-end/57467
10450 * passes.c (for_per_function): Skip unanalyzed functions.
10452 2013-06-01 Jan Hubicka <jh@suse.cz>
10454 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
10455 (lto_symtab_merge_symbols_1): ... this one.
10456 (lto_symtab_merge_cgraph_nodes): Rename to ...
10457 (lto_symtab_merge_symbols): ... this one; simplify.
10458 * cgraph.c (same_body_aliases_done): Rename to ...
10459 (cpp_implicit_aliases_done): ... this one.
10460 (cgraph_create_function_alias): Update.
10461 (cgraph_same_body_alias): Update.
10462 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
10463 (verify_edge_corresponds_to_fndecl): Simplify.
10464 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
10465 (cgraph_node): Remove same_body_alias.
10466 (varpool_node): Remove alias_of and extra_name_alias.
10467 (same_body_aliases_done): Rename to ..
10468 (cpp_implicit_aliases_done): ... this one.
10469 (symtab_alias_ultimate_target): Add default parameter.
10470 (symtab_resolve_alias): New function.
10471 (fixup_same_cpp_alias_visibility): Declare.
10472 (cgraph_function_node): Add default parameter.
10473 (cgraph_node_asm_name): Likewise.
10474 (cgraph_function_or_thunk_node): Add default parameter; do
10475 not ICE when it is NULL.
10476 (varpool_variable_node): Likewise.
10477 * tree-emutls.c (create_emultls_var): Update.
10478 (ipa_lower_emutls): Update.
10479 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10480 (cgraph_reset_node): Reset alias info.
10481 (cgraph_finalize_function): Update.
10482 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10483 (analyze_function): Simplify.
10484 (cgraph_process_same_body_aliases): Simplify.
10485 (analyze_functions): Fixup same body aliases.
10486 (handle_alias_pairs): Simplify.
10487 (assemble_thunk): Update.
10488 (assemble_thunks_and_aliases): Update.
10489 (output_weakrefs): Rewrite.
10490 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10491 (lto_output_varpool_node): Likewise.
10492 (compute_ltrans_boundary): Remve assert.
10493 (get_alias_symbol): New functoin.
10494 (input_node): Rewrite alias handling.
10495 (input_varpool_node): Likewise.
10496 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10497 * ipa.c (process_references): Handle weakrefs correctly.
10498 (symtab_remove_unreachable_nodes): Likewise.
10499 * trans-mem.c (get_cg_data): Update.
10500 (ipa_tm_create_version_alias): Update.
10501 (ipa_tm_execute): Update.
10502 * symtab.c (dump_symtab_base): Dump aliases.
10503 (verify_symtab_base): Verify aliases.
10504 (symtab_node_availability): New function.
10505 (symtab_alias_ultimate_target): Simplify.
10506 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10507 handle all the fixup cases.
10508 (symtab_resolve_alias): New function.
10509 * passes.c (ipa_write_summaries): Handle weakrefs.
10510 * varpool.c (varpool_analyze_node): Simplify.
10511 (assemble_aliases): Update.
10512 (varpool_create_variable_alias): Simplify.
10513 (varpool_extra_name_alias): Simplify.
10514 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10515 (lto_symtab_merge_symbols): ... this one.
10517 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10520 PR rtl-optimization/57268
10521 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10522 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10524 2013-06-01 Tobias Burnus <burnus@net-b.de>
10526 Partially reverted:
10527 2013-05-31 Tobias Burnus <burnus@net-b.de>
10529 PR middle-end/57073
10530 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10533 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10535 PR rtl-optimization/57268
10536 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10537 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10539 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10541 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10542 unordered comparison operators when -fno-trapping-math is in effect
10544 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10545 and implement unordered comparison operators properly on the e500.
10547 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10549 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10550 for constant scalar integers.
10551 (simplify_relational_operation_1): Likewise.
10553 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10555 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10556 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10559 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10560 Igor Zamyatin <igor.zamyatin@intel.com>
10562 Silvermont (SLM) architecture performance tuning.
10563 * config/i386/i386.h (enum ix86_tune_indices): Add
10564 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10565 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10567 * config/i386/i386.c (initial_ix86_tune_features)
10568 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10569 (ix86_lea_outperforms): Handle Silvermont tuning.
10570 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10572 (ix86_use_lea_for_mov): Likewise.
10573 (ix86_avoid_lea_for_addr): Likewise.
10574 (ix86_lea_for_add_ok): Likewise.
10575 (exact_dependency_1): New function.
10576 (exact_store_load_dependency): Likewise.
10577 (ix86_adjust_cost): Handle Silvermont tuning.
10578 (do_reoder_for_imul): Likewise.
10579 (swap_top_of_ready_list): New function.
10580 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10582 * config/i386/i386.md (peepholes that split memory operand in fp
10585 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10587 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10588 Remove un-necessary braces.
10590 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10592 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10593 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10595 2013-05-31 Tobias Burnus <burnus@net-b.de>
10597 PR middle-end/57073
10598 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
10600 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10603 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10604 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10605 * config/arm/neon.md (iordi3_neon): Remove.
10606 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10607 * config/arm/predicates.md (imm_for_neon_logic_operand):
10608 Move to earlier in the file.
10609 (neon_logic_op2): Likewise.
10610 (arm_iordi_operand_neon): New predicate.
10612 2013-05-31 Richard Biener <rguenther@suse.de>
10614 PR tree-optimization/57478
10615 PR tree-optimization/57453
10616 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10619 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10621 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10622 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10624 2013-05-30 Tobias Burnus <burnus@net-b.de>
10625 Thomas Koenig <tkoenig@gcc.gnu.org>
10627 PR middle-end/57073
10628 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10629 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10631 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10633 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
10635 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10637 * target.def (register_usage_leveling_p): New hook.
10638 * targhooks.c (default_register_usage_leveling_p): New.
10639 * targhooks.h (default_register_usage_leveling_p): New prototype.
10640 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10641 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10642 * doc/tm.texi: Update.
10643 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10645 2013-05-30 Ian Bolton <ian.bolton@arm.com>
10647 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10648 (*insv_reg<mode>): New define_insn.
10650 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
10652 PR rtl-optimization/57439
10653 * postreload.c (move2add_valid_value_p): Check that we have
10654 a zero subreg_regno_offset when accessing the register in
10655 the requested mode.
10657 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10658 Igor Zamyatin <igor.zamyatin@intel.com>
10660 Silvermont (SLM) architecture pipeline model, tuning and
10662 * config.gcc: Add slm config options and target.
10664 * config/i386/slm.md: New.
10666 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10668 * config/i386/i386-c.c (ix86_target_macros_internal): New case
10670 (ix86_target_macros_internal): Likewise.
10672 * config/i386/i386.c (slm_cost): New cost.
10673 (m_SLM): New macro flag.
10674 (initial_ix86_tune_features): Set m_SLM.
10675 (x86_accumulate_outgoing_args): Likewise.
10676 (x86_arch_always_fancy_math_387): Likewise.
10677 (processor_target_table): Add slm cost.
10678 (cpu_names): Add slm cpu name.
10679 (x86_option_override_internal): Set SLM ISA.
10680 (ix86_issue_rate): New case PROCESSOR_SLM.
10681 (ia32_multipass_dfa_lookahead): Likewise.
10682 (fold_builtin_cpu): Add slm.
10684 * config/i386/i386.h (TARGET_SLM): New target macro.
10685 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10686 (processor_type): Add PROCESSOR_SLM.
10688 * config/i386/i386.md (cpu): Add new value "slm".
10689 (slm.md): Include slm.md.
10691 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10692 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10694 * config/arm/arm-protos.h: Add and update function protos.
10695 * config/arm/arm.c (use_simple_return_p): New added.
10696 (thumb2_expand_return): Check simple_return flag.
10697 * config/arm/arm.md: Add simple_return and conditional simple_return.
10698 * config/arm/iterators.md: Add iterator for return and simple_return.
10700 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10702 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10703 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10704 (arm_emit_vfp_multi_reg_pop): Likewise.
10705 (thumb2_emit_ldrd_pop): Likewise.
10706 (arm_expand_epilogue): Add misc REG_CFA notes.
10707 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10709 2013-05-29 Lawrence Crowl <crowl@google.com>
10711 * config/arm/t-arm: Update for below.
10713 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10714 Change type to hash_table. Update dependent calls and types.
10716 * config/i386/t-cygming: Update for below.
10718 * config/i386/t-interix: Update for below.
10720 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10721 Change type to hash_table. Update dependent calls and types.
10722 (i386_find_on_wrapper_list::wrappers): Likewise.
10724 * config/ia64/t-ia64: Update for below.
10726 * config/ia64/ia64.c (bundle_state_table):
10727 Change type to hash_table. Update dependent calls and types.
10729 * config/mips/mips.c (mips_reorg_process_insns::htab):
10730 Change type to hash_table. Update dependent calls and types.
10732 * config/sol2.c (solaris_comdat_htab):
10733 Change type to hash_table. Update dependent calls and types.
10735 * config/t-sol2: Update for above.
10737 2013-05-29 Teresa Johnson <tejohnson@google.com>
10739 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10740 functions are not yet marked as defined.
10742 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
10743 Pat Haugen <pthaugen@us.ibm.com>
10744 Peter Bergner <bergner@vnet.ibm.com>
10746 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10750 (vrotl<mode>3): Likewise.
10751 (vashl<mode>3): Likewise.
10752 (vlshr<mode>3): Likewise.
10753 (vashr<mode>3): Likewise.
10755 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10756 support for power8 V2DI builtins.
10758 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10759 power8 V2DI builtins.
10760 (vupkhsw): Likewise.
10761 (vupklsw): Likewise.
10762 (vaddudm): Likewise.
10763 (vminsd): Likewise.
10764 (vmaxsd): Likewise.
10765 (vminud): Likewise.
10766 (vmaxud): Likewise.
10767 (vpkudum): Likewise.
10768 (vpksdss): Likewise.
10769 (vpkudus): Likewise.
10770 (vpksdus): Likewise.
10775 (vsubudm): Likewise.
10776 (vcmpequd): Likewise.
10777 (vcmpgtsd): Likewise.
10778 (vcmpgtud): Likewise.
10779 (vcmpequd_p): Likewise.
10780 (vcmpgtsd_p): Likewise.
10781 (vcmpgtud_p): Likewise.
10782 (vupkhsw): Likewise.
10783 (vupklsw): Likewise.
10784 (vaddudm): Likewise.
10785 (vmaxsd): Likewise.
10786 (vmaxud): Likewise.
10787 (vminsd): Likewise.
10788 (vminud): Likewise.
10789 (vpksdss): Likewise.
10790 (vpksdus): Likewise.
10791 (vpkudum): Likewise.
10792 (vpkudus): Likewise.
10797 (vsubudm): Likewise.
10799 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10800 support for power8 V2DI instructions.
10802 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10803 power8 V2DI instructions. Combine pack and unpack insns to use an
10804 iterator for each mode. Check whether a particular mode supports
10805 Altivec instructions instead of just checking TARGET_ALTIVEC.
10806 (UNSPEC_VPKUWUM): Likewise.
10807 (UNSPEC_VPKSHSS): Likewise.
10808 (UNSPEC_VPKSWSS): Likewise.
10809 (UNSPEC_VPKUHUS): Likewise.
10810 (UNSPEC_VPKSHUS): Likewise.
10811 (UNSPEC_VPKUWUS): Likewise.
10812 (UNSPEC_VPKSWUS): Likewise.
10813 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10814 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10815 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10816 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10817 (UNSPEC_VUPKHSB): Likewise.
10818 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10819 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10820 (UNSPEC_VUPKHSH): Likewise.
10821 (UNSPEC_VUPKLSB): Likewise.
10822 (UNSPEC_VUPKLSH): Likewise.
10824 (VI_char): Likewise.
10825 (VI_scalar): Likewise.
10826 (VI_unit): Likewise.
10828 (VP_small): Likewise.
10829 (VP_small_lc): Likewise.
10830 (VU_char): Likewise.
10831 (add<mode>3): Likewise.
10832 (altivec_vaddcuw): Likewise.
10833 (altivec_vaddu<VI_char>s): Likewise.
10834 (altivec_vadds<VI_char>s): Likewise.
10835 (sub<mode>3): Likewise.
10836 (altivec_vsubcuw): Likewise.
10837 (altivec_vsubu<VI_char>s): Likewise.
10838 (altivec_vsubs<VI_char>s): Likewise.
10839 (altivec_vavgs<VI_char>): Likewise.
10840 (altivec_vcmpbfp): Likewise.
10841 (altivec_eq<mode>): Likewise.
10842 (altivec_gt<mode>): Likewise.
10843 (altivec_gtu<mode>): Likewise.
10844 (umax<mode>3): Likewise.
10845 (smax<mode>3): Likewise.
10846 (umin<mode>3): Likewise.
10847 (smin<mode>3): Likewise.
10848 (altivec_vpkuhum): Likewise.
10849 (altivec_vpkuwum): Likewise.
10850 (altivec_vpkshss): Likewise.
10851 (altivec_vpkswss): Likewise.
10852 (altivec_vpkuhus): Likewise.
10853 (altivec_vpkshus): Likewise.
10854 (altivec_vpkuwus): Likewise.
10855 (altivec_vpkswus): Likewise.
10856 (altivec_vpks<VI_char>ss): Likewise.
10857 (altivec_vpks<VI_char>us): Likewise.
10858 (altivec_vpku<VI_char>us): Likewise.
10859 (altivec_vpku<VI_char>um): Likewise.
10860 (altivec_vrl<VI_char>): Likewise.
10861 (altivec_vsl<VI_char>): Likewise.
10862 (altivec_vsr<VI_char>): Likewise.
10863 (altivec_vsra<VI_char>): Likewise.
10864 (altivec_vsldoi_<mode>): Likewise.
10865 (altivec_vupkhsb): Likewise.
10866 (altivec_vupkhs<VU_char>): Likewise.
10867 (altivec_vupkls<VU_char>): Likewise.
10868 (altivec_vupkhsh): Likewise.
10869 (altivec_vupklsb): Likewise.
10870 (altivec_vupklsh): Likewise.
10871 (altivec_vcmpequ<VI_char>_p): Likewise.
10872 (altivec_vcmpgts<VI_char>_p): Likewise.
10873 (altivec_vcmpgtu<VI_char>_p): Likewise.
10874 (abs<mode>2): Likewise.
10875 (vec_unpacks_hi_v16qi): Likewise.
10876 (vec_unpacks_hi_v8hi): Likewise.
10877 (vec_unpacks_lo_v16qi): Likewise.
10878 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10879 (vec_unpacks_lo_v8hi): Likewise.
10880 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10881 (vec_pack_trunc_v8h): Likewise.
10882 (vec_pack_trunc_v4si): Likewise.
10883 (vec_pack_trunc_<mode>): Likewise.
10885 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10887 (vec_vmaxsd): Likewise.
10888 (vec_vmaxud): Likewise.
10889 (vec_vminsd): Likewise.
10890 (vec_vminud): Likewise.
10891 (vec_vpksdss): Likewise.
10892 (vec_vpksdus): Likewise.
10893 (vec_vpkudum): Likewise.
10894 (vec_vpkudus): Likewise.
10895 (vec_vrld): Likewise.
10896 (vec_vsld): Likewise.
10897 (vec_vsrad): Likewise.
10898 (vec_vsrd): Likewise.
10899 (vec_vsubudm): Likewise.
10900 (vec_vupkhsw): Likewise.
10901 (vec_vupklsw): Likewise.
10903 2013-05-29 Jan Hubicka <jh@suse.cz>
10905 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10906 flags; reorder rest of fields in more consistent way.
10907 (varpool_node): Remove analyzed, finalized and alias.
10908 (cgraph_ndoe): Likewise.
10909 (symtab_alias_ultimate_target): New function.
10910 (cgraph_function_node): Move offline.
10911 (cgraph_reset_node): Declare.
10912 (cgraph_comdat_can_be_unshared_p): Remove.
10913 (varpool_remove_initializer): Declare.
10914 (varpool_first_defined_variable, varpool_next_defined_variable
10915 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10916 (cgraph_function_with_gimple_body_p): Update.
10917 (varpool_all_refs_explicit_p): Update.
10918 (symtab_alias_target): New function.
10919 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10920 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
10921 (cgraph_function_or_thunk_node): Simplify using
10922 symtab_alias_ultimate_target.
10923 (varpool_variable_node): Likewise.
10924 * cgraph.c (cgraph_create_function_alias): Update.
10925 (cgraph_add_thunk): Update.
10926 (cgraph_remove_node): Update.
10927 (dump_cgraph_node): Do not dump removed flags.
10928 (cgraph_function_body_availability): Update.
10929 (cgraph_propagate_frequency): Update.
10930 (verify_cgraph_node): Check sanity of local flag.
10931 (cgraph_function_node): Move here from cgraph.h; revamp for
10932 cgraph_function_or_thunk_node.
10933 * lto-symtab.c (lto_varpool_replace_node): Update.
10934 (lto_symtab_resolve_can_prevail_p): Update.
10935 (lto_symtab_merge_cgraph_nodes): Update.
10936 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10937 propagate_constants_accross_call, devirtualization_time_bonus,
10938 ipcp_propagate_stage): Update.
10939 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
10940 * ipa-inline-transform.c (clone_inlined_nodes,
10941 preserve_function_body_p): Update.
10942 * ipa-reference.c (propagate): Update.
10943 (write_node_summary_p): Update.
10944 * toplev.c (wrapup_global_declaration_2): Update.
10945 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10946 (analyze_function) ... this one.
10947 (cgraph_process_new_functions): Update.
10948 (cgraph_reset_node): Export.
10949 (cgraph_finalize_function): Update.
10950 (cgraph_add_new_function): Update.
10951 (process_function_and_variable_attributes): Update.
10952 (varpool_finalize_decl): Update.
10953 (symbol_finalized): Remove.
10954 (symbol_finalized_and_needed): Rename to ...
10955 (symbol_defined_and_needed): ... update.
10956 (cgraph_analyze_functions): Update.
10957 (handle_alias_pairs): Update.
10958 (mark_functions_to_output): Update.
10959 (assemble_thunk): Update.
10960 (output_in_order): Update.
10961 (output_weakrefs): Update.
10962 (finalize_compilation_unit): Update.
10963 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10964 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10965 input_node, input_varpool_node): Update.
10966 * dbxout.c (dbxout_expand_expr): Update.
10967 * cgraphclones.c (cgraph_clone_node): Update.
10968 (cgraph_copy_node_for_versioning): Update.
10969 (cgraph_materialize_clone): Update.
10970 (cgraph_materialize_all_clones): Update.
10971 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10972 propagate_pure_const, propagate_nothrow): Update.
10973 * lto-streamer-out.c (lto_output, write_symbol): Update.
10974 * ipa-utils.c (ipa_reverse_postorder): Update.
10975 * ipa-inline.c (can_inline_edge_p): Update.
10976 (update_caller_keys, ipa_inline): Update.
10977 * dwarf2out.c (reference_to_unused,
10978 premark_types_used_by_global_vars_helper): Update.
10979 * tree-eh.c (tree_could_trap_p): Update.
10980 * ipa-split.c (consider_split, execute_split_functions): Update.
10981 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
10982 has_addr_references_p): Update; move ahead in file for better
10984 (process_references): Simplify.
10985 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10986 bodies are removed.
10987 (cgraph_comdat_can_be_unshared_p): Make static.
10988 (cgraph_externally_visible_p): Update.
10989 (varpool_externally_visible_p): Update.
10990 (function_and_variable_visibility): Update.
10991 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10992 ipa_tm_mark_force_output_node): Update.
10993 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10994 estimate_edge_devirt_benefit, inline_generate_summary,
10995 inline_write_summary): Update.
10996 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10997 * ipa-prop.c (ipa_compute_jump_functions): Update.
10998 (ipa_print_node_params, ipa_prop_read_section,
10999 ipa_update_after_lto_read, read_replacements_section): Update.
11000 * varasm.c (mark_decl_referenced): Update.
11001 (assemble_alias, dump_tm_clone_pairs): Update.
11002 * tree-inline.c (copy_bb): Update.
11003 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
11005 * symtab.c (dump_symtab_base): Print new flags.
11006 (verify_symtab_base): Verify new flags.
11007 (symtab_alias_ultimate_target): New function.
11008 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
11009 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
11011 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
11013 * i386.c (ix86_get_function_versions_dispatcher,
11014 ix86_generate_version_dispatcher_body): Update.
11015 (fold_builtin_cpu): Use varpool_add_new_variable.
11016 * varpool.c (varpool_remove_initializer): Break out from ...
11017 (varpool_remove_node): ... this one.
11018 (dump_varpool_node, varpool_node_for_asm,
11019 cgraph_variable_initializer_availability, varpool_analyze_node,
11020 varpool_assemble_decl, varpool_remove_unreferenced_decls,
11021 varpool_finalize_named_section_flags, varpool_create_variable_alias):
11024 2013-05-29 Jan Hubicka <jh@suse.cz>
11026 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
11028 2013-05-29 Easwaran Raman <eraman@google.com>
11030 PR tree-optimization/57442
11031 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
11032 when control exits the main loop.
11034 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
11036 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
11038 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
11039 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
11040 * rx/t-rx: Add rx100 under multi library matches option for nofpu
11043 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11045 PR tree-optimization/57441
11046 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
11047 Don't limit size of incr_vec to number of candidates.
11049 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
11051 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
11052 and mips16 directories.
11053 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
11054 (MULTILIB_DIRNAMES): Ditto.
11055 (MULTILIB_EXCEPTIONS): Add new exceptions.
11056 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
11057 (MULTILIB_DIRNAMES): Ditto.
11058 (MULTILIB_EXCEPTIONS): Add new exceptions.
11060 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11061 Marcus Shawcroft <marcus.shawcroft@arm.com>
11063 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
11064 SYMBOL_TINY_ABSOLUTE.
11065 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
11066 SYMBOL_TINY_ABSOLUTE.
11067 (aarch64_expand_mov_immediate): Likewise.
11068 (aarch64_classify_symbol): Likewise.
11069 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
11070 Permit SYMBOL_TINY_ABSOLUTE.
11071 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
11073 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11074 Marcus Shawcroft <marcus.shawcroft@arm.com>
11076 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
11077 Refactor if/switch. Replace gcc_assert with if.
11079 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11081 * config/i386/i386.c (initial_ix86_tune_features): Enable
11082 FP Reassociation for AMD bdver1 and bdver2.
11084 2013-05-29 Martin Jambor <mjambor@suse.cz>
11086 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
11087 and IMAGPART_EXPR do not occur within other handled_components.
11089 2013-05-29 Richard Biener <rguenther@suse.de>
11091 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
11092 access on whether the use is in the BB we currently try to
11094 (vect_bb_vectorization_profitable_p): Pass the BB we currently
11095 vectorize to vect_bb_slp_scalar_cost.
11097 2013-05-29 Richard Biener <rguenther@suse.de>
11099 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
11100 computing scalar cost offsetted by stmts that are kept live
11102 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
11103 for computation of scalar cost.
11105 2013-05-28 Steve Ellcey <sellcey@mips.com>
11107 * config/mips/mips-cpus.def (mips32r2): Change processor type.
11109 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
11111 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
11112 array notation built-in reduction functions.
11113 * doc/passes.texi (Passes): Added documentation about changes done
11115 * doc/invoke.texi (C Dialect Options): Added documentation about
11116 the -fcilkplus flag.
11117 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
11118 (BUILTINS_DEF): Depend on cilkplus.def.
11119 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
11120 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
11121 * cilkplus.def: New file.
11123 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
11125 PR rtl-optimization/57439
11126 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
11128 2013-05-28 Easwaran Raman <eraman@google.com>
11130 PR tree-optimization/57337
11131 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
11132 (find_insert_point): Correctly identify the insertion point
11133 when two statements with the same UID is compared.
11135 2013-05-28 Richard Biener <rguenther@suse.de>
11137 PR tree-optimization/56787
11138 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
11139 from the list of data references.
11140 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
11142 (vect_analyze_loop_operations): Likewise.
11143 (vect_transform_loop): Remove clobbers.
11145 2013-05-28 Martin Jambor <mjambor@suse.cz>
11147 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
11148 and REALPART_EXPRs have scalar type.
11150 2013-05-28 Richard Biener <rguenther@suse.de>
11152 PR tree-optimization/57411
11153 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
11155 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
11156 virtual operand propagation.
11158 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11160 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
11161 destination register for bmasksi_vis.
11162 (vector_init_bshuffle): Likewise.
11163 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
11165 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
11167 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
11168 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
11169 mode if the instruction isn't available in the original mode.
11170 * config/sparc/sparc.opt (mfix-ut699): New option.
11171 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
11172 (divdf3): Turn into expander.
11173 (divdf3_nofix): New insn.
11174 (divdf3_fix): Likewise.
11175 (divsf3): Disable if -mfix-ut699.
11176 (sqrtdf2): Turn into expander.
11177 (sqrtdf2_nofix): New insn.
11178 (sqrtdf2_fix): Likewise.
11179 (sqrtsf2): Disable if -mfix-ut699.
11181 2013-05-27 Richard Biener <rguenther@suse.de>
11183 PR middle-end/57412
11184 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
11185 block for the new loop.
11187 2013-05-27 Richard Biener <rguenther@suse.de>
11189 PR tree-optimization/57343
11190 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
11191 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
11192 (number_of_iterations_cond): Do not build the folded tree.
11194 2013-05-27 Richard Biener <rguenther@suse.de>
11197 PR middle-end/57381
11198 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11199 OEP_CONSTANT_ADDRESS_OF retained.
11201 PR tree-optimization/57417
11202 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
11203 for unchanged base.
11204 (set_ssa_val_to): Compare addresses using
11205 get_addr_base_and_unit_offset.
11207 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
11209 PR rtl-optimization/56833
11210 * postreload.c (move2add_record_mode): New function.
11211 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
11212 (move2add_use_add2_insn): Use move2add_record_sym_value.
11213 (move2add_use_add3_insn): Likewise.
11214 (reload_cse_move2add): Use move2add_valid_value_p and
11215 move2add_record_mode. Invalidate call-clobbered and REG_INC
11216 affected regs by setting reg_mode to VOIDmode.
11217 (move2add_note_store): Don't pretend the inside of a SUBREG is
11218 the actual destination. Invalidate single/leading registers by
11219 setting reg_mode to VOIDmode.
11220 Use move2add_record_sym_value, move2add_valid_value_p and
11221 move2add_record_mode.
11223 2013-05-27 Richard Biener <rguenther@suse.de>
11225 PR tree-optimization/57396
11226 * tree-affine.c (double_int_constant_multiple_p): Properly
11227 return false for val == 0 and div != 0.
11229 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11231 * config/mips/mips.h: Use #elif in preprocessor conditions.
11233 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11236 * config/mips/constraints.md (kl): New constraint.
11237 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
11238 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
11239 constraint for operand 0. Split after CSE for MIPS16. Emit a move
11240 from LO for MIPS16.
11241 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
11243 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
11246 * config/mips/mips.c (mips_can_inline_p): New function.
11247 (TARGET_CAN_INLINE_P): Define.
11249 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
11251 * sched-int.h (ds_t, dw_t): Make unsigned int.
11252 Fix documentation that describes how all the ds_t bits are used.
11253 Reserve the last bit for delayed-branch scheduling.
11254 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
11255 (BITS_PER_DEP_WEAK): Fix definition and documentation.
11256 (gen_dep_weak_1): Remove prototype.
11257 * sched-deps.c (get_dep_weak_1): Make static.
11258 * target.def (speculate_insn, needs_block_p, gen_spec_check,
11259 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
11260 * doc/tm.texi: Regenerate.
11261 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
11263 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
11266 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
11268 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
11269 Sandra Loosemore <sandra@codesourcery.com>
11271 * config.gcc (powerpc-*): Allow native for with-cpu.
11273 2013-05-24 Jeff Law <law@redhat.com>
11275 PR tree-optimization/57124
11276 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
11277 conversion feeding a condition if the range has an overflow
11278 if -fstrict-overflow. Add warnings for when we do make the
11281 2013-05-24 Dehao Chen <dehao@google.com>
11283 * tree-cfg.c (locus_discrim_map): Fix the typo.
11284 (locus_discrim_hasher): Likewise.
11285 (locus_discrim_hasher::hash): Likewise.
11286 (locus_discrim_hasher::equal): Likewise.
11288 2013-05-24 Martin Jambor <mjambor@suse.cz>
11290 PR tree-optimization/57294
11291 * cgraph.h (ipa_record_stmt_references): Declare.
11292 * cgraphbuild.c (ipa_record_stmt_references): New function.
11293 (build_cgraph_edges): Use ipa_record_stmt_references.
11294 (rebuild_cgraph_edges): Likewise.
11295 (cgraph_rebuild_references): Likewise.
11296 * ipa-prop.c (ipa_modify_call_arguments): Discard references
11297 associated with the old statement and build references from the
11298 newly built statements.
11299 * ipa-ref.c (ipa_remove_stmt_references): New function.
11300 * ipa-ref.h (ipa_remove_stmt_references): Declare.
11302 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
11304 * lra-constraints.c (emit_spill_move): Use smaller mode for
11306 (check_and_process_move): Consider mem-reg moves for secondary
11308 (curr_insn_transform): Don't lose insns emitted before for
11309 secondary memory moves.
11310 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
11311 reg set up in the current insn.
11313 2013-05-24 Dehao Chen <dehao@google.com>
11315 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
11317 (locus_descrim_hasher::equal): Likewise.
11318 (build_gimple_cfg): New discriminator assignment algorithm.
11319 (make_edges): Likewise.
11320 (next_discriminator_for_locus): Likewise.
11321 (same_line_p): Likewise.
11322 (assign_discriminators): Likewise.
11323 (make_cond_expr_edges): Likewise.
11324 (make_gimple_switch_edges): Likewise.
11325 (make_goto_expr_edges): Likewise.
11326 (make_gimple_asm_edges): Likewise.
11328 2013-05-24 Ian Bolton <ian.bolton@arm.com>
11330 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
11331 X format specifier to only display bottom 16 bits.
11332 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
11333 immediate to match for operand 2, since it will be masked.
11335 2013-05-24 Richard Biener <rguenther@suse.de>
11337 PR tree-optimization/57287
11338 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
11339 all SSA names that occur in abnormal PHIs.
11341 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
11344 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
11345 that index is not negative.
11347 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
11349 PR rtl-optimization/55177
11350 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
11351 (simplify_byte_swapping_operation): New.
11352 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
11353 (simplify_relational_operation_1): Deal with BSWAP.
11355 2013-05-23 Richard Henderson <rth@redhat.com>
11358 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
11359 (ix86_reorg): Call it.
11361 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
11364 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
11365 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
11366 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
11368 2013-05-23 Christian Bruel <christian.bruel@st.com>
11371 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
11373 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11374 Marcus Shawcroft <marcus.shawcroft@arm.com>
11376 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
11377 * config/aarch64/constraints.md (Usa): Remove.
11378 * doc/md.texi (AArch64 Usa): Remove.
11380 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
11381 Marcus Shawcroft <marcus.shawcroft@arm.com>
11383 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
11384 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
11385 * config/aarch64/predicates.md (aarch64_const_address): Remove.
11386 (aarch64_mov_operand): Use aarch64_mov_operand_p.
11388 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
11390 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
11391 instruction (AdvSIMD).
11392 * config/aarch64/aarch64-builtins.c
11393 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
11394 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
11396 2013-05-23 Martin Jambor <mjambor@suse.cz>
11398 PR middle-end/57347
11399 * tree.h (contains_bitfld_component_ref_p): Declare.
11400 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
11401 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
11403 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
11404 not access a bit-field. Assert all final offsets are byte-aligned.
11406 2013-05-23 Richard Biener <rguenther@suse.de>
11408 PR tree-optimization/57380
11409 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
11410 least one invariant or re-used load.
11411 * passes.c (init_optimization_passes): Move pass_phiprop before
11414 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
11416 * config/aarch64/aarch64-simd.md
11417 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
11419 2013-05-23 Richard Biener <rguenther@suse.de>
11421 PR middle-end/57381
11422 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11423 OEP_CONSTANT_ADDRESS_OF retained.
11425 2013-05-23 Jakub Jelinek <jakub@redhat.com>
11427 PR middle-end/57344
11428 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
11429 don't lower unit. Handle unit not being always BITS_PER_WORD.
11431 2013-05-23 Richard Biener <rguenther@suse.de>
11433 PR rtl-optimization/57341
11434 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
11435 instead of true_dependence.
11437 2013-05-22 David Malcolm <dmalcolm@redhat.com>
11439 * bb-reorder.c (branch_threshold): Make const.
11440 (exec_threshold): Ditto.
11442 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11443 Pat Haugen <pthaugen@us.ibm.com>
11444 Peter Bergner <bergner@vnet.ibm.com>
11446 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
11447 documentation for the power8 crypto builtins.
11449 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
11451 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
11452 macros for defining power8 builtin functions.
11453 (BU_P8V_AV_2): Likewise.
11454 (BU_P8V_AV_P): Likewise.
11455 (BU_P8V_VSX_1): Likewise.
11456 (BU_P8V_OVERLOAD_1): Likewise.
11457 (BU_P8V_OVERLOAD_2): Likewise.
11458 (BU_CRYPTO_1): Likewise.
11459 (BU_CRYPTO_2): Likewise.
11460 (BU_CRYPTO_3): Likewise.
11461 (BU_CRYPTO_OVERLOAD_1): Likewise.
11462 (BU_CRYPTO_OVERLOAD_2): Likewise.
11463 (XSCVSPDP): Fix typo, point to the correct instruction.
11464 (VCIPHER): Add power8 crypto builtins.
11465 (VCIPHERLAST): Likewise.
11466 (VNCIPHER): Likewise.
11467 (VNCIPHERLAST): Likewise.
11468 (VPMSUMB): Likewise.
11469 (VPMSUMH): Likewise.
11470 (VPMSUMW): Likewise.
11471 (VPERMXOR_V2DI): Likewise.
11472 (VPERMXOR_V4SI: Likewise.
11473 (VPERMXOR_V8HI: Likewise.
11474 (VPERMXOR_V16QI: Likewise.
11475 (VSHASIGMAW): Likewise.
11476 (VSHASIGMAD): Likewise.
11477 (VPMSUM): Likewise.
11478 (VPERMXOR): Likewise.
11479 (VSHASIGMA): Likewise.
11481 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11482 __CRYPTO__ if the crypto instructions are available.
11483 (altivec_overloaded_builtins): Add support for overloaded power8
11486 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11487 support for power8 crypto builtins.
11488 (builtin_function_type): Likewise.
11489 (altivec_init_builtins): Add support for builtins that take vector
11490 long long (V2DI) arguments.
11492 * config/rs6000/crypto.md: New file, define power8 crypto
11495 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
11496 Pat Haugen <pthaugen@us.ibm.com>
11497 Peter Bergner <bergner@vnet.ibm.com>
11499 * doc/invoke.texi (Option Summary): Add power8 options.
11500 (RS/6000 and PowerPC Options): Likewise.
11502 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11503 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11504 wm, wn, wr documentation.
11506 * config/rs6000/constraints.md (wm): New constraint for VSX
11507 registers if direct move instructions are enabled.
11508 (wn): New constraint for no registers.
11509 (wq): New constraint for quad word even GPR registers.
11510 (wr): New constraint if 64-bit instructions are enabled.
11511 (wv): New constraint if power8 vector instructions are enabled.
11512 (wQ): New constraint for quad word memory locations.
11514 * config/rs6000/predicates.md (const_0_to_15_operand): New
11515 constraint for 0..15 for crypto instructions.
11516 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11517 as GPR and floating point registers.
11518 (int_reg_operand): New predicate to match only GPR registers.
11519 (base_reg_operand): New predicate to match base registers.
11520 (quad_int_reg_operand): New predicate to match even GPR registers
11521 for quad memory operations.
11522 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11523 operations in both GPR and VSX registers.
11524 (quad_memory_operand): New predicate for quad memory operations.
11525 (reg_or_indexed_operand): New predicate for direct move support.
11527 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
11528 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11529 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11530 (POWERPC_MASKS): Add power8 options.
11531 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11534 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11535 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11537 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
11538 (-mpower8-fusion): New power8 options.
11539 (-mpower8-fusion-sign): Likewise.
11540 (-mpower8-vector): Likewise.
11541 (-mcrypto): Likewise.
11542 (-mdirect-move): Likewise.
11543 (-mquad-memory): Likewise.
11545 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
11546 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11548 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
11549 (rs6000_debug_vector_unit): Add p8_vector.
11550 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11551 definitions. Also print fusion state.
11552 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11553 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11554 (rs6000_option_override_internal): Add support for power8.
11555 (rs6000_common_init_builtins): Add debugging for skipped builtins
11556 if -mdebug=builtin.
11557 (rs6000_adjust_cost): Add power8 support.
11558 (rs6000_issue_rate): Likewise.
11559 (insn_must_be_first_in_group): Likewise.
11560 (insn_must_be_last_in_group): Likewise.
11561 (force_new_group): Likewise.
11562 (rs6000_register_move_cost): Likewise.
11563 (rs6000_opt_masks): Likewise.
11565 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11566 power8 capable assembler, default to power7 options.
11567 (TARGET_DIRECT_MOVE): Likewise.
11568 (TARGET_CRYPTO): Likewise.
11569 (TARGET_P8_VECTOR): Likewise.
11570 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11571 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11572 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11573 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11574 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11575 (TARGET_XSCVDPSPN): Likewise.
11576 (TARGET_XSCVSPDPN): Likewsie.
11577 (TARGET_SYNC_HI_QI): Likewise.
11578 (TARGET_SYNC_TI): Likewise.
11579 (MASK_CRYPTO): Likewise.
11580 (MASK_DIRECT_MOVE): Likewise.
11581 (MASK_P8_FUSION): Likewise.
11582 (MASK_P8_VECTOR): Likewise.
11583 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11584 temporary used by some of the direct move instructions to get two FP
11585 temporary registers does not force creation of a stack frame.
11586 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11587 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11588 that any VSX registers are tieable, even if they are also an
11589 Altivec vector mode.
11590 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11591 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11592 (RS6000_BTM_CRYPTO): Likewise.
11593 (RS6000_BTM_COMMON): Likewise.
11595 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11596 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11597 (enum rs6000_vector): Add power8 vector support.
11599 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11603 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11604 (any_sibcall_could_use_r3): this and handle indirect calls.
11605 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11607 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11609 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11611 2013-05-22 Richard Biener <rguenther@suse.de>
11613 PR middle-end/57349
11614 * profile.c (branch_prob): Do not split blocks that are
11615 abnormally receiving from ECF_RETURNS_TWICE functions.
11617 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11619 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11620 address mode. Move pointer mode initialization to the same place.
11622 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11624 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11625 while it has any effect.
11627 2013-05-21 Easwaran Raman <eraman@google.com>
11629 PR tree-optimization/57322
11630 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11631 UID of the statement added to the BB to be 1.
11633 2013-05-21 Jakub Jelinek <jakub@redhat.com>
11635 PR tree-optimization/57331
11636 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11637 of conversion from pointer type to integral type with integer.
11639 2013-05-21 Martin Jambor <mjambor@suse.cz>
11642 * ipa-prop.c (ipa_read_node_info): Process param_used and
11643 controlled_uses in the same order as when writing.
11645 2013-05-21 Magnus Granberg <baldrick@free.fr>
11648 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
11650 2013-05-21 Richard Biener <rguenther@suse.de>
11652 PR tree-optimization/57318
11653 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11654 estimate stmts with side-effects as likely eliminated.
11656 2013-05-21 Richard Biener <rguenther@suse.de>
11658 PR tree-optimization/57330
11659 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11660 preserve the call stmts fntype.
11662 2013-05-21 Richard Biener <rguenther@suse.de>
11664 PR tree-optimization/57303
11665 * tree-ssa-sink.c (statement_sink_location): Improve killing
11666 stmt detection and properly handle self-assignments.
11668 2013-05-21 Christian Bruel <christian.bruel@st.com>
11670 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11671 spanning registers. LEAF_REG_REMAP is supported only for contiguous
11672 registers. Set register size out of the PARALLEL loop.
11674 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11677 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11678 (*fmasf4, *fmasf4_media): New insns.
11680 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11682 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11683 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11684 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11685 (mips_idiv_insns): Update the comments to say that the returned
11686 instruction counts are in units of BASE_INSN_LENGTH.
11687 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11688 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11689 using 2 rather than 4 as the length of indirect MIPS16 and
11690 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11691 length of a NOP. Don't divide MIPS16 lengths by 2.
11692 (mips16_split_long_branches): Assume a branch is long if the
11693 length is greater than 4 rather than 8.
11694 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11695 rather than multiplying them by 2. Multiply instruction counts
11696 by BASE_INSN_LENGTH rather than 4.
11697 (*jump_mips16, tls_get_tp_mips16_<mode>)
11698 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11700 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11702 * config/mips/mips.md (extended_mips16): Remove branch case.
11703 (length): Remove duplicated extended_mips16 test.
11705 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11707 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11709 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11711 * recog.h (Recog_data): Rename to...
11712 (recog_data_d): ...this.
11713 (recog_data): Update accordingly.
11714 * recog.c (recog_data): Likewise.
11715 * reload.c (save_recog_data): Likewise.
11716 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11717 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11719 2013-05-17 Julian Brown <julian@codesourcery.com>
11721 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11722 found in a REG_EQUAL note, invalidate it.
11724 2013-05-17 Easwaran Raman <eraman@google.com>
11726 * tree-ssa-reassoc.c (find_insert_point): New function.
11727 (insert_stmt_after): Likewise.
11728 (get_def_stmt): Likewise.
11729 (ensure_ops_are_available): Likewise.
11730 (not_dominated_by): Likewise.
11731 (rewrite_expr_tree): Do not move statements beyond what is
11732 necessary. Remove call to swap_ops_for_binary_stmt...
11733 (reassociate_bb): ... and move it here.
11734 (build_and_add_sum): Assign UIDs for new statements.
11735 (linearize_expr): Likewise.
11736 (do_reassoc): Renumber gimple statement UIDs.
11738 2013-05-17 Jan Hubicka <jh@suse.cz>
11740 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11742 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
11743 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11745 (output_weakrefs): Update.
11747 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11748 Martin Jambor <mjambor@suse.cz>
11750 PR middle-end/57276
11751 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11752 value that corresponds to the given aggval is found in values vector.
11754 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
11756 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11757 sse, sse2, sse3, ssse3 and sse4a flags to options.
11759 2013-05-17 David Malcolm <dmalcolm@redhat.com>
11761 * gengtype-state.c: (s_expr_writer): New class, to handle
11762 prettifying of output layout of s-expressions.
11763 (state_writer): New class, to write out gtype.state.
11764 (state_written_type_count): Move this variable into member data of
11766 (s_expr_writer::s_expr_writer): New code: constructor for new class
11767 (state_writer::state_writer(): ditto
11768 (s_expr_writer::write_new_line): New function
11769 (s_expr_writer::write_any_indent): ditto
11770 (s_expr_writer::begin_s_expr): ditto
11771 (s_expr_writer::end_s_expr): ditto
11772 (write_state_fileloc): convert to method of state_writer...
11773 (state_writer:: write_state_fileloc): ...and use methods of
11774 s_expr_writer to write indentation into the gtype.state output file
11775 to visually represent the hierarchical structure of the list
11777 (write_state_fields): ditto, renaming to...
11778 (state_writer::write_state_fields)
11779 (write_state_a_string): ditto, renaming to...
11780 (state_writer::write_state_a_string)
11781 (write_state_string_option): ditto, renaming to...
11782 (state_writer::write_state_string_option)
11783 (write_state_type_option): ditto, renaming to...
11784 (state_writer::write_state_type_option)
11785 (write_state_nested_option): ditto, renaming to...
11786 (state_writer::write_state_nested_option)
11787 (write_state_option): ditto, renaming to...
11788 (state_writer::write_state_option)
11789 (write_state_options): ditto, renaming to...
11790 (state_writer::write_state_options)
11791 (write_state_lang_bitmap): ditto, renaming to...
11792 (state_writer::write_state_lang_bitmap)
11793 (write_state_version): ditto, renaming to...
11794 (state_writer::write_state_version)
11795 (write_state_scalar_type): ditto, renaming to...
11796 (state_writer::write_state_scalar_type)
11797 (write_state_string_type): ditto, renaming to...
11798 (state_writer::write_state_string_type)
11799 (write_state_undefined_type): ditto, renaming to...
11800 (state_writer::write_state_undefined_type)
11801 (write_state_struct_union_type): ditto, renaming to...
11802 (state_writer::write_state_struct_union_type)
11803 (write_state_struct_type): ditto, renaming to...
11804 (state_writer::write_state_struct_type)
11805 (write_state_user_struct_type): ditto, renaming to...
11806 (state_writer::write_state_user_struct_type)
11807 (write_state_lang_struct_type): ditto, renaming to...
11808 (state_writer::write_state_lang_struct_type)
11809 (write_state_param_struct_type): ditto, renaming to...
11810 (state_writer::write_state_param_struct_type)
11811 (write_state_pointer_type): ditto, renaming to...
11812 (state_writer::write_state_pointer_type)
11813 (write_state_array_type): ditto, renaming to...
11814 (state_writer::write_state_array_type)
11815 (write_state_gc_used): ditto, renaming to...
11816 (state_writer::write_state_gc_used)
11817 (write_state_common_type_content): ditto, renaming to...
11818 (state_writer::write_state_common_type_content)
11819 (write_state_type): ditto, renaming to...
11820 (state_writer::write_state_type)
11821 (write_state_pair_list): ditto, renaming to...
11822 (state_writer::write_state_pair_list)
11823 (write_state_pair): ditto, renaming to...
11824 (state_writer::write_state_pair)
11825 (write_state_typedefs): ditto, renaming to...
11826 (state_writer::write_state_typedefs)
11827 (write_state_structures): ditto, renaming to...
11828 (state_writer::write_state_structures)
11829 (write_state_param_structs): ditto, renaming to...
11830 (state_writer::write_state_param_structs)
11831 (write_state_variables): ditto, renaming to...
11832 (state_writer::write_state_variables)
11833 (write_state_srcdir): ditto, renaming to...
11834 (state_writer::write_state_srcdir)
11835 (write_state_files_list): ditto, renaming to...
11836 (state_writer::write_state_files_list)
11837 (write_state_languages): ditto, renaming to...
11838 (state_writer::write_state_languages)
11839 (write_state): create a state_writer instance and use it when
11840 writing out the state file
11842 2013-05-17 Mike Stump <mikestump@comcast.net>
11844 PR rtl-optimization/57304
11845 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11846 accessing DF_REF_REAL_LOC.
11848 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11850 PR rtl-optimization/57281
11851 PR rtl-optimization/57300
11852 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11853 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11854 what the other splitter did if the registers are dead.
11856 2013-05-17 Richard Biener <rguenther@suse.de>
11858 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11861 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11863 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11866 2013-05-17 Marek Polacek <polacek@redhat.com>
11868 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11869 length when doing non-zero store of storing '\0' to '\0'.
11871 2013-05-17 Jakub Jelinek <jakub@redhat.com>
11873 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11874 vect_external_def oprnd1 with loop_vinfo, try to emit
11875 optional cast, negation and and stmts on the loop preheader
11876 edge instead of into the pattern def seq.
11878 PR tree-optimization/57051
11879 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11880 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11882 2013-05-16 Nick Clifton <nickc@redhat.com>
11884 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11885 (rl78_is_naked_func): New function.
11886 (rl78_expand_prologue): Skip prologue generation for naked functions.
11887 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
11888 * doc/extend.texi (naked): Add RL78 to the list of processors
11889 that supports this attribute.
11891 2013-05-16 Jeff Law <law@redhat.com>
11893 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11895 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11897 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11898 cache parameters using detect_caches_amd also for CYRIX,
11899 NSC and TM2 signatures.
11901 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
11902 Dzianis Kahanovich <mahatma@eu.by>
11906 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11907 VIA/Centaur processors and determine their cache parameters
11908 using detect_caches_amd.
11910 2013-05-16 Teresa Johnson <tejohnson@google.com>
11912 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11913 (rtl_verify_edges): New function.
11914 (rtl_verify_bb_insns): Ditto.
11915 (rtl_verify_bb_pointers): Ditto.
11916 (rtl_verify_bb_insn_chain): Ditto.
11917 (rtl_verify_fallthru): Ditto.
11918 (rtl_verify_bb_layout): Ditto.
11919 (rtl_verify_flow_info_1): Outline checks into new functions.
11920 (rtl_verify_flow_info): Ditto.
11922 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
11924 * cfghooks.c (copy_bbs): Add update_dominance argument.
11925 * cfghooks.h (copy_bbs): Update prototype.
11926 * tree-cfg.c (gimple_duplicate_sese_region):
11927 Add update_dominance argument.
11928 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11929 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11930 gimple_duplicate_sese_region call.
11931 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11932 Update copy_bbs call.
11933 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11934 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11936 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11938 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11939 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11940 vect_recog_rotate_pattern.
11941 (vect_recog_rotate_pattern): New function.
11943 2013-05-16 Jason Merrill <jason@redhat.com>
11945 * Makefile.in (LLINKER): New variable.
11946 (mostlyclean): Remove link mutex.
11947 * configure.ac: Handle --enable-link-mutex.
11948 * lock-and-run.sh: New script.
11950 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11953 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11956 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11958 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11960 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11962 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11963 * config/arm/arm.c (next_consecutive_mem): New function.
11964 (gen_movmem_ldrd_strd): Likewise.
11965 * config/arm/arm.md (movmemqi): Update condition and code.
11966 (unaligned_loaddi, unaligned_storedi): New patterns.
11968 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11970 * config.gcc: Obsolete *-*-solaris2.9*.
11971 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11973 2013-05-16 Richard Biener <rguenther@suse.de>
11975 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11976 earlier, after GRAPHITE transforms and IV canonicalization.
11978 2013-05-16 Jakub Jelinek <jakub@redhat.com>
11980 * omp-low.c (extract_omp_for_data): For collapsed loops,
11981 if at least one of the loops is known at compile time to
11982 iterate zero times, set count to 0.
11983 (expand_omp_regimplify_p): New function.
11984 (expand_omp_for_generic): For collapsed loops, if at least
11985 one of the loops isn't known to iterate at least once,
11986 add runtime check with setting count to 0.
11987 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11988 For unsigned types if it isn't known at compile time that
11989 the loop will iterate at least once, add runtime check to bypass
11990 the whole loop if initial condition isn't true.
11992 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11994 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11996 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
11998 PR middle-end/57286
11999 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
12000 transformations to avoid an infinite loop.
12002 2013-05-16 Marek Polacek <polacek@redhat.com>
12004 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
12006 2013-05-15 Leif Ekblad <leif@rdos.net>
12008 * config/i386/i386.c (ix86_decompose_address): Use
12009 DEFAULT_TLS_SEG_REG to access TLS segment register.
12010 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
12011 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
12012 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
12014 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
12017 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
12018 sibling calls to functions that would normally be lazily bound,
12019 unless $gp is call-clobbered.
12021 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
12023 * config/i386/i386.c (ix86_option_override_internal): Update
12024 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
12025 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
12026 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
12027 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
12029 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
12031 2013-05-15 Andreas Schwab <schwab@suse.de>
12033 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
12034 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
12037 2013-05-15 Teresa Johnson <tejohnson@google.com>
12039 * loop-unroll.c (report_unroll_peel): Check decision before
12040 emitting unroll/peel message.
12042 2013-05-15 Teresa Johnson <tejohnson@google.com>
12044 * function.h (has_bb_partition): New rtl_data flag.
12045 (bb_reorder_complete): Ditto.
12046 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
12047 instead of flag_reorder_blocks_and_partition.
12048 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
12049 with some enhancements.
12050 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
12051 * bb-reorder.c (connect_traces): Check for has_bb_partition
12052 instead of flag_reorder_blocks_and_partition.
12053 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
12054 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
12055 verify_hot_cold_block_grouping.
12056 (partition_hot_cold_basic_blocks): Set has_bb_partition.
12058 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12061 * config/arm/predicates.md (call_insn_operand): New predicate.
12062 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
12063 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
12064 if insn is not a tail call.
12065 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
12067 * config/arm/arm.h (enum reg_class): New caller save register class.
12068 (REG_CLASS_NAMES): Likewise.
12069 (REG_CLASS_CONTENTS): Likewise.
12070 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
12073 2013-05-15 Richard Biener <rguenther@suse.de>
12075 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
12076 of MSG_OPTIMIZED_LOCATIONS.
12077 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
12078 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
12080 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
12081 of MSG_OPTIMIZED_LOCATIONS.
12082 (execute_vect_slp): Likewise.
12083 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
12084 (vect_create_cond_for_alias_checks): Likewise.
12085 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
12086 (vect_recog_widen_mult_pattern): Likewise.
12087 (vect_recog_widen_sum_pattern): Likewise.
12088 (vect_recog_over_widening_pattern): Likewise.
12089 (vect_recog_widen_shift_pattern): Likewise.
12090 (vect_recog_vector_vector_shift_pattern): Likewise.
12091 (vect_recog_divmod_pattern): Likewise.
12092 (vect_recog_mixed_size_cond_pattern): Likewise.
12093 (vect_recog_bool_pattern): Likewise.
12094 (vect_pattern_recog_1): Likewise.
12096 2013-05-15 Martin Jambor <mjambor@suse.cz>
12098 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
12099 non-functions to builtin_unreachable.
12100 * ipa-inline-transform.c (inline_call): Do not assert estimates were
12101 correct when new direct edges were discovered.
12103 2013-05-15 Martin Jambor <mjambor@suse.cz>
12105 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
12106 header, print symbol order instead of node uid, print more information
12107 about indirect edge targets.
12108 (ipa_make_edge_direct_to_target): Print symbol order instead of node
12110 (ipa_make_edge_direct_to_target): Likewise.
12111 (remove_described_reference): Likewise.
12112 (propagate_controlled_uses): Likewise.
12113 (ipa_print_node_params): Also print symbol order.
12114 (ipcp_transform_function): Print symbol order instead of node uids.
12115 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
12116 (cgraph_get_create_real_symbol_node): Likewise.
12117 * ipa-cp.c (print_lattice): Likewise.
12118 (print_all_lattices): Likewise.
12119 (determine_versionability): Likewise.
12120 (initialize_node_lattices): Likewise.
12121 (estimate_local_effects): Likewise.
12122 (update_profiling_info): Likewise.
12123 (create_specialized_node): Likewise.
12124 (perhaps_add_new_callers): Likewise.
12125 (decide_about_value): Likewise.
12126 (decide_whether_version_node): Likewise.
12127 (identify_dead_nodes): Likewise.
12128 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
12129 (dump_inline_summary): Likewise.
12130 (estimate_node_size_and_time): Likewise.
12131 (inline_analyze_function): Likewise.
12132 * ipa-inline.c (report_inline_failed_reason): Likewise.
12133 (want_early_inline_function_p): Likewise.
12134 (edge_badness): Likewise.
12135 (update_edge_key): Likewise.
12136 (inline_small_functions): Likewise. Add dumping of order to two other
12138 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
12139 instead of node uids.
12140 (propagate_pure_const): Likewise.
12141 (propagate_pure_const): Likewise.
12142 * ipa-utils.c (dump_cgraph_node_set): Likewise.
12143 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
12144 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
12146 * tree-pretty-print.c (dump_function_header): Likewise.
12147 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
12148 Print symbol order instead of node uids.
12150 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12152 * config/s390/s390.c (s390_register_move_cost): Don't impose the
12153 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
12155 2013-05-15 Richard Biener <rguenther@suse.de>
12157 PR tree-optimization/57275
12158 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
12159 return value for fail to do runtime alias checks for gather loads.
12161 2013-05-15 Jan Hubicka <jh@suse.cz>
12165 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
12166 weakrefs are not external.
12167 (lto_symtab_merge_decls): Fix thinko when dealing with
12168 non-lto_symtab decls.
12169 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
12170 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
12171 * varpool.c (dump_varpool_node): Dump more flags.
12173 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12175 * config/i386/i386.c (processor_alias_table): Add instruction
12176 FSGSBASE for AMD bdver3 architecture.
12178 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12180 * tree.c (warn_deprecated_use): Print file:line using locus color.
12181 * diagnostic.c (diagnostic_report_current_module): Print file:line
12182 and file:line:column using locus color.
12184 2013-05-14 Mike Stump <mikestump@comcast.net>
12186 * gdbinit.in: Add __null.
12188 2013-05-14 Mike Stump <mikestump@comcast.net>
12190 * recog.h: Rename struct recog_data to Recog_data.
12191 * recog.c: Likewise.
12192 * reload.c (can_reload_into): Likewise.
12193 * config/picochip/picochip.c: Likewise.
12195 2013-05-14 Mike Stump <mikestump@comcast.net>
12197 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
12199 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
12201 * resource.h (struct resources): Remove unch_memory member.
12202 (CLEAR_RESOURCE): Don't clear unch_memory.
12203 * resource.c (mark_referenced_resources): Don't set it.
12204 (mark_set_resources): Likewise.
12205 (mark_target_live_regs): Don't clear it.
12206 (init_resource_info): Likewise.
12207 * reorg.c (resource_conflicts_p): Don't compare it.
12208 (redundant_insn): Don't set it.
12210 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
12212 * emit-rtl.c (next_label): Remove unused function.
12213 (skip_consecutive_labels, link_cc0_insns): Move to ...
12214 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
12215 only place where these functions are used, and make them static.
12217 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12219 * fold-const.c (fold_negate_expr): Handle vectors.
12220 (fold_truth_not_expr): Make it static.
12221 (fold_invert_truthvalue): New static function.
12222 (invert_truthvalue_loc): Handle vectors. Do not call
12223 fold_truth_not_expr directly.
12224 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
12225 <TRUTH_NOT_EXPR>: Do not cast to boolean.
12226 (fold_comparison): Handle vector constants.
12227 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
12228 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
12229 * tree.h (fold_truth_not_expr): Remove declaration.
12231 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
12233 * config/aarch64/aarch64-simd.md
12234 (aarch64_vcond_internal<mode>): Rename to...
12235 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
12236 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
12237 float modes. Clarify all iterator modes.
12238 (vcond<mode><mode>): Use new name for vcond expanders.
12239 (vcond<v_cmp_result><mode>): Likewise.
12240 (vcondu<mode><mode>: Likewise.
12241 * config/aarch64/iterators.md (VDQF_COND): New.
12243 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
12246 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
12247 variable for the shift amount. Check that we shift by non-negative
12250 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
12253 * config/arm/arm.h (EPILOGUE_USES): Only return true
12254 for LR_REGNUM after epilogue_completed.
12256 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
12258 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
12259 is error_mark_node.
12261 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12264 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
12265 and Solaris 11+/x86 with gld.
12266 * configure: Regenerate.
12268 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12270 * expmed.c (expand_shift_1): Canonicalize rotates by
12271 constant bitsize / 2 to bitsize - 1.
12272 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
12273 case ROTATERT>: Likewise.
12276 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12278 * config/i386/i386.md (rotateinv): New code attr.
12279 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12280 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12281 roll $31, %eax, etc.
12283 2013-05-14 Richard Biener <rguenther@suse.de>
12285 PR middle-end/57235
12286 * tree-eh.c (sink_clobbers): Give up for successors with
12287 multiple predecessors and no virtual uses.
12289 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
12291 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
12292 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
12294 2013-05-14 Jakub Jelinek <jakub@redhat.com>
12296 PR middle-end/57251
12297 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
12298 the case when both op0 and op1 have VOIDmode.
12300 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12302 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
12303 in multiply-accumulate mode.
12305 2013-05-13 Guozhi Wei <carrot@google.com>
12307 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
12309 2013-05-13 Kai Tietz <ktietz@redhat.com>
12312 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
12313 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
12314 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
12315 * config/i386/i386.c (ix86_option_override_internal): Likewise.
12316 (ix86_expand_prologue): Likewise.
12317 (ix86_expand_split_stack_prologue): Likewise.
12318 (legitimate_pic_address_disp_p): Likewise.
12319 (legitimize_pic_address): Likewise.
12320 (legitimize_tls_address): Likewise.
12321 (legitimize_pe_coff_symbol): Likewise.
12322 (output_pic_addr_const): Likewise.
12323 (construct_plt_address): Likewise.
12324 (ix86_expand_call): Likewise.
12325 (x86_output_mi_thunk): Likewise.
12326 (x86_function_profiler): Likewise.
12328 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
12330 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
12331 similar switch cases.
12332 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
12333 (aarch64_simd_mov_to_<mode>low): Delete.
12334 (aarch64_simd_mov_to_<mode>high): Delete.
12335 (move_lo_quad_<mode>): Add w<-r alternative.
12336 (aarch64_simd_move_hi_quad_<mode>): Likewise.
12337 (aarch64_simd_mov_from_*): Update type attribute.
12338 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
12341 2013-05-13 Jan Hubicka <jh@suse.cz>
12343 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
12344 * config/i386/i386.c (ix86_compute_frame_layout,
12345 ix86_expand_epilogue, emit_i387_cw_initialization,
12346 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
12347 ix86_local_alignment): Fix use of size/speed predicates.
12349 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12351 PR tree-optimization/45216
12352 PR tree-optimization/57157
12353 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
12354 the (-Y) & (B - 1) variant if OP is |.
12355 * expmed.c (expand_shift_1): For rotations by const0_rtx just
12356 return shifted. Use (-op1) & (prec - 1) as other_amount
12357 instead of prec - op1.
12359 2013-05-13 Martin Jambor <mjambor@suse.cz>
12361 PR middle-end/42371
12362 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
12363 (ipa_constant_data): New type.
12364 (ipa_jump_func): Use ipa_constant_data to hold information about
12365 constant jump functions.
12366 (ipa_get_jf_constant): Adjust to jump function type changes.
12367 (ipa_get_jf_constant_rdesc): New function.
12368 (ipa_param_descriptor): New field controlled_uses.
12369 (ipa_get_controlled_uses): New function.
12370 (ipa_set_controlled_uses): Likewise.
12371 * ipa-ref.h (ipa_find_reference): Declare.
12372 * ipa-prop.c (ipa_cst_ref_desc): New type.
12373 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
12375 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
12376 New parameter cs. Adjust all callers.
12377 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
12378 (remove_described_reference): New function.
12379 (jfunc_rdesc_usable): Likewise.
12380 (try_make_edge_direct_simple_call): Decrement controlled use count,
12381 attempt to remove reference if it hits zero.
12382 (combine_controlled_uses_counters): New function.
12383 (propagate_controlled_uses): Likewise.
12384 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
12385 (ipa_edge_duplication_hook): Duplicate reference descriptions.
12386 (ipa_print_node_params): Print described use counter.
12387 (ipa_write_jump_function): Adjust to jump function type changes.
12388 (ipa_read_jump_function): New parameter CS, pass it to
12389 ipa_set_jf_constant. Adjust caller.
12390 (ipa_write_node_info): Stream controlled use count
12391 (ipa_read_node_info): Likewise.
12392 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
12394 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
12395 count. Remove cloning-added reference if it reaches zero.
12396 * ipa-ref.c (ipa_find_reference): New function.
12398 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12400 * config/i386/i386.c (processor_target_table): Modified default
12401 alignment values for AMD BD and BT architectures.
12403 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
12405 * tree-vect-generic.c (uniform_vector_p): Move ...
12406 * tree.c (uniform_vector_p): ... here.
12407 * tree.h (uniform_vector_p): Declare it.
12408 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
12411 2013-05-13 Jakub Jelinek <jakub@redhat.com>
12413 PR tree-optimization/57230
12414 * tree-ssa-strlen.c (handle_char_store): Record length for
12415 array store from STRING_CST.
12417 PR tree-optimization/57230
12418 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
12421 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
12423 * config/epiphany/epiphany.c (epiphany_init): Check size of
12424 NUM_MODES_FOR_MODE_SWITCHING.
12425 (epiphany_expand_prologue):
12426 Remove CONFIG_REGNUM initial value handling code.
12427 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
12428 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
12429 (emit_set_fp_mode, epiphany_mode_after): Likewise.
12430 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
12431 Don't return 1 for FP_MODE_NONE.
12432 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
12433 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
12434 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
12435 * config/epiphany/epiphany.md (save_config): New pattern.
12437 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
12439 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
12441 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
12443 * config/i386/i386.md (memory): Handle sseishft1.
12444 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
12445 (*vec_extractv2di_1): Ditto.
12447 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
12449 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
12452 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
12454 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
12455 Add mthumb/march=armv7-a multilib.
12456 Add mthumb/march=armv7-r multilib.
12457 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12459 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12461 * config/v850/t-rtems: Add more multilibs.
12463 2013-05-10 Richard Biener <rguenther@suse.de>
12465 PR tree-optimization/57214
12466 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12467 not propagate from SSA names that occur in abnormal PHI nodes.
12469 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
12471 * stor-layout.c (element_precision): New function.
12472 * machmode.h (element_precision): Declare it.
12473 * tree.c (build_minus_one_cst): New function.
12474 (element_precision): Likewise.
12475 * tree.h (build_minus_one_cst): Declare new function.
12476 (element_precision): Likewise.
12477 * fold-const.c (operand_equal_p): Use element_precision.
12478 (fold_binary_loc): Handle vector types.
12479 * convert.c (convert_to_integer): Use element_precision.
12480 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12483 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12485 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12486 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12487 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12488 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12489 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12490 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12491 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12492 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12493 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12494 (Uuw8): New constraints.
12495 (Usb4): Move into alphabetical order.
12496 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12497 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12498 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12499 previously unnamed patterns.
12500 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12501 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12502 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12503 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12504 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12505 of specific lengths.
12507 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12509 * config/i386/i386.md (rotateinv): New code attr.
12510 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12511 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12512 roll $31, %eax, etc.
12514 PR tree-optimization/45216
12515 PR tree-optimization/57157
12516 * tree-ssa-forwprop.c (simplify_rotate): New function.
12517 (ssa_forward_propagate_and_combine): Call it.
12519 2013-05-10 Richard Biener <rguenther@suse.de>
12521 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12522 disable peeling when we version for aliasing.
12523 (vector_alignment_reachable_p): Honor explicit user alignment.
12524 (vect_supportable_dr_alignment): Likewise.
12525 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12526 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12527 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12528 then peeling to arrange for the cost-model check to come first.
12530 2013-05-10 Alan Modra <amodra@gmail.com>
12532 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12533 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12534 * configure: Regenerate.
12536 2013-05-10 Alan Modra <amodra@gmail.com>
12539 * varasm.c (default_elf_select_section): Move !DECL_P check..
12540 (get_named_section): ..to here before calling get_section_name.
12542 (default_section_type_flags): Add DECL_P check.
12543 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12544 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12546 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12548 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12549 When using gen_stack_adjust_str with a register offset, add a
12550 REG_FRAME_RELATED_EXPR note.
12552 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
12554 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12555 (*vec_extractv4si_zext_mem): Ditto.
12556 (*vec_extractv2di): Add 0->x and x->x alternatives.
12557 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12558 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12560 2013-05-09 Jason Merrill <jason@redhat.com>
12562 N3639 C++1y VLA support
12563 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12566 * tree.c (build_constructor_va): New.
12567 * tree.h: Declare it.
12569 2013-05-09 Martin Jambor <mjambor@suse.cz>
12572 * gimple-fold.c (canonicalize_constructor_val): Call
12573 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12575 2013-05-09 Jan Hubicka <jh@suse.cz>
12576 Richard Biener <rguenther@suse.de>
12579 * symtab.c (symtab_make_decl_local): Do not add private names.
12581 2013-05-09 Jan Hubicka <jh@suse.cz>
12584 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12585 (unlink_from_assembler_name_hash): Likewise.
12586 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12587 symtab_unregister_node, symtab_initialize_asm_name_hash,
12588 change_decl_assembler_name): Update.
12590 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12592 * config/aarch64/aarch64.md: New movtf split.
12593 (*movtf_aarch64): Update.
12594 (aarch64_movdi_tilow): Handle TF modes and rename to
12595 aarch64_movdi_<mode>low.
12596 (aarch64_movdi_tihigh): Handle TF modes and rename to
12597 aarch64_movdi_<mode>high
12598 (aarch64_movtihigh_di): Handle TF modes and rename to
12599 aarch64_mov<mode>high_di
12600 (aarch64_movtilow_di): Handle TF modes and rename to
12601 aarch64_mov<mode>low_di
12602 (aarch64_movtilow_tilow): Remove spurious whitespace.
12603 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12605 (aarch64_print_operand): Update.
12607 2013-05-09 Alan Modra <amodra@gmail.com>
12609 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12611 * configure: Regenerate.
12613 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12615 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12616 splitter preparation statements.
12617 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12618 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12621 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12623 * gimple-ssa-strength-reduction.c (count_candidates): Change
12624 return value to int.
12625 (analyze_candidates_and_replace): Change type of length to int.
12627 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
12629 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12630 (*vec_extract<mode>): Use VI12_128 mode iterator.
12631 (*vec_extract<mode>_mem): Ditto.
12632 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12635 2013-05-08 Diego Novillo <dnovillo@google.com>
12640 2012-08-17 Diego Novillo <dnovillo@google.com>
12643 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12644 * config.in: Regenerate.
12645 * configure: Regenerate.
12646 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
12648 2013-05-08 Jan Hubicka <jh@suse.cz>
12651 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12652 * cgraph.h (symtab_node_base): Add unique_name.
12653 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12654 input_overwrite_node, input_varpool_node): Stream unique_name.
12655 * cgraphclones.c (cgraph_create_virtual_clone,
12656 cgraph_function_versioning): Set unique_name.
12657 * ipa.c (function_and_variable_visibility): Set unique_name.
12659 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12661 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12662 (alloc_cand_and_find_basis): Restrict conditional candidate
12663 processing to CAND_MULTs.
12665 2013-05-08 Jan Hubicka <jh@suse.cz>
12668 lto-symtab.c (lto_symtab_symbol_p): New function.
12669 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12670 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12671 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12672 Skip static symbols.
12674 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12676 PR tree-optimization/57200
12677 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12678 Only call inform if the preceding warning_at returns true.
12680 2013-05-07 Han Shen <shenhan@google.com>
12682 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12683 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12684 * common.opt (fstack-protector-strong): New option.
12685 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12686 * doc/invoke.texi (Optimization Options): Document
12687 "-fstack-protector-strong".
12688 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12690 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12692 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12694 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12696 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12697 it is not lhs of the insn.
12698 (setup_reg_equiv): Remove insn having equiv memory which it is not
12700 * lra-constraints.c (process_address): Try to improve generation
12701 code for address base + disp.
12702 (lra_constraints): Make correct the code for checking insn setting
12703 up backward equivalence. Remove insn only if it is in the init
12705 * lra-eliminations.c (update_reg_eliminate): Change return value.
12706 (lra_eliminate): Use the result.
12708 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
12710 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12711 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12712 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12713 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12714 register target operands.
12715 (*vec_extractv8hi_sse2): New pattern.
12716 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12717 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12718 (*vec_extract<mode>_mem): New insn and split pattern.
12720 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12722 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12723 (TARGET_ASAN_SHADOW_OFFSET): Define.
12724 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12725 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12727 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12729 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12730 (incr_vec_index): Return -1 if increment not found.
12731 (create_add_on_incoming_edge): Assert if increment not found.
12732 (record_increment): Limit number of increments recorded.
12733 (all_phi_incrs_profitable): Return false if an increment not found.
12734 (replace_profitable_candidates): Don't process increments that were
12736 (analyze_candidates_and_replace): Limit size of incr_vec.
12738 2013-05-07 Richard Biener <rguenther@suse.de>
12740 * calls.c (special_function_p): setjmp-like functions are leaf.
12741 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12742 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12744 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12746 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12747 (aarch64_simd_mov<mode>): New expander.
12748 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12749 (aarch64_simd_mov_to_<mode>high): Likewise.
12750 (aarch64_simd_mov_from_<mode>low): Likewise.
12751 (aarch64_simd_mov_from_<mode>high): Likewise.
12752 (aarch64_dup_lane<mode>): Update.
12753 (aarch64_dup_lanedi): New instruction pattern.
12754 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12755 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12757 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12759 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12760 (replace_mult_candidate): Remove unnecessary argument; remove
12761 unnecessary parameter from call to introduce_cast_before_cand.
12762 (replace_unconditional_candidate): Remove unnecessary parameter
12763 from call to replace_mult_candidate.
12764 (replace_conditional_candidate): Likewise.
12765 (insert_initializers): Use make_temp_ssa_name.
12766 (introduce_cast_before_cand): Remove unnecessary argument; use
12767 make_temp_ssa_name.
12768 (replace_one_candidate): Remove unnecessary argument; remove
12769 unnecessary parameter from calls to introduce_cast_before_cand.
12770 (replace_profitable_candidates): Remove unnecessary parameters
12771 from calls to replace_one_candidate.
12773 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12775 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12776 phi def as possibly hiding a basis for a CAND_ADD whose operands
12777 have been commuted in the analysis.
12778 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12780 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12782 * config/aarch64/aarch64.md
12783 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12784 shift value between 0-4.
12786 2013-05-07 Richard Biener <rguenther@suse.de>
12788 * double-int.h (rshift): New overload.
12789 * double-int.c (rshift): New function.
12790 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12791 (create_reference_ops_from_ref): Remove.
12792 (vn_reference_insert): Use shared ops for constructing the
12793 reference and copy it.
12795 2013-05-07 Richard Biener <rguenther@suse.de>
12797 PR middle-end/57190
12798 * tree-eh.c (sink_clobbers): Properly propagate
12799 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12801 2013-05-07 Jakub Jelinek <jakub@redhat.com>
12803 PR tree-optimization/57149
12804 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12805 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12806 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12807 uninit_undefined_value_p instead of ssa_undefined_value_p.
12810 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12811 for modifier == EXPAND_INITIALIZER.
12813 2013-05-07 Anton Blanchard <anton@samba.org>
12815 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12816 for powerpc64 little endian.
12817 * configure: Regenerate.
12819 2013-05-06 Graham Stott <grahams@btinternet.com>
12821 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12822 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12823 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12824 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12826 2013-05-06 Graham Stott <grahams@btinternet.com>
12828 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12829 codes which allow non-lvalues.
12831 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
12833 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12834 components are all 1s.
12835 (integer_minus_onep): New function.
12836 * tree.h (integer_minus_onep): Declare it.
12837 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12838 integer_minus_onep instead of integer_all_onesp.
12840 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12843 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12844 variations of these patterns.
12846 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12848 * config/i386/i386.md (isa): Add x64_sse4 member.
12849 (enabled): Handle x64_sse4.
12850 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12851 instruction for 64bit SSE4_1 targets. Update insn attributes.
12852 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12853 instruction for SSE4_1 targets. Update insn attributes.
12854 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12855 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12856 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12858 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12859 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12860 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12862 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12865 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12867 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12869 * final.c (do_assembler_dialects): Don't handle curly braces and
12870 vertical bar escaped by % as dialect delimiters.
12871 (output_asm_insn): Print curly braces and vertical bar if escaped
12872 by % and ASSEMBLER_DIALECT defined.
12873 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12874 * doc/tm.texi: Regenerated.
12876 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12878 * config/mips/mips.c: Include tree-pass.h.
12879 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12880 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12881 (pass_mips_machine_reorg2): New machine specific pass.
12882 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12883 (mips_option_override): Register the new pass.
12884 * rtl.h (cleanup_barriers): Remove prototype.
12885 (dbr_schedule): Likewise.
12886 * jump.c (cleanup_barriers): Make static.
12887 * reorg.c (dbr_schedule): Likewise.
12889 2013-05-06 Richard Biener <rguenther@suse.de>
12891 PR tree-optimization/57185
12892 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12893 SSA names properly.
12895 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
12898 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12899 between "first_arg" and "insn", not "last" and "insn".
12901 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12903 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12904 (find_candidates_in_block): Re-enable slsr_process_phi.
12905 (create_phi_basis): Fix double counting of candidate adjustment.
12907 2013-05-06 Richard Biener <rguenther@suse.de>
12909 PR middle-end/57147
12910 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12911 the edge is also fallthru, preserve it and just clear the
12913 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12914 also complex, preserve that and just clear the fallthru flag.
12915 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12916 update virtual operands.
12918 2013-05-06 Alan Modra <amodra@gmail.com>
12920 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12921 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12922 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12923 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12924 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12925 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12927 2013-05-06 Alan Modra <amodra@gmail.com>
12929 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12930 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12931 (DEFAULT_ASM_ENDIAN): Define.
12932 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12933 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12934 Update -K PIC clause from sysv4.h.
12935 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12936 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12938 2013-05-06 Alan Modra <amodra@gmail.com>
12940 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12941 twice for little-endian.
12942 (ashrdi3_no_power, ashrdi3): Support little-endian.
12944 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12947 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12948 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12950 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12951 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12952 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12955 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12956 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12958 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12959 * config/pa/pa.opt: Make mbig-switch a no-op.
12960 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12961 (CASE_VECTOR_MODE): Always return SImode.
12962 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12963 for the !TARGET_BIG_SWITCH case.
12964 * config/pa/pa-linux.h: Likewise.
12965 * config/pa/pa-openbsd.h: Likewise.
12966 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12967 * config/pa/pa.md (short_jump): Remove define_insn.
12968 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12969 (casesi0): Remove define_insn.
12970 (type): Remove btable_branch.
12971 (pa_combine_type): Likewise.
12972 (in_nullified_branch_delay): Likewise.
12973 (in_call_delay): Likewise.
12974 (define_delay): Likewise.
12975 (define_insn_reservation "Z3"): Likewise.
12976 (define_insn_reservation "Z4"): Likewise.
12977 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12978 (pa_adjust_insn_length): Remove adjustment for btable branches.
12979 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12982 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
12984 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12985 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12986 Add m->r,x alternatives.
12987 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12988 splitters using SWI48x mode iterator.
12989 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12990 TARGET_64BIT. Add m->x alternative.
12991 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12992 Add o->x alternative. Enable for TARGET_SSE.
12993 (sse_storeq): Remove expander.
12994 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12995 with memory input operand.
12996 (*vec_extractv2di_1 splitter): New.
12997 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12998 * config/i386/i386.md (ssevecmodelower): New mode attribute.
13000 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13002 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
13003 (INT_LOWPART): Delete.
13004 (extract_MB): Adjust.
13005 (extract_ME): Adjust.
13006 (print_operand): Adjust.
13008 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
13010 * config/rs6000/predicates.md (reg_or_add_cint_operand,
13011 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13012 (reg_or_logical_cint_operand, easy_fp_constant,
13013 logical_const_operand): Delete "CONST_DOUBLE" case.
13014 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
13015 "HOST_BITS_PER_WIDE_INT == 64" test.
13016 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
13017 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
13018 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
13019 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
13021 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
13022 CONST_DOUBLE DImode/VOIDmode case.
13023 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
13024 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13025 CONST_DOUBLE VOIDmode case.
13026 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
13027 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
13028 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
13029 Delete CONST_DOUBLE case.
13030 (splitters for mov FMOVE64 const_double): Delete
13031 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
13032 "HOST_BITS_PER_WIDE_INT >= 64" test.
13033 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
13035 (mov DI const_double): Delete.
13037 2013-05-04 Jakub Jelinek <jakub@redhat.com>
13039 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
13040 on op shows all bits zero in mode of a lowpart subreg, return zero.
13042 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
13045 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
13046 to save TFmode registers and DImode to save TImode registers for
13047 caller save operations.
13048 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
13049 mark being partially clobbered since they only use the first
13052 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
13053 and TDmode only use the upper 64-bits of each VSX register.
13055 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13057 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
13058 (find_candidates_in_block): Disable slsr_process_phi.
13060 2013-05-03 Guozhi Wei <carrot@google.com>
13062 * coverage.c (coverage_obj_init): Move the construction of gcov
13064 (build_init_ctor): ... here.
13066 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
13068 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
13069 (slsr_cand_d): Redefine def_phi.
13070 (stride_status, phi_adjust_status, count_phis_status): New enums.
13071 (find_phi_def): New.
13072 (find_basis_for_base_expr): New.
13073 (find_basis_for_candidate): Handle hidden bases.
13074 (alloc_cand_and_find_basis): Handle phi candidates.
13075 (slsr_process_phi): New.
13076 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
13077 (create_mul_imm_cand): Likewise.
13078 (create_add_ssa_cand): Exclude phi base candidates.
13079 (create_add_imm_cand): Likewise.
13080 (slsr_process_cast): Likewise.
13081 (slsr_process_copy): Likewise.
13082 (find_candidates_in_block): Handle phi candidates.
13083 (dump_candidate): Likewise.
13084 (unconditional_cands): Delete.
13085 (unconditional_cands_with_known_stride_p): Delete.
13086 (phi_dependent_cand_p): New.
13087 (cand_increment): Handle phi-dependent candidates.
13088 (replace_dependent): Delete.
13089 (replace_mult_candidate): New.
13090 (replace_unconditional_candidate): New.
13091 (incr_vec_index): Move to avoid forward reference.
13092 (create_add_on_incoming_edge): New.
13093 (create_phi_basis): New.
13094 (replace_dependents): Delete.
13095 (replace_conditional_candidate): New.
13096 (phi_add_costs): New.
13097 (replace_uncond_cands_and_profitable_phis): New.
13098 (record_increment): Handle phi adjustments.
13099 (record_phi_increments): New.
13100 (record_increments): Handle phi adjustments.
13101 (phi_incr_cost): New.
13102 (lowest_cost_path): Handle phis.
13103 (total_savings): Likewise.
13104 (analyze_increments): Likewise.
13105 (ncd_with_phi): New.
13106 (ncd_of_cand_and_phis): New.
13107 (nearest_common_dominator_for_cands): Handle phi increments.
13108 (all_phi_incrs_profitable): New.
13109 (replace_profitable_candidates): Handle phi-dependent candidates.
13110 (analyze_candidates_and_replace): Likewise.
13112 2013-05-03 Teresa Johnson <tejohnson@google.com>
13115 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
13116 do not exceed REG_BR_PROB_BASE.
13118 2013-05-03 Jeff Law <law@redhat.com>
13120 PR tree-optimization/57144
13121 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
13122 operand of the condition will bit into the new type when eliminating
13123 a cast feeding a condition.
13125 2013-05-03 Jakub Jelinek <jakub@redhat.com>
13127 PR rtl-optimization/57130
13128 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
13129 of COMPARE as in_code to the recursive call if needed.
13131 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
13133 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
13134 (enabled): Handle new members.
13135 * config/i386/sse.md (*vec_concatv2si): Merge from
13136 *vec_concatv2si_sse2 and vec_concatv2si_sse.
13137 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
13139 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
13141 PR tree-optimization/57027
13142 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
13143 for fnms opportunity, check we got the prerequisite kind
13144 of tree / gimple before using accessor functions.
13146 2013-05-03 Richard Biener <rguenther@suse.de>
13148 * double-int.h (lshift): New overload without precision
13149 and arith argument.
13150 (operator *=, operator +=, operator -=): Move ...
13151 * double-int.c (operator *=, operator +=, operator -=): ... here
13152 and implement more efficiently.
13153 (mul_double_with_sign): Remove.
13154 (lshift_double): Adjust to take unsinged shift argument, push
13155 dispatching code to callers.
13156 (mul_double_wide_with_sign): Add early out for callers that
13157 are not interested in high parts or overflow.
13158 (lshift): New function.
13159 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
13160 dispatch code here.
13161 (lrotate, rrotate): Use logical shifts.
13162 * expr.c (get_inner_reference): Use lshift.
13163 * fixed-value.c (do_fixed_divide): Likewise.
13164 * tree-dfa.c (get_ref_base_and_extent): Likewise.
13165 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
13166 (indirect_refs_may_alias_p): Likewise.
13167 (stmt_kills_ref_p_1): Likewise.
13169 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13171 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
13173 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
13175 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
13176 scalar form of FABD instruction.
13178 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
13180 * lra-constraints.c (process_alt_operands): Add checking alt
13181 number to choose the best alternative.
13183 2013-05-02 Richard Biener <rguenther@suse.de>
13185 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
13186 bitmap and its handling.
13187 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
13189 2013-05-02 Richard Biener <rguenther@suse.de>
13191 PR middle-end/57140
13192 * tree-inline.c (copy_loops): Properly handle removed loops.
13193 (copy_cfg_body): Mark destination loops for fixup if source
13194 loops needed fixup.
13196 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
13199 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
13200 generating simple_return for naked functions.
13202 2013-05-02 Martin Jambor <mjambor@suse.cz>
13204 PR middle-end/56988
13205 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
13206 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
13208 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
13209 ipa_agg_replacement_value structures.
13210 (known_aggs_to_agg_replacement_list): Likewise.
13211 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
13212 (read_agg_replacement_chain): Likewise.
13213 (ipcp_transform_function): Also check that by_ref flags match.
13215 2013-05-02 Richard Biener <rguenther@suse.de>
13217 * graphds.h (struct graph): Add obstack member.
13218 * graphds.c (new_graph): Initialize obstack and allocate
13220 (add_edge): Allocate edge from the obstack.
13221 (free_graph): Free the obstack instead of all edges and vertices.
13223 2013-05-02 Teresa Johnson <tejohnson@google.com>
13225 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
13227 * cfg.c (update_bb_profile_for_threading): Ditto.
13228 * tree-inline.c (copy_bb): Ditto.
13229 (copy_edges_for_bb): Ditto.
13230 (initialize_cfun): Ditto.
13231 (copy_cfg_body): Ditto.
13232 (expand_call_inline): Ditto.
13233 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
13234 (estimate_node_size_and_time): Ditto.
13235 (inline_merge_summary): Ditto.
13236 * cgraphclones.c (cgraph_clone_edge): Ditto.
13237 (cgraph_clone_node): Ditto.
13238 * sched-rgn.c (compute_dom_prob_ps): Ditto.
13239 (compute_trg_info): Ditto.
13241 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13243 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
13244 S reg when fp attribute set.
13245 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
13247 2013-05-02 Ian Bolton <ian.bolton@arm.com>
13249 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
13251 (*and_one_cmplsi3_compare0_uxtw): Likewise.
13252 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
13253 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
13255 2013-05-02 Richard Biener <rguenther@suse.de>
13257 * tree-scalar-evolution.c (scev_info_hasher): Remove.
13258 (struct instantiate_cache_entry): New type.
13259 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
13260 (struct instantiate_cache_type): New type.
13261 (set_instantiated_value, get_instantiated_value): Remove.
13262 (get_instantiated_value_entry): New function.
13263 (instantiate_scev_name): Use the new cache and adjust.
13264 (instantiate_scev_poly): Adjust.
13265 (instantiate_scev_binary): Likewise.
13266 (instantiate_array_ref): Likewise.
13267 (instantiate_scev_convert): Likewise.
13268 (instantiate_scev_not): Likewise.
13269 (instantiate_scev_3): Likewise.
13270 (instantiate_scev_2): Likewise.
13271 (instantiate_scev_r): Likewise.
13272 (instantiate_scev): Likewise.
13273 (resolve_mixers): Likewise.
13275 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
13278 * lra-constraints.c (best_small_class_operands_num): Remove.
13279 (process_alt_operands): Remove small_class_operands_num. Take
13280 small classes operands into losers and only if the operand is not
13281 matched. Modify debugging output.
13282 (curr_insn_transform): Remove best_small_class_operands_num.
13285 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13287 * config/aarch64/aarch64-builtins.c
13288 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
13289 * config/aarch64/aarch64-simd-builtins.def
13290 (reduc_splus_): Add new modes.
13291 (reduc_uplus_): New.
13292 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
13293 (reduc_uplus_v4sf): Likewise.
13294 (reduc_splus_v4sf): Likewise.
13295 (aarch64_addv<mode>): Likewise.
13296 (reduc_uplus_<mode>): Likewise.
13297 (reduc_splus_<mode>): Likewise.
13298 (aarch64_addvv2di): Likewise.
13299 (reduc_uplus_v2di): Likewise.
13300 (reduc_splus_v2di): Likewise.
13301 (aarch64_addvv2si): Likewise.
13302 (reduc_uplus_v2si): Likewise.
13303 (reduc_splus_v2si): Likewise.
13304 (reduc_<sur>plus_<mode>): New.
13305 (reduc_<sur>plus_v2di): Likewise.
13306 (reduc_<sur>plus_v2si): Likewise.
13307 (reduc_<sur>plus_v4sf): Likewise.
13308 (aarch64_addpv4sf): Likewise.
13309 * config/aarch64/arm_neon.h
13310 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
13311 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
13312 add UNSPEC_SADDV, UNSPEC_UADDV.
13314 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
13316 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13318 * config/aarch64/arm_neon.h
13319 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
13321 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13323 * config/aarch64/aarch64-builtins
13324 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
13326 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13328 * config/aarch64/aarch64-simd-builtins.def
13329 (reduc_smax_): New.
13330 (reduc_smin_): Likewise.
13331 (reduc_umax_): Likewise.
13332 (reduc_umin_): Likewise.
13333 (reduc_smax_nan_): Likewise.
13334 (reduc_smin_nan_): Likewise.
13337 (smax): Update for V2SF, V4SF and V2DF modes.
13340 (smin_nan): Likewise.
13341 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
13342 (<su><maxmin><mode>3): ...This, refactor.
13343 (s<maxmin><mode>3): New.
13344 (<maxmin_uns><mode>3): Likewise.
13345 (reduc_<maxmin_uns>_<mode>): Refactor.
13346 (reduc_<maxmin_uns>_v4sf): Likewise.
13347 (reduc_<maxmin_uns>_v2si): Likewise.
13348 (aarch64_<fmaxmin><mode>: Remove.
13349 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
13351 (vmin<q>_f<32,64>): Likewise.
13352 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13354 (su): Add mappings for smax, smin, umax, umin.
13356 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13357 (FMAXMIN): Rename as...
13358 (FMAXMIN_UNS): ...This.
13360 (fmaxminv): Likewise.
13361 (fmaxmin): Likewise.
13363 (maxmin_uns_op): Likewise.
13365 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13367 * config/aarch64/arm_neon.h
13368 (vac<ge, gt><sd>_f<32, 64>): Rename to...
13369 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
13370 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
13372 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13374 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
13375 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
13377 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13379 * config/aarch64/aarch64-simd.md
13380 (vcond<mode>_internal): Handle special cases for constant masks.
13381 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
13382 (vcondu<mode><mode>): Likewise.
13383 (vcond<v_cmp_result><mode>): New.
13385 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13387 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
13388 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
13389 * config/aarch64/aarch64-simd-builtins.def
13390 (cmeq): Update to BUILTIN_VALLDI.
13395 * config/aarch64/arm_neon.h
13396 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
13397 to builtins or C as appropriate.
13399 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13401 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
13403 (cmhi): Rename to...
13405 * config/aarch64/aarch64-simd.md
13406 (simd_mode): Add SF.
13407 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
13408 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
13409 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
13410 (cstore<mode>_neg): ...This.
13411 * config/aarch64/iterators.md
13413 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
13414 (COMPARISONS): New.
13415 (UCOMPARISONS): Likewise.
13416 (optab): Add missing comparisons.
13422 (VCMP_S): Likewise.
13423 (VCMP_U): Likewise.
13424 (V_cmp_result): Add DF, SF modes.
13425 (v_cmp_result): Likewise.
13427 (vmtype): Likewise.
13428 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
13430 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
13432 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
13433 define_insn to define_insn_and_split.
13434 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
13435 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
13436 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
13437 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
13438 (thumb2_negscc): Likewise.
13440 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13442 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
13444 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13446 * config/arm/thumb2.md: Remove trailing whitespaces.
13448 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13450 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
13451 Use gen_int_mode rather than GEN_INT.
13453 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
13455 * value-prof.c (stream_in_histogram_value): Remove the strayed
13458 2013-04-30 Richard Biener <rguenther@suse.de>
13460 PR middle-end/57122
13461 * cfghooks.c (split_edge): Properly check for the loop latch edge.
13463 2013-04-30 Richard Biener <rguenther@suse.de>
13465 PR middle-end/57107
13466 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13468 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
13470 PR rtl-optimization/56957
13471 PR rtl-optimization/57105
13472 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13473 variable. Use just INSN_UID for determining whether an insn
13474 should be only disconnected from the insn stream.
13475 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13477 2013-04-30 Jakub Jelinek <jakub@redhat.com>
13479 PR tree-optimization/57104
13480 * tsan.c (instrument_expr): Don't instrument accesses to
13481 DECL_HARD_REGISTER VAR_DECLs.
13483 2013-04-30 Richard Biener <rguenther@suse.de>
13485 * function.h (loops_for_fn): New inline function.
13486 (set_loops_for_fn): Likewise.
13487 * cfgloop.h (place_new_loop): Add struct function parameter.
13488 (get_loop): Likewise.
13489 (get_loops): Likewise.
13490 (number_of_loops): Likewise.
13491 (fel_next): Adjust.
13492 (fel_init): Likewise.
13493 * cfg.c (get_loop_copy): Adjust.
13494 * cfgloop.c (flow_loops_dump): Likewise.
13495 (record_loop_exits): Likewise.
13496 (verify_loop_structure): Likewise.
13497 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13498 (estimate_reg_pressure_cost): Likewise.
13499 (mark_loop_exit_edges): Likewise.
13500 * cfgloopmanip.c (place_new_loop): Likewise.
13501 (add_loop): Likewise.
13502 (duplicate_loop): Likewise.
13503 * graph.c (draw_cfg_nodes): Likewise.
13504 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13505 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13506 (extract_affine_chrec): Likewise.
13507 (build_scop_iteration_domain): Likewise.
13508 * graphite.c (graphite_initialize): Likewise.
13509 * ira-build.c (create_loop_tree_nodes): Likewise.
13510 (more_one_region_p): Likewise.
13511 (rebuild_regno_allocno_maps): Likewise.
13512 (mark_loops_for_removal): Likewise.
13513 (mark_all_loops_for_removal): Likewise.
13514 (remove_unnecessary_regions): Likewise.
13515 (ira_build): Likewise.
13516 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13517 * loop-init.c (fix_loop_structure): Likewise.
13518 (gate_rtl_move_loop_invariants): Likewise.
13519 (gate_rtl_unswitch): Likewise.
13520 (gate_rtl_unroll_and_peel_loops): Likewise.
13521 (rtl_doloop): Likewise.
13522 * lto-streamer-in.c (input_cfg): Likewise.
13523 * lto-streamer-out.c (output_cfg): Likewise.
13524 * modulo-sched.c (sms_schedule): Likewise.
13525 * predict.c (tree_estimate_probability): Likewise.
13526 (tree_estimate_probability_driver): Likewise.
13527 (estimate_loops): Likewise.
13528 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13529 (move_sese_region_to_fn): Likewise.
13530 (debug_loop_num): Likewise.
13531 * tree-chrec.c (chrec_evaluate): Likewise.
13532 (hide_evolution_in_other_loops_than_loop): Likewise.
13533 (chrec_component_in_loop_num): Likewise.
13534 (reset_evolution_in_loop): Likewise.
13535 (evolution_function_is_invariant_rec_p): Likewise.
13536 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13537 * tree-inline.c (copy_loops): Likewise.
13538 (copy_cfg_body): Likewise.
13539 (tree_function_versioning): Likewise.
13540 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13541 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13543 (add_to_evolution_1): Likewise.
13544 (scev_const_prop): Likewise.
13545 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13546 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13547 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13548 (tree_ssa_lim_initialize): Likewise.
13549 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13550 (verify_loop_closed_ssa): Likewise.
13551 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13552 (tree_ssa_loop_im): Likewise.
13553 (tree_ssa_loop_unswitch): Likewise.
13554 (tree_vectorize): Likewise.
13555 (check_data_deps): Likewise.
13556 (tree_ssa_loop_ivcanon): Likewise.
13557 (tree_ssa_loop_bounds): Likewise.
13558 (tree_complete_unroll): Likewise.
13559 (tree_complete_unroll_inner): Likewise.
13560 (tree_parallelize_loops): Likewise.
13561 (tree_ssa_loop_prefetch): Likewise.
13562 (tree_ssa_loop_ivopts): Likewise.
13563 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13564 * tree-vectorizer.c (vectorize_loops): Likewise.
13566 2013-04-29 Mike Frysinger <vapier@gentoo.org>
13568 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13569 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13570 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13571 with EABI_LINK_SPEC.
13573 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13576 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13579 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13582 * lra-constraints.c (process_alt_operands): Discourage a bit more
13583 using memory for pseudos. Print cost dump for alternatives.
13584 Modify cost values for conflicts with early clobbers.
13585 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13587 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13590 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13592 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13594 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13595 from/to S register.
13596 (movdi_aarch64): Support LDR/STR from/to D register.
13598 2013-04-29 Ian Bolton <ian.bolton@arm.com>
13600 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13601 or higher by default.
13603 2013-04-29 Richard Biener <rguenther@suse.de>
13605 PR middle-end/57075
13606 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13607 even if not adding abnormal edges for calls that can make
13610 2013-04-29 Richard Biener <rguenther@suse.de>
13612 PR middle-end/57103
13613 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13615 (move_stmt_r): Remove redundant checking.
13617 2013-04-29 Teresa Johnson <tejohnson@google.com>
13620 * basic-block.h (apply_scale): New function.
13621 (apply_probability): Use apply_scale.
13622 * gimple-streamer-in.c (input_bb): Ditto.
13623 * lto-streamer-in.c (input_cfg): Ditto.
13624 * lto-cgraph.c (merge_profile_summaries): Ditto.
13625 * tree-optimize.c (execute_fixup_cfg): Ditto.
13626 * tree-inline.c (copy_bb): Update comment to use apply_scale.
13627 (copy_edges_for_bb): Ditto.
13628 (copy_cfg_body): Ditto.
13630 2013-04-29 Tom de Vries <tom@codesourcery.com>
13632 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13633 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13634 (tail_merge_optimize): Handle current_loops == NULL.
13636 2013-04-26 Jeff Law <law@redhat.com>
13638 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13639 (simplify_cond_using_ranges): Generalize code to simplify
13640 COND_EXPRs where one argument is a constant and the other
13641 is an SSA_NAME created by an integral type conversion.
13643 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13645 * config/arm/arm.md (store_minmaxsi): Use only when
13646 optimize_insn_for_size_p.
13648 2013-04-29 Christian Bruel <christian.bruel@st.com>
13651 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13653 2013-04-29 Richard Biener <rguenther@suse.de>
13655 PR middle-end/57089
13656 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13657 loop tree make sure to schedule a fixup for the child as well.
13658 (expand_omp_for_generic): Properly add loops.
13659 (expand_omp_for_static_nochunk): Likewise.
13660 (expand_omp_for_static_chunk): Likewise.
13661 (expand_omp_for): For the degenerate case fixup loops.
13662 (expand_omp_sections): Fix default bb placement in loops.
13663 (expand_omp_atomic_pipeline): Properly add loops.
13665 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13667 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13669 2013-04-29 Tom de Vries <tom@codesourcery.com>
13671 * tree-ssa-tail-merge.c: Update header comment.
13673 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13675 * config/aarch64/arm_neon.h
13676 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13677 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13678 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13679 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13680 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13681 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13682 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13683 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13685 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13687 * config/aarch64/aarch64-simd.md
13688 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13689 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13690 fix_trunc, fixuns_trunc.
13691 (ftrunc<VDQF:mode>2): New.
13692 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13693 (fix_trunc_optab): New.
13695 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13697 * config/aarch64/aarch64-builtins.c
13698 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13699 iceilf, lround, iroundf.
13701 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
13704 * config/i386/i386.h (enum ix86_tune_indices)
13705 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13706 New, split from X86_TUNE_INTER_UNIT_MOVES.
13707 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13708 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13709 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13710 (TARGET_INTER_UNIT_MOVES): Remove.
13711 * config/i386/i386.c (initial_ix86_tune_features): Update.
13712 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13713 (ix86_expand_convert_uns_didf_sse): Use
13714 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13715 (ix86_expand_vector_init_one_nonzero): Ditto.
13716 (ix86_expand_vector_init_interleave): Ditto.
13717 (inline_secondary_memory_needed): Return true for moves from SSE class
13718 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13719 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13720 * config/i386/constraints.md (Yi, Ym): Depend on
13721 TARGET_INTER_UNIT_MOVES_TO_VEC.
13722 (Yj, Yn): New constraints.
13723 * config/i386/i386.md (*movdi_internal): Change constraints of
13724 operand 1 from Yi to Yj and from Ym to Yn.
13725 (*movsi_internal): Ditto.
13726 (*movdf_internal): Ditto.
13727 (*movsf_internal): Ditto.
13728 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13729 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13730 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13731 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13732 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13733 * config/i386/sse.md (movdi_to_sse): Ditto.
13734 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13735 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13736 TARGET_INTER_UNIT_MOVES.
13737 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13738 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13739 instead of TARGET_INTER_UNIT_MOVES.
13740 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13741 operand 1 from Yi to Yj and from Ym to Yn.
13743 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13745 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13746 (float_truncate_hi_): Likewise.
13747 (float_extend_lo_): Likewise.
13748 (float_truncate_lo_): Likewise.
13749 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13750 (aarch64_float_extend_lo_v2df): Likewise.
13751 (vec_unpacks_hi_v4sf): Likewise.
13752 (aarch64_float_truncate_lo_v2sf): Likewise.
13753 (aarch64_float_truncate_hi_v4sf): Likewise.
13754 (vec_pack_trunc_v2df): Likewise.
13755 (vec_pack_trunc_df): Likewise.
13757 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13759 * config/aarch64/aarch64-builtins.c
13760 (aarch64_fold_builtin): Fold float conversions.
13761 * config/aarch64/aarch64-simd-builtins.def
13762 (floatv2si, floatv4si, floatv2di): New.
13763 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13764 * config/aarch64/aarch64-simd.md
13765 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13766 * config/aarch64/iterators.md (FLOATUORS): New.
13767 (optab): Add float, floatuns.
13768 (su_optab): Likewise.
13770 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13772 * config/aarch64/aarch64-builtins.c
13773 (aarch64_builtin_vectorized_function): Use new names for
13775 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13776 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13777 (fcvtzu): Split as...
13778 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13779 (fcvtas): Split as...
13780 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13781 (fcvtau): Split as...
13782 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13783 (fcvtps): Split as...
13784 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13785 (fcvtpu): Split as...
13786 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13787 (fcvtms): Split as...
13788 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13789 (fcvtmu): Split as...
13790 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13791 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13792 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13793 (lfrintnusf, lfrintnudf): Likewise.
13794 * config/aarch64/aarch64-simd.md
13795 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13797 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13798 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13799 (fcvt_pattern): Likewise.
13801 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13803 * config/aarch64/aarch64-simd.md
13804 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13805 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13807 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13809 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13810 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13811 (vrnd<a,m,n,p>_f32): Implement using builtins.
13812 (vrnd<i,x><q>_f<32, 64>): New.
13814 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13816 * config/aarch64/aarch64-builtins.c
13817 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13818 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13819 (frintz): Rename to...
13821 (frintp): Rename to...
13823 (frintm): Rename to...
13825 (frinti): Rename to...
13826 (nearbyint): ...this.
13827 (frintx): Rename to...
13829 (frinta): Rename to...
13831 * config/aarch64/aarch64-simd.md
13832 (aarch64_frint<frint_suffix><mode>): Delete.
13833 (<frint_pattern><mode>2): Convert to insn.
13834 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13835 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13836 (frint_pattern): Likewise.
13837 (frint_suffix): Likewise.
13839 2013-04-29 Richard Biener <rguenther@suse.de>
13841 PR tree-optimization/57081
13842 * loop-init.c: Include tree-flow.h.
13843 (loop_optimizer_finalize): Free number of iteration estimates.
13844 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13846 2013-04-29 Jakub Jelinek <jakub@redhat.com>
13848 PR tree-optimization/57083
13849 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13850 non-singleton shift count range, zero extend low_bound for uns case.
13852 * config/i386/predicates.md (general_vector_operand): New predicate.
13853 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13854 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13855 if they aren't nonimmediate operands. If their original values
13856 satisfy const_vector_equal_evenodd_p, don't shift them.
13857 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13858 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13859 if not nonimmediate_operand.
13860 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13861 instead of register_operand.
13862 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13864 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
13866 * stor-layout.c (finalize_size_functions): Allocate a structure and
13867 reset cfun before dumping the functions.
13869 2013-04-27 Jakub Jelinek <jakub@redhat.com>
13871 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13874 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13875 use xop_pmacsdqh if uns_p.
13876 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13877 the immediate rotate count.
13879 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13881 * rtl.h (struct rtx_def): Add comment for field jump.
13882 (LRA_SUBREG_P): New macro.
13883 * recog.c (register_operand): Check LRA_SUBREG_P.
13884 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13886 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13887 (lra_final_code_change): Skip subreg change for operators.
13888 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13889 if there are no operand changes.
13890 * lra-constraints.c (curr_insn_set): New.
13891 (match_reload): Set LRA_SUBREG_P.
13892 (emit_spill_move): Ditto.
13893 (check_and_process_move): Use curr_insn_set. Process only single
13894 set insns. Don't initialize sec_mem_p and change_p.
13895 (simplify_operand_subreg): Use LRA_SUBREG_P.
13896 (reg_in_class_p): New function.
13897 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13898 of #ifdef. Add code to remove cycling.
13899 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13900 non-null disp. Reload inner instead of disp when base and index
13901 are null. Try to put lo_sum into register.
13902 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13903 (check_and_process_move): Move code for move cost check to
13904 simple_move_p. Remove equiv_substitution.
13905 (simple_move_p): New function.
13906 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13907 curr_insn_set. Call check_and_process_move only for single set
13908 insns. Use the new function. Move call of check_and_process_move
13909 after operand equiv substitution and address process.
13911 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13914 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13915 with nonlocal goto receivers or returns twice calls, ignore
13916 unininitialized values from abnormal edges to nl goto receiver
13917 or returns twice call.
13919 2013-04-26 Jakub Jelinek <jakub@redhat.com>
13921 PR tree-optimization/57051
13922 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13923 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13926 2013-04-26 Richard Biener <rguenther@suse.de>
13928 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13929 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13930 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13931 (fixup_loop_arrays_after_move): New function.
13932 (move_sese_region_to_fn): Properly outline the loop tree parts
13933 of the SESE region.
13935 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
13937 * config/i386/i386.md (type, unit): Fix long lines.
13939 2013-04-26 Richard Biener <rguenther@suse.de>
13941 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13942 (lto-streamer-out.o): Likewise.
13943 * cfgloop.c (init_loops_structure): Export, add struct function
13944 argument and adjust.
13945 (flow_loops_find): Adjust.
13946 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13947 (init_loops_structure): Declare.
13948 * lto-streamer-in.c: Include cfgloop.h.
13949 (input_cfg): Input the loop tree.
13950 * lto-streamer-out.c: Include cfgloop.h.
13951 (output_cfg): Output the loop tree.
13952 (output_struct_function_base): Do not drop PROP_loops.
13954 2013-03-26 Richard Biener <rguenther@suse.de>
13956 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13957 (pass_build_cfg): Provide PROP_loops.
13958 (move_sese_region_to_fn): Remove loops that are outlined into fn
13960 * tree-inline.c: Include cfgloop.h.
13961 (initialize_cfun): Do not drop PROP_loops.
13962 (copy_loops): New function.
13963 (copy_cfg_body): Copy loop structure.
13964 (tree_function_versioning): Initialize destination loop tree.
13965 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13966 (pass_parallelize_loops): Do IL verification.
13967 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13968 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13969 the CFG make sure we fixup loops as well.
13970 * tree-ssa-tail-merge.c: Include cfgloop.h.
13971 (replace_block_by): When merging loop latches mark loops for fixup.
13972 * lto-streamer-out.c (output_struct_function_base): Drop
13973 PROP_loops for now.
13974 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13975 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13976 * ipa-split.c: Include cfgloop.h.
13977 (split_function): Add the new return block to the loop tree root.
13978 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13979 whether we have removed the forwarder block.
13980 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13981 * cfgloop.h (place_new_loop): Declare.
13982 * cfgloopmanip.c (place_new_loop): Export.
13983 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13984 (tree-switch-conversion.o): Likewise.
13985 (tree-complex.o): Likewise.
13986 (tree-inline.o): Likewise.
13987 (tree-ssa-tailmerge.o): Likewise.
13988 (ipa-split.o): Likewise.
13989 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13990 (tree-ssa-copy.o): Likewise.
13991 * tree-switch-conversion.c: Include cfgloop.h
13992 (process_switch): If we emit a bit-test cascade, schedule loops
13994 * tree-complex.c: Include cfgloop.h.
13995 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13996 * asan.c: Include cfgloop.h.
13997 (create_cond_insert_point): Properly add new basic-blocks to
13998 loops, schedule loop fixup.
13999 * cfgloop.c (verify_loop_structure): Check that looks are not
14001 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
14003 (expand_omp_for_generic): Likewise.
14004 (expand_omp_sections): Likewise.
14005 (expand_omp_atomic_pipeline): Schedule loops for fixup.
14006 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
14007 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
14008 is initialized, not when loops are present.
14009 * tree-parloops.c (parallelize_loops): Remove checking here.
14010 * passes.c (init_optimization_passes): Schedule a copy-propagation
14011 pass before complete unrolling of inner loops.
14013 2013-04-26 Jakub Jelinek <jakub@redhat.com>
14015 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
14016 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
14017 (colorize_init): Add argument to _WIN32 version.
14018 * toplev.c: Include diagnostic-color.h.
14019 (process_options): Default to -fdiagnostics-color=auto if
14020 GCC_COLORS env var is in the environment.
14021 * common.opt (fdiagnostics-color=): Add Var and Init.
14022 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
14023 env var is in the environment, the default is auto rather than never.
14025 * diagnostic.h (file_name_as_prefix): Add context argument.
14026 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
14027 the string as locus.
14028 * langhooks.c (lhd_print_error_function): Adjust caller.
14030 2013-04-25 Lawrence Crowl <crowl@google.com>
14032 * var-tracking.c (shared_hash_def::htab):
14033 Change type to hash_table. Update dependent calls and types.
14035 2013-04-25 Lawrence Crowl <crowl@google.com>
14037 * Makefile.in: Update as needed below.
14039 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
14040 Move declaration to after the type's method definitons.
14042 * attribs.c (htab_t scoped_attributes::attribute_hash):
14043 Change type to hash_table. Update dependent calls and types.
14045 * bitmap.c (htab_t bitmap_desc_hash):
14046 Change type to hash_table. Update dependent calls and types.
14048 * cselib.c (htab_t cselib_hash_table):
14049 Change type to hash_table. Update dependent calls and types.
14051 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
14052 (hash_string_slot_node): Move implementation into lto-streamer.h
14053 struct string_slot_hasher.
14054 (eq_string_slot_node): Likewise.
14056 * data-streamer-out.c: Update output_block::string_hash_table
14057 dependent calls and types.
14059 * dwarf2cfi.c (htab_t trace_index):
14060 Change type to hash_table. Update dependent calls and types.
14062 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
14063 Change type to hash_table. Update dependent calls and types.
14064 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
14065 (htab_t optimize_external_refs::map): Likewise.
14066 (htab_t output_comp_unit::extern_map): Likewise.
14067 (htab_t output_comdat_type_unit::extern_map): Likewise.
14068 (htab_t output_macinfo::macinfo_htab): Likewise.
14069 (htab_t optimize_location_lists::htab): Likewise.
14070 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
14072 * except.c (htab_t ehspec_hash_type):
14073 Change type to hash_table. Update dependent calls and types.
14074 (assign_filter_values::ttypes): Likewise.
14075 (assign_filter_values::ehspec): Likewise.
14076 (sjlj_assign_call_site_values::ar_hash): Likewise.
14077 (convert_to_eh_region_ranges::ar_hash): Likewise.
14079 * gcse.c (htab_t pre_ldst_table):
14080 Change type to hash_table. Update dependent calls and types.
14082 * ggc-common.c (htab_t saving_htab):
14083 Change type to hash_table. Update dependent calls and types.
14084 (htab_t loc_hash): Likewise.
14085 (htab_t ptr_hash): Likewise.
14086 (call_count): Rename ggc_call_count.
14087 (call_alloc): Rename ggc_call_alloc.
14088 (loc_descriptor): Rename make_loc_descriptor.
14089 (add_statistics): Rename ggc_add_statistics.
14091 * ggc-common.c (saving_htab):
14092 Change type to hash_table. Update dependent calls and types.
14094 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
14095 (push_gimplify_context): Likewise.
14096 (pop_gimplify_context): Likewise.
14097 (struct gimple_temp_hash_elt): Added.
14098 (struct gimplify_hasher): Likewise.
14099 (struct gimplify_ctx.temp_htab):
14100 Change type to hash_table. Update dependent calls and types.
14102 * gimple-fold.c: Include gimplify-ctx.h.
14104 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
14105 Change type to hash_table. Update dependent calls and types.
14106 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
14107 avoid potential global name collision.
14109 * gimplify.c: Include gimplify-ctx.h.
14110 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
14111 (htab_t gimplify_ctx::temp_htab):
14112 Update dependent calls and types for new type hash_table.
14113 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
14114 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
14116 * gimplify-ctx.h: New.
14117 (struct gimple_temp_hash_elt): Move from gimplify.c.
14118 (class gimplify_hasher): New.
14119 (struct gimplify_ctx): Move from gimple.h.
14120 (htab_t gimplify_ctx::temp_htab):
14121 Change type to hash_table. Update dependent calls and types.
14123 * graphite-clast-to-gimple.c: Include graphite-htab.h.
14124 (htab_t ivs_params::newivs_index):
14125 Change type to hash_table. Update dependent calls and types.
14126 (htab_t ivs_params::params_index): Likewise.
14127 (htab_t print_generated_program::params_index): Likewise.
14128 (htab_t gloog::newivs_index): Likewise.
14129 (htab_t gloog::params_index): Likewise.
14131 * graphite.c: Include graphite-htab.h.
14132 4htab_t graphite_transform_loops::bb_pbb_mapping):
14133 Change type to hash_table. Update dependent calls and types.
14135 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
14136 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
14137 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
14139 * graphite-dependences.c: Include graphite-htab.h.
14140 (loop_is_parallel_p): Change hash table type of parameter.
14142 * graphite-htab.h: New.
14143 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
14144 (extern find_pbb_via_hash): Move from graphite-poly.h.
14145 (extern loop_is_parallel_p): Move from graphite-poly.h.
14146 (extern get_loop_body_pbbs): Move from graphite-poly.h.
14148 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
14149 (extern loop_is_parallel_p): Move to graphite-htab.h.
14150 (extern get_loop_body_pbbs): Move to graphite-htab.h.
14152 * haifa-sched.c (htab_t delay_htab):
14153 Change type to hash_table. Update dependent calls and types.
14154 (htab_t delay_htab_i2): Likewise.
14156 * ira-color.c (htab_t allocno_hard_regs_htab):
14157 Change type to hash_table. Update dependent calls and types.
14159 * ira-costs.c (htab_t cost_classes_htab):
14160 Change type to hash_table. Update dependent calls and types.
14162 * loop-invariant.c (htab_t merge_identical_invariants::eq):
14163 Change type to hash_table. Update dependent calls and types.
14165 * loop-iv.c (htab_t bivs):
14166 Change type to hash_table. Update dependent calls and types.
14168 * loop-unroll.c (htab_t opt_info::insns_to_split):
14169 Change type to hash_table. Update dependent calls and types.
14170 (htab_t opt_info::insns_with_var_to_expand): Likewise.
14172 * lto-streamer.h (struct string_slot): Move from data-streamer.h
14173 (struct string_slot_hasher): New.
14174 (htab_t output_block::string_hash_table):
14175 Change type to hash_table. Update dependent calls and types.
14177 * lto-streamer-in.c (freeing_string_slot_hasher): New.
14178 (htab_t file_name_hash_table):
14179 Change type to hash_table. Update dependent calls and types.
14181 * lto-streamer-out.c: Update output_block::string_hash_table dependent
14184 * lto-streamer.c (htab_t tree_htab):
14185 Change type to hash_table. Update dependent calls and types.
14187 * omp-low.c: Include gimplify-ctx.h.
14189 * passes.c (htab_t name_to_pass_map):
14190 Change type to hash_table. Update dependent calls and types.
14191 (pass_traverse): Rename to passes_pass_traverse.
14193 * plugin.c (htab_t event_tab):
14194 Change type to hash_table. Update dependent calls and types.
14196 * postreload-gcse.c (htab_t expr_table):
14197 Change type to hash_table. Update dependent calls and types.
14198 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
14200 * sese.c (debug_rename_map_1): Make extern.
14201 (htab_t copy_bb_and_scalar_dependences::rename_map):
14202 Change type to hash_table. Update dependent calls and types.
14204 * sese.h (extern debug_rename_map): Move to .c file.
14206 * store-motion.c (htab_t store_motion_mems_table):
14207 Change type to hash_table. Update dependent calls and types.
14209 * trans-mem.c (htab_t tm_new_mem_hash):
14210 Change type to hash_table. Update dependent calls and types.
14212 * tree-browser.c (htab_t TB_up_ht):
14213 Change type to hash_table. Update dependent calls and types.
14215 * tree-cfg.c (htab_t discriminator_per_locus):
14216 Change type to hash_table. Update dependent calls and types.
14218 * tree-complex.c: Include tree-hasher.h
14219 (htab_t complex_variable_components):
14220 Change type to hash_table. Update dependent calls and types.
14222 * tree-eh.c (htab_t finally_tree):
14223 Change type to hash_table. Update dependent calls and types.
14225 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
14226 struct int_tree_hasher.
14227 (extern int_tree_map_eq): Likewise.
14228 (uid_decl_map_hash): Removed.
14229 (extern decl_tree_map_eq): Likewise.
14231 * tree-hasher.h: New.
14232 (struct int_tree_hasher): New.
14233 (typedef int_tree_htab_type): New.
14235 * tree-inline.c: Include gimplify-ctx.h.
14237 * tree-mudflap.c: Include gimplify-ctx.h.
14239 * tree-parloops.c: Include tree-hasher.h.
14240 (htab_t eliminate_local_variables_stmt::decl_address):
14241 Change type to hash_table. Update dependent calls and types.
14242 (htab_t separate_decls_in_region::decl_copies): Likewise.
14244 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
14245 Change type to hash_table. Update dependent calls and types.
14247 * tree-sra.c (candidates):
14248 Change type to hash_table. Update dependent calls and types.
14250 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
14252 (int_tree_map_hash): Likewise.
14254 * tree-ssa-dom.c (htab_t avail_exprs):
14255 Change type to hash_table. Update dependent calls and types.
14257 * tree-ssa-live.c (var_map_base_init::tree_to_index):
14258 Change type to hash_table. Update dependent calls and types.
14260 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
14261 Change type to hash_table. Update dependent calls and types.
14263 * tree-ssa-phiopt.c (seen_ssa_names):
14264 Change type to hash_table. Update dependent calls and types.
14266 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
14267 Change type to hash_table. Update dependent calls and types.
14269 * tree-ssa-uncprop.c (equiv):
14270 Change type to hash_table. Update dependent calls and types.
14272 2013-04-25 Jakub Jelinek <jakub@redhat.com>
14274 PR rtl-optimization/57003
14275 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
14276 call note_stores with kill_clobbered_value callback again after
14277 killing regs_invalidated_by_call.
14279 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14281 * config/aarch64/aarch64-simd.md
14282 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
14283 (aarch64_simd_bsl<mode>): Likewise.
14284 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
14286 2013-04-25 Marek Polacek <polacek@redhat.com>
14288 PR tree-optimization/57066
14289 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
14291 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
14293 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
14295 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14297 * config/aarch64/aarch64-builtins.c
14298 (aarch64_fold_builtin): New.
14299 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
14300 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
14301 * config/aarch64/aarch64-simd-builtins.def (abs): New.
14302 * config/aarch64/arm_neon.h
14303 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
14305 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14306 Tejas Belagod <tejas.belagod@arm.com>
14308 * config/aarch64/aarch64-builtins.c
14309 (aarch64_gimple_fold_builtin): New.
14310 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
14311 * config/aarch64/aarch64-simd-builtins.def (addv): New.
14312 * config/aarch64/aarch64-simd.md (addpv4sf): New.
14313 (addvv4sf): Update.
14314 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
14316 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14318 * config/aarch64/aarch64.md
14319 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
14321 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14323 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
14324 (*ngcsi_uxtw): New pattern.
14326 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14327 Julian Brown <julian@codesourcery.com>
14329 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
14330 (TB_DREG): Add T_V4HF.
14331 (v4hf_UP): New macro.
14332 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14333 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14334 Handle initialisation of V4HF. Adjust initialisation of reinterpret
14336 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
14337 (arm_vector_mode_supported_p): Handle V4HF.
14338 (arm_mangle_map): Handle V4HFmode.
14339 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
14340 * config/arm/arm_neon_builtins.def: Add entries for
14341 vcvtv4hfv4sf, vcvtv4sfv4hf.
14342 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
14343 (neon_vcvtv4hfv4sf): Likewise.
14344 * config/arm/neon-gen.ml: Handle half-precision floating point
14346 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
14347 * config/arm/arm_neon.h: Regenerate.
14348 * config/arm/neon.ml (type elts): Add F16.
14349 (type vectype): Add T_float16x4, T_floatHF.
14350 (type vecmode): Add V4HF.
14351 (type features): Add Requires_FP_bit feature.
14352 (elt_width): Handle F16.
14353 (elt_class): Likewise.
14354 (elt_of_class_width): Likewise.
14355 (mode_of_elt): Refactor.
14356 (type_for_elt): Handle F16, fix error messages.
14357 (vectype_size): Handle T_float16x4.
14358 (vcvt_sh): New function.
14359 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
14360 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
14361 (string_of_mode): Handle V4HF.
14362 * doc/arm-neon-intrinsics.texi: Regenerate.
14364 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14366 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
14367 format specifier in 'X' case.
14369 2013-04-25 Alan Modra <amodra@gmail.com>
14372 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
14373 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
14374 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
14375 Repeat for many other rotate/shift and mask patterns using subregs.
14376 Name lshiftrt insns.
14377 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
14378 on WORDS_BIG_ENDIAN.
14380 2013-04-25 Alan Modra <amodra@gmail.com>
14382 * config.gcc: Support little-endian powerpc-linux targets.
14383 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
14384 (LINK_OS_LINUX_SPEC): Define.
14385 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
14386 Preserve MASK_LITTLE_ENDIAN.
14387 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
14388 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
14389 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
14390 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
14391 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
14392 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
14393 Correct fp word order for little-endian. Don't shift toc entries
14394 smaller than a word for little-endian.
14395 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
14396 (bswapdi2 splits): Correct low-part subreg for little-endian.
14397 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
14398 low/high where such is correct only for be.
14399 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
14400 little-endian for -mcall-aixdesc.
14402 2013-04-25 Alan Modra <amodra@gmail.com>
14404 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
14405 replace_equiv_address_nv.
14407 2013-04-25 Alan Modra <amodra@gmail.com>
14409 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
14411 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14414 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14415 * rtl.h (struct rtx_def): ...
14417 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14419 PR rtl-optimizations/57046
14420 * lra-constraints (split_reg): Set up lra_risky_transformations_p
14421 for multi-reg splits.
14423 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
14425 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
14427 2013-04-24 Sterling Augustine <saugustine@google.com>
14429 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
14430 (comp_dir_string, debug_str_dwo_section): New.
14431 (DEBUG_STR_DWO_SECTION): Rename to ...
14432 (DEBUG_DWO_STR_SECTION): ... this.
14433 (DEBUG_NORM_STR_SECTION): Delete.
14434 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
14435 (DEBUG_STR_DWO_SECTION_FLAGS): New.
14436 (find_AT_string): Move most logic to ...
14437 (find_AT_string_in_table): ... here. New.
14438 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
14439 add_skeleton_AT_string. Delete logic.
14440 (output_skeleton_debug_sections): Remove call to
14441 add_top_level_skeleton_die_attrs.
14442 (add_comp_dir_attribute): Move logic to comp_dir_string.
14443 (dwarf2out_init): Initialize debug_str_dwo_section.
14444 (output_indirect_string): Call find_string_form.
14445 (output_indirect_strings): Rewrite.
14446 (prune_unused_types): Empty skeleton_debug_str_hash.
14447 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
14448 (dwarf2out_finish): Call output_indirect_strings.
14450 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14452 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
14454 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14456 * rtl.h (struct rtx_def): Add comment for field jump.
14457 (LRA_SUBREG_P): New macro.
14458 * recog.c (register_operand): Check LRA_SUBREG_P.
14459 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14461 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14462 (lra_final_code_change): Skip subreg change for operators.
14463 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14464 if there are no operand changes.
14465 * lra-constraints.c (curr_insn_set): New.
14466 (match_reload): Set LRA_SUBREG_P.
14467 (emit_spill_move): Ditto.
14468 (check_and_process_move): Use curr_insn_set. Process only single
14469 set insns. Don't initialize sec_mem_p and change_p.
14470 (simplify_operand_subreg): Use LRA_SUBREG_P.
14471 (reg_in_class_p): New function.
14472 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14473 of #ifdef. Add code to remove cycling.
14474 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14475 non-null disp. Reload inner instead of disp when base and index
14476 are null. Try to put lo_sum into register.
14477 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
14478 (check_and_process_move): Move code for move cost check to
14479 simple_move_p. Remove equiv_substitution.
14480 (simple_move_p): New function.
14481 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14482 curr_insn_set. Call check_and_process_move only for single set
14483 insns. Use the new function. Move call of check_and_process_move
14484 after operand equiv substitution and address process.
14486 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14488 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14489 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14490 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14492 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14494 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14496 2013-04-24 Marek Polacek <polacek@redhat.com>
14498 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14499 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14500 (select_loops_exit_conditions): Likewise.
14501 (number_of_iterations_for_all_loops): Likewise.
14502 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14503 (scev_analysis): Likewise.
14505 2013-04-02 Catherine Moore <clm@codesourcery.com>
14506 Chao-ying Fu <fu@mips.com>
14508 * config/mips/micromips.md (jraddiusp): New pattern.
14509 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14510 instruction if possible.
14512 2013-04-24 Alan Modra <amodra@gmail.com>
14514 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14516 2013-04-24 Julian Brown <julian@codesourcery.com>
14517 Chung-Lin Tang <cltang@codesourcery.com>
14519 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
14520 dependency behavior in enumeration type DIE generation. Add TODO note
14521 to comments about future DW_FORM_sdata/udata re-work of related code.
14523 2013-04-23 Lawrence Crowl <crowl@google.com>
14525 * Makefile.in: Update as needed below.
14527 * hash-table.h (class hash_table):
14528 Correct many methods with parameter types compare_type to the correct
14529 value_type. (Correct code was unlikely to notice the change.)
14530 (hash_table::elements_with_deleted) New.
14531 (class hashtable::iterator): New.
14532 (hashtable::begin()): New.
14533 (hashtable::end()): New.
14534 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14536 * statistics.c (statistics_hashes):
14537 Change type to hash_table. Update dependent calls and types.
14539 * tree-into-ssa.c (var_infos):
14540 Change type to hash_table. Update dependent calls and types.
14542 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14543 Change type to hash_table. Update dependent calls and types.
14545 * tree-ssa-loop-im.c (struct mem_ref.refs):
14546 Change type to hash_table. Update dependent calls and types.
14548 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14549 Change type to hash_table. Update dependent calls and types.
14551 * tree-ssa-sccvn.c (vn_tables_s::nary):
14552 Change type to hash_table. Update dependent calls and types.
14553 (vn_tables_s::phis): Likewise.
14554 (vn_tables_s::references): Likewise.
14556 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14557 (vn_reference_eq): Update parameter and return types.
14559 * tree-ssa-structalias.c (pointer_equiv_class_table):
14560 Change type to hash_table. Update dependent calls and types.
14561 (location_equiv_class_table): Likewise.
14563 * tree-vect-data-refs.c: Consequential changes for making
14564 peeling a hash_table.
14566 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14567 (destroy_loop_vec_info): Dependent hash_table update.
14569 * tree-vectorizer.h (peeling_htab):
14570 Change type to hash_table. Update dependent calls and types.
14572 2013-04-23 Shiva Chen <shiva0217@gmail.com>
14574 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14575 to check the register content is equal or not.
14576 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14577 to assign register content record.
14578 * lra-eliminations.c (update_reg_eliminate): Use
14579 lra_update_reg_val_offset to update register content offset.
14580 * lra-int.h (struct lra_reg): Add offset member.
14581 (lra_reg_val_equal_p): New static inline function.
14582 (lra_update_reg_val_offset): New static inline function.
14583 (lra_assign_reg_val): New static inline function.
14584 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14585 to assign register content record.
14586 (initialize_lra_reg_info_element): Initial offset to zero.
14588 2013-04-23 Catherine Moore <clm@codesourcery.com>
14590 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14591 operands. Record compression.
14593 2013-04-23 Xinliang David Li <davidxl@google.com>
14595 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14597 2013-04-23 Richard Biener <rguenther@suse.de>
14599 PR middle-end/57036
14600 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14601 parameter, only add abnormal goto edges from the copied body
14602 if the call could perform abnormal gotos.
14603 (copy_cfg_body): Adjust.
14605 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14607 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14609 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14611 * coretypes.h (gimple_stmt_iterator): Add struct to make
14614 2013-04-23 Richard Biener <rguenther@suse.de>
14616 PR tree-optimization/57026
14617 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14618 from SSA names occuring in abnormal PHI nodes.
14620 2013-04-22 Andi Kleen <ak@linux.intel.com>
14622 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14624 2013-04-22 Andi Kleen <ak@linux.intel.com>
14626 * lto/lto.c (print_lto_report_1): Declare early.
14627 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14629 2013-04-22 Andi Kleen <ak@linux.intel.com>
14631 * common.opt (-flto-report-wpa): Add.
14632 * doc/invoke.texi (-flto-report-wpa): Add.
14633 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14636 2013-04-22 Xinliang David Li <davidxl@google.com>
14638 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14639 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14640 * Makefile.in: New dependency
14642 David Daney <ddaney.cavm@gmail.com>
14644 * configure.ac (gcc_cv_as_micromips_support): Use the
14645 --fatal-warnings option.
14646 * configure: Regenerate.
14648 2013-04-22 Marek Polacek <polacek@redhat.com>
14651 * tsan.c (instrument_expr): Don't instrument expression
14652 in case its size is zero.
14654 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
14658 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14660 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14662 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14664 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14665 (gimple_stmt_iterator): New typedef.
14666 * gimple.h (gimple_stmt_iterator): Rename to...
14667 (gimple_stmt_iterator_d): ... This.
14668 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14669 trees be valid for GIMPLE and GENERIC.
14670 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14671 * gimple-fold.c (gimple_fold_call): Call target hook
14672 gimple_fold_builtin.
14673 * hooks.c (hook_bool_gsiptr_false): New.
14674 * hooks.h (hook_bool_gsiptr_false): New.
14675 * target.def (fold_stmt): New.
14676 * doc/tm.texi: Regenerate.
14678 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14681 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14682 a set sp if no stack realignment.
14684 2013-04-22 Nick Clifton <nickc@redhat.com>
14686 * config.gcc (tilegx-linux): Extend extra_objs rather than
14688 (tilepro-linux): Likewise.
14690 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14692 * config/aarch64/aarch64-builtins.c
14694 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14695 (VAR<1-12>): Add MAP parameter.
14696 (BUILTIN_*): Likewise.
14697 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14698 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14699 (aarch64_ushl_n<mode>): Likewise.
14700 (aarch64_sshr_n<mode>): Likewise.
14701 (aarch64_ushr_n<mode>): Likewise.
14702 (aarch64_<maxmin><mode>): Likewise.
14703 (aarch64_sqrt<mode>): Likewise.
14704 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14705 (vshr<q>_n_*): Likewise.
14707 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14709 * config/aarch64/aarch64-builtins.c
14710 (aarch64_simd_builtin_type_mode): Handle SF types.
14712 (BUILTIN_GPF): Define.
14713 (aarch64_init_simd_builtins): Handle SF types.
14714 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14715 (frecps): Likewise.
14716 (frecpx): Likewise.
14717 * config/aarch64/aarch64-simd.md
14718 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14719 (aarch64_frecpe<mode>): New.
14720 (aarch64_frecps<mode>): Likewise.
14721 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14722 (v8type): Add frecp<esx>.
14723 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14724 (aarch64_frecps<mode>): Likewise.
14725 * config/aarch64/iterators.md (FRECP): New.
14726 (frecp_suffix): Likewise.
14727 * config/aarch64/arm_neon.h
14728 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14730 2013-04-22 Christian Bruel <christian.bruel@st.com>
14733 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14734 (REG_CLASS_NAMES): Idem.
14735 (REG_CLASS_CONTENTS): Idem.
14736 (REGCLASS_HAS_FP_REG): Idem.
14737 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14738 (sh_conditional_register_usage): Idem.
14740 2013-04-21 Jeff Law <law@redhat.com>
14742 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14743 (ssa_forward_propagate_and_combine): Use it.
14745 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14747 * lra.c: Update the flow chart diagram.
14749 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14751 PR rtl-optimization/56847
14752 * lra-constraints.c (process_alt_operands): Discourage alternative
14753 with non-matche doffsettable memory constraint fro memory with
14756 2013-04-19 Richard Biener <rguenther@suse.de>
14758 PR tree-optimization/56982
14759 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14761 * gimplify.c (gimplify_call_expr): Notice special calls.
14762 (gimplify_modify_expr): Likewise.
14763 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14764 abnormal control flow receivers.
14765 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14766 in the same way as cfun->has_nonlocal_labels.
14767 (gimple_purge_dead_abnormal_call_edges): Likewise.
14768 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14769 receivers start a basic-block.
14771 2013-04-19 Richard Biener <rguenther@suse.de>
14773 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14775 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14776 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14777 (SLP_TREE_LOAD_PERMUTATION): Add.
14778 (vect_transform_slp_perm_load): Adjust prototype.
14779 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14780 (vect_free_slp_instance): Likewise.
14781 (vect_create_new_slp_node): Likewise.
14782 (vect_supported_slp_permutation_p): Remove.
14783 (vect_slp_rearrange_stmts): Adjust.
14784 (vect_supported_load_permutation_p): Likewise. Inline
14785 vect_supported_slp_permutation_p here.
14786 (vect_analyze_slp_instance): Compute load permutations per
14787 slp node instead of per instance.
14788 (vect_get_slp_defs): Adjust.
14789 (vect_transform_slp_perm_load): Likewise.
14790 (vect_schedule_slp_instance): Remove redundant code.
14791 (vect_schedule_slp): Remove hack for PR56270, add it ...
14792 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14793 CSE loads for SLP. Adjust.
14795 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14797 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14798 spelling in two comments.
14800 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14803 * config/arm/arm.c (load_multiple_sequence): Require SP
14804 as base register for loads if SP is in the register list.
14806 2013-04-19 Martin Jambor <mjambor@suse.cz>
14808 PR tree-optimization/56718
14809 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14810 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14811 and made public. Adjusted all callers.
14812 (ipa_intraprocedural_devirtualization): New function.
14813 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14814 (ipa_intraprocedural_devirtualization): Likewise.
14815 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14817 2013-04-19 Richard Biener <rguenther@suse.de>
14819 PR tree-optimization/57000
14820 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14822 2013-04-19 Terry Guo <terry.guo@arm.com>
14824 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14826 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14827 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14828 (cortex_m4_fmacs): Use new reservations.
14829 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14831 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14833 PR rtl-optimization/56999
14834 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14836 (lra_coalesce): Remove split_origin_bitmap and related code.
14837 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14838 ranges if necessary.
14840 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
14842 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14844 (ix86_expand_call): Remove clobbered_registers array and use
14845 x86_64_ms_sysv_extra_clobbered_registers instead.
14846 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14848 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14850 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14851 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14852 (*call_value_rex64_ms_sysv): Ditto.
14854 2013-04-18 Cary Coutant <ccoutant@google.com>
14856 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14857 parent instead of die_mark.
14859 2013-04-18 Diego Novillo <dnovillo@google.com>
14861 * gimple.c (create_gimple_tmp): New.
14862 (get_expr_type): New.
14863 (build_assign): New.
14864 (build_type_cast): New.
14865 * gimple.h (enum ssa_mode): Define.
14866 (gimple_seq_set_location): New.
14867 * asan.c (build_check_stmt): Change some gimple_build_* calls
14868 to use build_assign and build_type_cast.
14870 2013-04-18 Richard Biener <rguenther@suse.de>
14872 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14873 handle negative step. Remove redundant checks.
14874 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14875 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14876 for negative step and grouped loads fail to vectorize.
14878 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14880 * emit-rtl.c (reset_insn_used_flags): New function.
14881 (reset_all_used_flags): Use it.
14882 (verify_insn_sharing): New function.
14883 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14885 2013-04-18 Jakub Jelinek <jakub@redhat.com>
14887 PR tree-optimization/56984
14888 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14889 and (x >> M) >= N don't register any assertion if N << M is the
14892 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14894 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14895 do not delete_insn insns that have not yet been emitted, only
14896 unlink them with remove_insn.
14897 * df-scan.c (df_insn_delete): Revert r197492.
14899 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14901 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14902 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14904 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14906 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14907 define_insn_and_split.
14908 (and_scc,ior_scc,negscc): Likewise.
14909 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14911 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
14913 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14914 can benefit from using a sequence of LDRD instructions in epilogue
14915 instead of a single LDM instruction.
14917 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14920 * doc/extend.texi: Fix typo.
14922 2013-04-17 Richard Biener <rguenther@suse.de>
14924 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14925 (vect_build_slp_tree): ... here.
14926 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14927 match. Remove special-casing of mismatched complex loads.
14928 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14929 re-try the match with swapped commutative operands.
14930 (vect_supported_load_permutation_p): Remove special-casing of
14931 mismatched complex loads.
14932 (vect_analyze_slp_instance): Adjust.
14934 2013-04-17 Richard Biener <rguenther@suse.de>
14936 PR rtl-optimization/56921
14937 * cfgloop.h (struct loop): Add simple_loop_desc member.
14938 (struct niter_desc): Mark with GTY(()).
14939 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14940 * loop-iv.c (get_simple_loop_desc): Likewise.
14941 (free_simple_loop_desc): Likewise.
14944 2013-04-16 Richard Biener <rguenther@suse.de>
14946 PR rtl-optimization/56921
14947 * loop-init.c (pass_rtl_move_loop_invariants): Add
14948 TODO_do_not_ggc_collect to todo_flags_finish.
14949 (pass_rtl_unswitch): Same.
14950 (pass_rtl_unroll_and_peel_loops): Same.
14951 (pass_rtl_doloop): Same.
14953 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14955 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14956 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14957 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14959 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14960 * tree-streamer.c (record_common_node): Adjust reference in comment.
14962 2013-04-17 Terry Guo <terry.guo@arm.com>
14964 * config/arm/cortex-m4.md: Add a new bypass.
14966 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14968 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14970 (*subs_<optab><mode>_multp2): New pattern.
14971 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14972 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14974 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14976 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14977 (*subs_mul_imm_<mode>): New pattern.
14979 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
14982 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14983 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14984 (vsx_movti_32bit): Change j->wa to O->wa.
14986 2013-04-16 Richard Biener <rguenther@suse.de>
14988 PR rtl-optimization/56921
14989 * loop-init.c (pass_rtl_move_loop_invariants): Add
14990 TODO_do_not_ggc_collect to todo_flags_finish.
14991 (pass_rtl_unswitch): Same.
14992 (pass_rtl_unroll_and_peel_loops): Same.
14993 (pass_rtl_doloop): Same.
14995 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
14997 * config/arm/arm.c (emit_multi_reg_push): New declaration
14998 for an existing function.
14999 (arm_emit_strd_push): New function.
15000 (arm_expand_prologue): Used here.
15001 (arm_emit_ldrd_pop): New function.
15002 (arm_expand_epilogue): Used here.
15003 (arm_get_frame_offsets): Update condition.
15004 (arm_emit_multi_reg_pop): Add a special case for load of a single
15005 register with writeback.
15007 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
15009 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
15012 2013-04-16 Richard Biener <rguenther@suse.de>
15014 PR tree-optimization/56756
15015 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
15016 (first_mem_ref_loc): New.
15017 (execute_sm): Place the load temporarily before a previous
15018 access instead of in the latch edge to ensure its SSA dependencies
15019 are defined at points dominating the load.
15021 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
15023 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
15024 correct fix by moving header and footer insn to the footer of
15025 the merged basic block. Clear BB_END of the merged-away block.
15027 PR middle-end/43631
15028 * emit-rtl.c (make_note_raw): New function.
15029 (link_insn_into_chain): New static inline function.
15030 (add_insn): Use it.
15031 (add_insn_before, add_insn_after): Factor insn chain linking code...
15032 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
15033 using link_insn_into_chain.
15034 (note_outside_basic_block_p): New helper function for emit_note_after
15035 and emit_note_before.
15036 (emit_note_after): Use nobb variant of add_insn_after if the note
15037 should not be contained in a basic block.
15038 (emit_note_before): Use nobb variant of add_insn_before if the note
15039 should not be contained in a basic block.
15040 (emit_note_copy): Use make_note_raw.
15041 (emit_note): Likewise.
15042 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
15043 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
15044 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
15045 the moved barrier the tail of the basic block it follows.
15046 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
15048 2013-04-15 Jakub Jelinek <jakub@redhat.com>
15050 PR tree-optimization/56962
15051 * gimple-ssa-strength-reduction.c (record_increment): Only set
15052 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
15053 either rhs1 or rhs2 is equal to c->base_expr.
15055 2013-04-15 Richard Biener <rguenther@suse.de>
15057 PR tree-optimization/56933
15058 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
15060 (GROUP_READ_WRITE_DEPENDENCE): Remove.
15061 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
15062 * tree-vect-data-refs.c (vect_analyze_group_access): Move
15063 dependence check ...
15064 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
15066 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
15067 GROUP_READ_WRITE_DEPENDENCE.
15069 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15071 * emit-rtl.c (reset_all_used_flags): New function.
15072 (verify_rtl_sharing): Call reset_all_used_flags before and after
15073 performing the checks.
15075 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15077 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
15078 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
15079 * config/arm/constraints.md (De): New constraint.
15080 * config/arm/neon.md (anddi3_neon): Delete.
15081 (neon_vand<mode>): Expand to standard anddi3 pattern.
15082 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
15083 Move earlier in the file.
15084 (neon_inv_logic_op2): Likewise.
15085 (arm_anddi_operand_neon): New predicate.
15087 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15089 * configure.ac (gcc_cv_ld_as_needed): Set
15090 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
15091 Use -z ignore, -z record on *-*-solaris2*.
15092 (HAVE_LD_AS_NEEDED): Update comment.
15093 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
15094 * configure: Regenerate.
15095 * config.in: Regenerate.
15096 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
15097 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
15098 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
15099 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
15100 equivalents. Fix markup.
15101 * doc/tm.texi: Regenerate.
15103 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
15105 * config/i386/i386.opt: New option mstack-protector-guard=.
15106 * config/i386/i386-opts.h: Add enum stack_protector_guard.
15107 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
15108 TARGET_SSP_TLS_GUARD.
15109 * config/i386/i386.c (ix86_option_override_internal): Set
15110 ix86_stack_protector_guard.
15111 * config/i386/i386.md (stack_protect_set): Enable for
15112 TARGET_SSP_TLS_GUARD only.
15113 (stack_protect_set_<mode>): Ditto.
15114 (stack_protect_test): Ditto.
15115 (stack_protect_test_<mode>): Ditto.
15116 * doc/invoke.texi (i386 Option): Document.
15118 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
15121 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
15122 (S_MODES): Set H_MODE bit.
15123 (SF_MODES): Set only S_MODE and SF_MODE bits.
15124 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
15125 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
15126 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
15127 <MODE_FLOAT>: Likewise.
15129 2013-04-15 Joey Ye <joey.ye@arm.com>
15131 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
15133 2013-04-15 Joey Ye <joey.ye@arm.com>
15135 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
15137 (thumb_far_jump_used_p): Count instruction size and set
15140 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
15142 * reorg.c (fill_simple_delay_slots): Reindent block of code.
15143 * resource.c (mark_target_live_regs): Reformat conditional block.
15145 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15147 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
15148 notes, they are emitted only just before final.
15149 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
15151 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
15153 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
15154 * cfgrtl.c (delete_insn): Call it here instead.
15155 * lra-spills.c (lra_final_code_change): Use delete_insn.
15156 * haifa-sched.c (sched_remove_insn): Likewise.
15157 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
15158 returning to the nop pool.
15159 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
15160 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
15162 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15164 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
15165 * doc/tm.texi: Regenerated.
15167 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
15169 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
15172 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
15174 * df-core.c (df_find_def): Compare register numbers.
15175 (df_find_use): Likewise.
15177 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
15180 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
15181 lra_in_progress for return.
15183 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15185 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
15186 define_insn into define_insn_and_split and emit movsicc patterns.
15188 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
15190 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
15192 2013-04-12 Richard Biener <rguenther@suse.de>
15194 * tree-pass.h (TODO_do_not_ggc_collect): New.
15195 * passes.c (execute_one_ipa_transform_pass): Honor
15196 TODO_do_not_ggc_collect.
15197 (execute_one_pass): Likewise.
15200 2013-04-10 Richard Biener <rguenther@suse.de>
15202 * passes.c (init_optimization_passes): Remove reload pass.
15203 * ira.c (do_reload): Merge into ...
15205 (rest_of_handle_reload): Remove.
15206 (pass_reload): Likewise.
15207 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15208 of reload for vzeroupper pass placement.
15210 2013-04-12 Jakub Jelinek <jakub@redhat.com>
15212 PR tree-optimization/56918
15213 PR tree-optimization/56920
15214 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
15215 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
15216 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
15217 use wide_mul_with_sign method.
15219 2013-04-12 Richard Biener <rguenther@suse.de>
15221 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
15222 not be considered a gimple constant.
15224 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
15226 * fold-const.c (const_binop): Handle vector shifts by a scalar.
15227 (fold_binary_loc): Call const_binop also for mixed vector-scalar
15230 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
15231 Jakub Jelinek <jakub@redhat.com>
15233 * opts.c: Include diagnostic-color.h.
15234 (common_handle_option): Handle OPT_fdiagnostics_color_.
15235 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
15236 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
15237 (diagnostic-color.o): New.
15238 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
15239 (diagnostic_color_rule): New enum.
15240 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
15241 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
15242 the location string.
15243 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
15244 either NULL, or color kind.
15245 * diagnostic-color.c: New file.
15246 * diagnostic-color.h: New file.
15247 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
15249 * doc/invoke.texi (-fdiagnostics-color): Document.
15250 * pretty-print.h (pp_show_color): Define.
15251 (struct pretty_print_info): Add show_color field.
15252 * diagnostic.c: Include diagnostic-color.h.
15253 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
15254 macros. Colorize error:, warning: etc. strings and also the location
15256 (diagnostic_show_locus): Colorize the caret line.
15257 * pretty-print.c: Include diagnostic-color.h.
15258 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
15259 inside of %< %> quotes or quoted through q format modifier.
15261 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15263 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
15265 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15267 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
15268 code in CC_NZ mode.
15269 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
15272 2013-04-11 Marek Polacek <polacek@redhat.com>
15274 PR tree-optimization/48184
15275 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
15277 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
15279 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
15280 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
15281 (skip_simple_arithmetic): Tidy up.
15282 * tree.h (skip_simple_constant_arithmetic): Declare.
15284 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
15286 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
15288 2013-04-11 Richard Biener <rguenther@suse.de>
15290 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15291 generate vector constants.
15293 2013-04-11 Richard Biener <rguenther@suse.de>
15295 PR tree-optimization/56878
15296 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
15297 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
15299 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
15300 Prefer to align the DR with the most invariant base address.
15302 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15304 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
15307 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
15309 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
15310 floating-point vector comparisons against 0.
15312 2013-04-11 Jakub Jelinek <jakub@redhat.com>
15314 PR tree-optimization/56899
15315 * fold-const.c (extract_muldiv_1): Apply distributive law
15316 only if TYPE_OVERFLOW_WRAPS (ctype).
15318 2013-04-11 Bin Cheng <bin.cheng@arm.com>
15321 * ira-costs.c (scan_one_insn): Check whether the source rtx of
15322 loading has side effect.
15324 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
15326 * config/sparc/sparc.c: Include tree-pass.h.
15327 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
15328 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
15329 head of file. Change return type. Split off gate function.
15330 (sparc_gate_work_around_errata): New function.
15331 (pass_work_around_errata): New pass definition.
15332 (insert_pass_work_around_errata) New pass insert definition to
15333 insert pass_work_around_errata just after delayed-branch scheduling.
15334 (sparc_option_override): Insert the pass.
15335 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
15337 2013-04-10 David S. Miller <davem@davemloft.net>
15339 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
15340 or -mcpu=hypersparc.
15342 * target.def (cstore_mode): New hook.
15343 * target.h: Include insn-codes.h
15344 * targhooks.c: Likewise.
15345 (default_cstore_mode): New function.
15346 * targhooks.h: Declare it.
15347 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
15348 * doc/tm.texi: Rebuild.
15349 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
15350 target hook, rather than inspecting the insn_data.
15351 * config/sparc/sparc.c (sparc_cstore_mode): New function.
15352 (TARGET_CSTORE_MODE): Redefine.
15353 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
15355 * config/sparc/predicates.md (cstore_result_operand): New special
15357 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
15358 Use it for operand 0.
15359 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
15360 (*snesi_special): Likewise.
15361 (*snesi_zero): Likewise.
15362 (*seqsi_zero): Likewise.
15363 (*sltu_insn): Likewise.
15364 (*sgeu_insn): Likewise.
15365 (*seqdi_special): Make operand 0 and comparison operation be of
15367 (*snedi_special): Likewise.
15368 (*snedi_special_vis3): Likewise.
15369 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
15370 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
15371 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
15372 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
15373 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
15374 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
15375 (*sltu_extend_sp64): Likewise.
15376 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
15377 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
15378 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
15379 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
15380 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
15382 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
15384 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
15385 (aarch64_start_file): Use the new function.
15387 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15388 Jason Merrill <jason@redhat.com>
15390 * common.opt: Add -gdwarf.
15391 * opts.c (common_handle_option): Handle it.
15392 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
15394 2013-04-10 Richard Biener <rguenther@suse.de>
15396 * passes.c (execute_todo): Do not call ggc_collect conditional here.
15397 (execute_one_ipa_transform_pass): But unconditionally here.
15398 (execute_one_pass): And here.
15399 (init_optimization_passes): Remove reload pass.
15400 * tree-pass.h (TODO_ggc_collect): Remove.
15401 (pass_reload): Likewise.
15402 * ira.c (do_reload): Merge into ...
15404 (rest_of_handle_reload): Remove.
15405 (pass_reload): Likewise.
15406 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15407 of reload for vzeroupper pass placement.
15408 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
15409 and todo_flags_finish of all passes.
15411 2013-04-10 Richard Biener <rguenther@suse.de>
15413 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
15414 first_const_oprnd field, rename first_def_type to first_op_type.
15415 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
15416 (vect_get_and_check_slp_defs): Always use the type of the
15417 operand. Allow mixed vect_external_def, vect_constant_def types.
15418 (vect_get_constant_vectors): Handle mixed vect_external_def,
15419 vect_constant_def types.
15421 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
15423 PR tree-optimization/55524
15424 * tree-ssa-math-opts.c
15425 (convert_mult_to_fma): Don't use an fms construct
15426 when we don't have an fms operation, but fnma, and it looks
15427 likely that we'll be able to use the latter.
15429 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
15431 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
15433 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
15434 inline fail caused by overwritable functions.
15436 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
15438 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
15439 unnecessary bits in the constant power of two case.
15441 2013-04-10 Richard Biener <rguenther@suse.de>
15443 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
15444 broken code swapping operands.
15445 (vect_build_slp_tree): Do not compute load permutations here.
15446 (vect_analyze_slp_instance): Compute load permutations here,
15447 after building the SLP tree.
15449 2013-04-09 Christian Bruel <christian.bruel@st.com>
15451 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
15452 of next/prev_real_insn.
15454 2013-04-09 Jan Hubicka <jh@suse.cz>
15456 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15457 Drop aliased parameter.
15458 (function_and_variable_visibility): Do not handle alias pairs.
15459 * cgraph.c (varpool_externally_visible_p): Update prototype.
15460 * varpool.c (varpool_add_new_variable): Update.
15462 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15464 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15466 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15468 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15470 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15472 2013-04-09 Marek Polacek <polacek@redhat.com>
15474 PR tree-optimization/48762
15475 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
15477 2013-04-09 Richard Biener <rguenther@suse.de>
15479 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15481 (vect_build_slp_tree): Likewise.
15482 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15483 calculating the cost of a SLP instance.
15484 (vect_analyze_slp_instance): Use it from here, after building
15487 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15489 PR middle-end/56883
15490 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15491 expand_omp_for_static_chunk): Use simple_p = true in
15492 force_gimple_operand_gsi calls when assigning to addressable decls.
15494 2013-04-09 Jeff Law <law@redhat.com>
15496 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15497 when the boolean was created by converting a wider object which
15498 had a boolean range.
15500 2013-04-09 Richard Biener <rguenther@suse.de>
15502 * tree-vectorizer.h (slp_void_p): Remove.
15503 (slp_tree): Typedef before _slp_tree declaration.
15504 (struct _slp_tree): Use a vector of slp_tree as children.
15505 (vect_get_place_in_interleaving_chain): Remove.
15506 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15508 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15510 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15511 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15512 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15513 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15514 Use slp_node instead of slp_void_p and adjust.
15516 2013-04-09 Richard Biener <rguenther@suse.de>
15518 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15519 work that is not necessary.
15521 2013-04-09 Jakub Jelinek <jakub@redhat.com>
15523 PR tree-optimization/56854
15524 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15525 forward into clobber stmts if it would change MEM_REF lhs into
15528 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15530 * tree.c (type_hash_lookup, type_hash_add): Make static.
15531 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15533 2013-04-09 Richard Biener <rguenther@suse.de>
15535 * tree.h (unsave_expr_now): Remove.
15536 * tree-inline.c (mark_local_for_remap_r): Remove.
15537 (unsave_expr_1): Likewise.
15538 (unsave_r): Likewise.
15539 (unsave_expr_now): Likewise.
15540 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15541 (propagate_tree_value): Likewise.
15543 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15545 * doc/rtl.texi (sequence): Rewrite documentation to match the
15546 current use of SEQUENCE rtl objects.
15547 * rtl.def (SEQUENCE): Likewise.
15549 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15550 Update documentation.
15551 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15552 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15554 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15556 2013-04-08 Teresa Johnson <tejohnson@google.com>
15558 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15559 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
15561 (estimate_edge_size_and_time): Add comment to suggest using rounding
15563 (estimate_node_size_and_time): Ditto.
15564 (remap_edge_change_prob): Use helper rounding divide methods.
15565 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15566 (gimple_mod_pow2_value_transform): Ditto.
15567 (gimple_mod_subtract_transform): Ditto.
15568 (gimple_ic_transform): Ditto.
15569 (gimple_stringops_transform): Ditto.
15570 * stmt.c (conditional_probability): Ditto.
15571 (emit_case_dispatch_table): Ditto.
15572 * lto-cgraph.c (merge_profile_summaries): Ditto.
15573 * tree-optimize.c (execute_fixup_cfg): Ditto.
15574 * cfgcleanup.c (try_forward_edges): Ditto.
15575 * cfgloopmanip.c (scale_loop_profile): Ditto.
15577 (duplicate_loop_to_header_edge): Ditto.
15578 (lv_adjust_loop_entry_edge): Ditto.
15579 * tree-vect-loop.c (vect_transform_loop): Ditto.
15580 * profile.c (compute_branch_probabilities): Ditto.
15581 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15582 * lto-streamer-in.c (input_cfg): Ditto.
15583 * gimple-streamer-in.c (input_bb): Ditto.
15584 * ipa-cp.c (update_profiling_info): Ditto.
15585 (update_specialized_profile): Ditto.
15586 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15587 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
15589 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15590 (compute_trg_info): Ditto.
15591 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15592 (purge_dead_edges): Ditto.
15593 * loop-unswitch.c (unswitch_loop): Ditto.
15594 * cgraphclones.c (cgraph_clone_edge): Ditto.
15595 (cgraph_clone_node): Ditto.
15596 * tree-inline.c (copy_bb): Ditto.
15597 (copy_edges_for_bb): Ditto.
15598 (initialize_cfun): Ditto.
15599 (copy_cfg_body): Ditto.
15600 (expand_call_inline): Ditto.
15602 2013-04-08 Kai Tietz <ktietz@redhat.com>
15604 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15605 TARGET_CYGWIN64 by TARGET_64BIT.
15607 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15609 * config/epiphany/epiphany.md (GPR_1): New constant.
15610 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15611 * config/epiphany/epiphany.c (gen_compare_reg):
15612 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15613 is already in place.
15614 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15615 Don't require being called during rtl expansion; If y operlaps r0,
15617 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15618 (epiphany_expand_epilogue): Likewise.
15620 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15621 Don't use CC_FPmode for ORDERED / UNORDERED.
15622 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
15624 * config/epiphany/constraints.md (CnL): New constraint.
15625 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15626 * config/epiphany/predicates.md (add_operand): Allow 1024.
15628 * config/epiphany/epiphany.md (logical_op): New code iterator.
15629 (op_mnc): New code attribute.
15630 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15631 (mov_f+1, mov_f+2): New peephole2 patterns.
15633 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15634 (cstoresi4): Also allow re-use of zero result when doing a NE
15635 comparison to a non-zero operand.
15636 Use (clobber (scratch)) for first insn if the gpr output is not needed.
15638 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15639 Use gen_addsi3_i / gen_subsi3_i.
15641 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15645 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15646 contain anything but clobbers, at most one __builtin_stack_restore,
15647 optionally debug stmts and final resx, and if it has at least one
15648 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15649 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15650 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15651 which isn't defaut definition, remove them.
15652 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15653 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15654 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15655 with MEM_REF LHS with SSA_NAME address.
15657 2013-04-08 Jeff Law <law@redhat.com>
15659 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
15661 2013-04-08 Richard Biener <rguenther@suse.de>
15663 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15665 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15666 determined vector type.
15667 (vect_analyze_data_refs): Likewise.
15668 (vect_get_new_vect_var): Adjust.
15669 (vect_create_destination_var): Preserve SSA name versions.
15670 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15671 not dump anything here.
15673 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15675 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15676 Add member lr_slot_known.
15677 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15679 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15680 Remove code that sets lr_slot_offset according to what a previous
15681 version of epiphany_emit_save_restore used to do.
15682 (epiphany_emit_save_restore): When doing an lr save or restore,
15683 set/verify lr_slot_known and lr_slot_offset.
15685 2013-04-08 Xinyu Qi <xyqi@marvell.com>
15688 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
15691 2013-04-08 Richard Biener <rguenther@suse.de>
15693 * alias.c (find_base_term): Fix thinko in previous change.
15695 2013-04-08 Jakub Jelinek <jakub@redhat.com>
15697 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15698 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15699 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15700 if possible to compute val.
15701 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15702 For QImode integers don't require anything about precision. Use
15703 const_with_all_bytes_same to find out if the constant doesn't have
15704 repeated bytes in it.
15706 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15708 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15711 2013-04-08 Marek Polacek <polacek@redhat.com>
15713 PR rtl-optimization/48182
15714 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15717 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15720 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15721 nuses, make sure we have a label.
15723 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15726 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15727 (rs6000_emit_swdiv_low_precision): Remove.
15728 (rs6000_emit_swdiv): Rewrite to handle between one and four
15729 iterations of Newton-Raphson generally; modify required number of
15730 iterations for some cases.
15731 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15733 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15735 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15736 set-but-unused variable.
15738 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15739 basic blocks of released function bodies garbage-collectable.
15741 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15742 (struct rtl_opt_pass): Add TODO_df_finish.
15744 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15746 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15748 * config/arm/constraints.md (q): New constraint.
15749 * config/arm/ldrdstrd.md: New file.
15750 * config/arm/arm.md (ldrdstrd.md) New include.
15751 (arm_movdi): Use "q" instead of "r" constraint
15752 for double-word memory access.
15753 (movdf_soft_insn): Likewise.
15754 * config/arm/vfp.md (movdi_vfp): Likewise.
15755 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
15756 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
15757 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15758 (mem_ok_for_ldrd_strd): Likewise.
15759 (output_move_double): Update assertion.
15761 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15763 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15765 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15767 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15768 define_insn_and_split.
15769 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15771 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15773 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15774 define_insn_and_split.
15775 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
15776 (shiftsi3_compare): New pattern.
15777 (rrx): New pattern.
15778 * config/arm/unspecs.md (UNSPEC_RRX): New.
15780 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15782 * config/arm/arm.md (negdi_extendsidi): New pattern.
15783 (negdi_zero_extendsidi): Likewise.
15785 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15787 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15788 define_insn_and_split.
15789 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15790 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15792 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15794 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15795 define_insn_and_split.
15796 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15797 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15799 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15801 * config/arm/arm.md (subsi3_carryin): New pattern.
15802 (subsi3_carryin_const): Likewise.
15803 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15804 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15806 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15808 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15810 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15812 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
15813 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
15815 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15817 * config/arm/arm.c (arm_expand_builtin): Change fcode
15818 type to unsigned int.
15820 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15822 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15824 2013-04-04 Ian Lance Taylor <iant@google.com>
15826 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15829 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15831 PR middle-end/56729
15832 * df-scan.c (df_insn_delete): Disable failing assert.
15834 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15836 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15837 New function prototype.
15838 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15839 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15840 (arm_builtin_vectorized_function): New function.
15842 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15844 * config/arm/arm_neon_builtins.def: New file.
15845 * config/arm/arm.c (neon_builtin_data): Move contents to
15846 arm_neon_builtins.def.
15847 (enum arm_builtins): Include neon builtin definitions.
15848 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
15849 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
15851 2013-04-04 Marek Polacek <polacek@redhat.com>
15853 PR tree-optimization/48186
15854 * predict.c (maybe_hot_frequency_p): Return false if
15855 HOT_BB_FREQUENCY_FRACTION is 0.
15856 (cgraph_maybe_hot_edge_p): Likewise.
15858 2013-04-04 Richard Biener <rguenther@suse.de>
15860 PR tree-optimization/56826
15861 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15864 2013-04-04 Richard Biener <rguenther@suse.de>
15866 PR tree-optimization/56213
15867 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
15868 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
15870 2013-04-04 Richard Biener <rguenther@suse.de>
15872 PR tree-optimization/56837
15873 * tree-loop-distribution.c (classify_partition): For non-zero
15874 values require that the value has the same precision as its
15875 mode to be useful as memset value.
15877 2013-04-03 Nick Clifton <nickc@redhat.com>
15879 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
15880 (fmssf4): Use fmsf.s on E3V5 architectures.
15881 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15882 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15884 2013-04-03 Jeff Law <law@redhat.com>
15886 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15887 (lra-eliminations.o): Likewise.
15889 2013-04-03 Teresa Johnson <tejohnson@google.com>
15891 * gcov-io.c (compute_working_sets): Moved most of body of old
15892 compute_working_sets here from profile.c.
15893 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
15894 (gcov_working_set_t): Moved typedef here from basic-block.h
15895 (compute_working_set): Declare.
15896 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15897 (get_working_sets): Renamed from compute_working_set,
15898 replace most of body with call to new compute_working_sets.
15899 (get_exec_counts): Replace call to compute_working_sets
15900 to get_working_sets.
15901 * profile.h (get_working_sets): Renamed from compute_working_set.
15902 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
15903 to get_working_sets.
15904 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15905 * gcov-dump.c (dump_working_sets): New function.
15907 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
15909 * hwint.c (sext_hwi, zext_hwi): New functions.
15910 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15911 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15912 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15913 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15914 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15915 (sext_hwi, zext_hwi): New functions.
15917 2013-04-03 Jeff Law <law@redhat.com>
15919 PR tree-optimization/56799
15920 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15921 back test for widening conversion erroneously dropped in prior change.
15923 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15926 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15927 instead of next_real_insn.
15929 2013-04-03 Marek Polacek <polacek@redhat.com>
15932 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
15934 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15937 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15939 (thumb1_output_casesi): Likewise.
15940 (thumb2_output_casesi): Likewise.
15942 2013-04-03 Richard Biener <rguenther@suse.de>
15944 PR tree-optimization/56817
15945 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15947 (tree_unroll_loops_completely_1): ... new function to manually
15948 walk the loop tree, properly defering outer loops of unrolled
15949 loops to later iterations.
15951 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15953 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15954 (vectorizable_load): Likewise.
15955 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15956 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15958 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15960 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15963 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15965 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15967 2013-04-03 Bin Cheng <bin.cheng@arm.com>
15969 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15971 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15973 PR tree-optimization/56790
15974 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15977 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
15979 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15981 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15982 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15985 2013-04-03 Jakub Jelinek <jakub@redhat.com>
15988 * tree.h (force_folding_builtin_constant_p): New decl.
15989 * builtins.c (force_folding_builtin_constant_p): New variable.
15990 (fold_builtin_constant_p): Fold immediately also if
15991 force_folding_builtin_constant_p.
15993 2013-04-03 Richard Biener <rguenther@suse.de>
15995 PR tree-optimization/56812
15996 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15997 DRs of the same interleaving chain are independent.
15999 2013-04-02 Jason Merrill <jason@redhat.com>
16001 * gdbinit.in (pbb): Use debug fn.
16003 2013-04-02 Lawrence Crowl <crowl@google.com>
16005 * sese.h (struct ivtype_map_elt_s): Remove unused.
16006 (extern debug_ivtype_map): Remove unused.
16007 (extern eq_ivtype_map_elts): Remove unused.
16008 * sese.c (debug_ivtype_map): Removed unused.
16009 (debug_ivtype_map_1): Removed unused.
16010 (debug_ivtype_elt): Remove unused.
16011 (eq_ivtype_map_elts): Remove unused.
16014 2013-04-02 Kai Tietz <ktietz@redhat.com>
16017 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
16018 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
16019 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
16021 (legitimize_pe_coff_symbol): Likewise.
16022 (is_imported_p): New helper-function.
16023 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
16024 for Windows x64 targets.
16025 (ix86_expand_prologue): Optimize for pe-coff targets.
16026 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
16027 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
16028 medium/large code-model.
16029 (legitimize_pic_address): Likewise.
16030 (legitimize_tls_address): Likewise.
16031 (ix86_expand_call): Likewise.
16032 (x86_output_mi_thunk): Likewise.
16033 (get_dllimport_decl): Add new beimport argument.
16034 (construct_plt_address): Don't assert for x64 pe-coff targets.
16035 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
16037 (SYMBOL_FLAG_STUBVAR): New macro.
16038 (SYMBOL_REF_STUBVAR_P): Likewise.
16039 * config/i386/winnt.c (stub_list): New structure.
16040 (stub_head): New local variable.
16041 (i386_pe_record_stub): New function.
16042 (i386_pe_file_end): Emit refptr-stubs.
16044 2013-04-02 Jakub Jelinek <jakub@redhat.com>
16046 PR rtl-optimization/56745
16047 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
16048 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
16051 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
16052 and both of them are MEM_REFs, just compare first argument for
16053 equality and attempt to deal even with differing offsets.
16056 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
16057 of gimple_clobber_p to be MEM_REF.
16058 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
16059 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
16060 after gimplification.
16061 * asan.c (get_mem_ref_of_assignment): Don't instrument
16062 gimple_clobber_p stmts.
16063 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
16064 gimple_clobber_p stmt if they have MEM_REF lhs and
16065 are dead because of another gimple_clobber_p stmt.
16066 * tree-ssa-live.c (clear_unused_block_pointer): Treat
16067 gimple_clobber_p stmts like debug stmts.
16068 (remove_unused_locals): Remove clobbers with MEM_REF lhs
16069 that refer to unused VAR_DECLs or uninitialized values.
16070 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
16071 gimple_clobber_p stmts if they refer to removed parameters.
16072 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
16075 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
16077 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
16078 using SWI48 mode attribute.
16080 2013-04-02 Wei Mi <wmi@google.com>
16082 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
16083 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
16084 *<rotate_insn><mode>3_mask in i386.md.
16086 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
16088 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
16090 2013-04-02 Richard Biener <rguenther@suse.de>
16092 PR tree-optimization/56778
16093 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
16094 Runtime alias tests are not supported for gather loads.
16095 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
16096 stmts referenced from SSA operands before updating SSA form.
16098 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
16099 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16101 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
16102 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
16103 * config/arm/cortex-a53.md: New file.
16104 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
16105 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
16106 * config/arm/arm.c (arm_issue_rate): Likewise.
16107 * config/arm/arm-tune.md: Regenerate
16108 * config/arm/arm-tables.opt: Regenerate.
16109 * config/arm/arm-cores.def: Add cortex-a53.
16111 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
16113 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
16116 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16118 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
16119 scalar load/store operations using B/H registers.
16120 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
16122 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
16124 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
16126 * config/aarch64/aarch64.c
16127 (aarch64_simd_scalar_immediate_valid_for_move): New.
16128 * config/aarch64/aarch64-protos.h
16129 (aarch64_simd_scalar_immediate_valid_for_move): New.
16130 * config/aarch64/constraints.md (Dh, Dq): New.
16131 * config/aarch64/iterators.md (hq): New.
16133 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
16135 * reorg.c (get_branch_condition): Deal with conditional returns.
16136 (fill_simple_delay_slots): Remove dead code dealing with jumps.
16138 2013-04-01 Wei Mi <wmi@google.com>
16140 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
16141 Truncate operand 2 using %b asm operand modifier.
16142 (*<shift_insn><mode>3_mask): Ditto.
16143 (*<rotate_insn><mode>3_mask): Ditto.
16145 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
16147 PR middle-end/56798
16148 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
16150 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
16152 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
16154 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
16156 2013-03-30 Lawrence Crowl <crowl@google.com>
16158 * dse.c (clear_alias_sets): Remove never set.
16159 (disqualified_clear_alias_sets): Remove never set.
16160 (clear_alias_mode_pool): Remove never set.
16161 (dse_step0): Remove condition that is never true.
16162 (canon_address): Remove condition that is never true.
16163 (dse_step7): Remove condition that is never true.
16164 (rest_of_handle_dse): Remove condition that is never true.
16165 (rest_of_handle_dse::did_global): Remove never read from above.
16166 (dse_step2_spill): Remove never called from above.
16167 (dse_step5_spill): Remove never called from above.
16169 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
16171 * doc/md.texi (Standard Names) <casesi>: Update documentation for
16172 JUMP_TABLE_DATA changes.
16173 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
16174 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
16175 (Insns) <jump_table_data>: New entry.
16176 * doc/tm.texi: Regenerate.
16178 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
16180 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
16181 for table jump at the end of a basic block using tablejump_p.
16182 * targhooks.c (default_invalid_within_doloop): Likewise.
16183 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
16184 target hook implementation that is identical to the default hook.
16185 (rs6000_invalid_within_doloop): Remove.
16187 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
16188 unused variable from tablejump_p call.
16190 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
16191 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
16192 (INSN_DELETED_P): Likewise.
16193 (emit_jump_table_data): New prototype.
16194 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
16195 after 4th as unused.
16196 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
16197 * sched-vis.c (print_insn): Likewise.
16198 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
16199 insn for compatibility with back ends that use next_active_insn to
16200 identify jump table data.
16201 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
16202 (remove_insn): Likewise.
16203 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
16205 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
16206 (emit_jump_table_data): New function.
16208 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
16209 basic block, a JUMP_TABLE_DATA never is.
16210 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
16211 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
16212 off from code handling real insns.
16213 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
16214 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
16215 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
16216 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
16217 is not a NONDEBUG_INSN_P.
16218 * ira-costs.c (scan_one_insn): Likewise.
16219 * jump.c (mark_all_labels): Likewise.
16220 (mark_jump_label_1): Likewise.
16221 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16222 * lra.c (get_insn_freq): Expect all insns reaching here to be in
16224 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
16225 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
16226 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
16227 JUMP_TABLE_DATA_P insns.
16228 (calculate_elim_costs_all_insns): Likewise.
16229 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
16230 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
16231 (delete_output_reload): Code style fixups.
16232 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
16233 insn flags on this non-insn.
16234 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
16235 as scheduling barriers, for pre-change compatibility.
16236 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
16237 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
16239 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
16240 redundant JUMP_TABLE_DATA_P test.
16241 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
16242 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16243 (frv_for_each_packet): Likewise.
16244 * config/i386/i386.c (min_insn_size): Likewise.
16245 (ix86_avoid_jump_mispredicts): Likewise.
16246 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16247 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
16248 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16249 (mips16_insn_length): Robustify.
16250 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
16251 (mips16_split_long_branches): Likewise.
16252 * config/pa/pa.c (pa_combine_instructions): Likewise.
16253 * config/rs6000/rs6000.c (get_next_active_insn): Treat
16254 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
16255 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
16256 as contributing to pool range lengths.
16257 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
16258 Remove redundant JUMP_TABLE_DATA_P test.
16259 (sh_loop_align): Likewise.
16260 (split_branches): Likewise.
16261 (sh_insn_length_adjustment): Likewise.
16262 * config/spu/spu.c (get_branch_target): Likewise.
16264 2013-03-29 Jan Hubicka <jh@suse.cz>
16266 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
16267 gcov streaming; stream hot bb threshold to ltrans.
16268 * predict.c (get_hot_bb_threshold): Break out from ....
16269 (maybe_hot_count_p): ... here.
16270 (set_hot_bb_threshold): New function.
16271 * lto-section-in.c (lto_section_name): Add profile.
16272 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
16273 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
16274 and data-streamer.h
16275 (histogram_entry): New structure.
16276 (histogram, histogram_pool): New global vars.
16277 (histogram_hash): New structure.
16278 (histogram_hash::hash): New method.
16279 (histogram_hash::equal): Likewise.
16280 (account_time_size): New function.
16281 (cmp_counts): New function.
16282 (dump_histogram): New function.
16283 (ipa_profile_generate_summary): New function.
16284 (ipa_profile_write_summary): New function.
16285 (ipa_profile_read_summary): New function.
16286 (ipa_profile): Decide on threshold.
16287 (pass_ipa_profile): Add ipa_profile_write_summary and
16288 ipa_profile_read_summary.
16289 * Makefile.in (ipa.o): Update dependencies.
16290 * lto-streamer.h (LTO_section_ipa_profile): New section.
16292 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
16294 * tree.h (VAR_P): New.
16296 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
16299 * doc/invoke.texi ([-fwhole-program]): Fix typo.
16301 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
16303 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
16304 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16305 (control_flow_insn_p): Likewise.
16306 * cfgrtl.c (duplicate_insn_chain): Likewise.
16307 * final.c (get_attr_length_1): Likewise.
16308 (shorten_branches): Likewise.
16309 (final_scan_insn): Likewise.
16310 * function.c (instantiate_virtual_regs): Likewise.
16311 * gcse.c (insert_insn_end_basic_block): Likewise.
16312 * ira-costs.c (scan_one_insn): Likewise.
16313 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16314 * lra.c (check_rtl): Likewise.
16315 * reload1.c (elimination_costs_in_insn): Likewise.
16316 * reorg.c (follow_jumps): Likewise.
16318 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
16319 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16320 (thumb_far_jump_used_p): Likewise.
16321 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
16322 (workaround_speculation): Likewise.
16323 (add_sched_insns_for_speculation): Likewise.
16324 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
16325 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16326 (frv_for_each_packet): Likewise.
16327 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
16328 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
16329 (final_emit_insn_group_barriers): Likewise.
16330 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16331 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16332 (mips16_insn_length): Likewise.
16333 * config/pa/pa.c (pa_reorg): Likewise.
16334 (pa_combine_instructions): Likewise.
16335 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
16336 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
16337 (sh_reorg): Likewise.
16338 (split_branches): Likewise.
16339 * config/spu/spu.c (get_branch_target): Likewise.
16341 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
16344 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
16346 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
16347 Fix declaration name.
16349 2013-03-28 Lawrence Crowl <crowl@google.com>
16351 * graphds.h (struct graph.indicies): Remove unused.
16352 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
16353 (SCOP_ORIGINAL_PDDRS): Remove unused.
16354 * sese.h (extern insert_loop_close_phis): Removed unused.
16355 (extern insert_guard_phis): Removed unused.
16356 (extern ivtype_map_elt_info): Removed unused.
16357 (new_ivtype_map_elt): Removed unused.
16358 * sese.c (ivtype_map_elt_info): Removed unused.
16360 2013-03-28 Lawrence Crowl <crowl@google.com>
16362 * Makefile.in: Add several missing include dependences.
16364 (test-dump.o): New. This object is not added to any executable,
16365 but is present for ad-hoc testing.
16367 (debug (const bitmap_head_def &)): New.
16368 (debug (const bitmap_head_def *)): New.
16370 (extern debug (const bitmap_head_def &)): New.
16371 (extern debug (const bitmap_head_def *)): New.
16373 (debug (edge_def &)): New.
16374 (debug (edge_def *)): New.
16376 (debug (basic_block_def &)): New.
16377 (debug (basic_block_def *)): New.
16379 (dump_node (const_tree, int, FILE *)): Correct source file.
16381 (debug (die_struct &)): New.
16382 (debug (die_struct *)): New.
16384 (extern debug (die_struct &)): New.
16385 (extern debug (die_struct *)): New.
16386 * gimple-pretty-print.c
16387 (debug (gimple_statement_d &)): New.
16388 (debug (gimple_statement_d *)): New.
16389 * gimple-pretty-print.h
16390 (extern debug (gimple_statement_d &)): New.
16391 (extern debug (gimple_statement_d *)): New.
16393 (debug (ira_allocno_copy &)): New.
16394 (debug (ira_allocno_copy *)): New.
16395 (debug (ira_allocno &)): New.
16396 (debug (ira_allocno *)): New.
16398 (extern debug (ira_allocno_copy &)): New.
16399 (extern debug (ira_allocno_copy *)): New.
16400 (extern debug (ira_allocno &)): New.
16401 (extern debug (ira_allocno *)): New.
16403 (debug (live_range &)): New.
16404 (debug (live_range *)): New.
16406 (debug (lra_live_range &)): New.
16407 (debug (lra_live_range *)): New.
16409 (debug (lra_live_range &)): New.
16410 (debug (lra_live_range *)): New.
16412 (debug (omega_pb_d &)): New.
16413 (debug (omega_pb_d *)): New.
16415 (extern debug (omega_pb_d &)): New.
16416 (extern debug (omega_pb_d *)): New.
16418 (debug (const rtx_def &)): New.
16419 (debug (const rtx_def *)): New.
16421 (debug_tree (tree): Move within file.
16422 (debug_raw (const tree_node &)): New.
16423 (debug_raw (const tree_node *)): New.
16424 (dump_tree_via_hooks (const tree_node *, int)): New.
16425 (debug (const tree_node &)): New.
16426 (debug (const tree_node *)): New.
16427 (debug_verbose (const tree_node &)): New.
16428 (debug_verbose (const tree_node *)): New.
16429 (debug_head (const tree_node &)): New.
16430 (debug_head (const tree_node *)): New.
16431 (debug_body (const tree_node &)): New.
16432 (debug_body (const tree_node *)): New.
16433 (debug_vec_tree (tree): Move and reimplement in terms of dump.
16434 (debug (vec<tree, va_gc> &)): New.
16435 (debug (vec<tree, va_gc> *)): New.
16437 (extern debug (const rtx_def &)): New.
16438 (extern debug (const rtx_def *)): New.
16440 (debug_raw (simple_bitmap_def &)): New.
16441 (debug_raw (simple_bitmap_def *)): New.
16442 (debug (simple_bitmap_def &)): New.
16443 (debug (simple_bitmap_def *)): New.
16445 (extern debug (simple_bitmap_def &)): New.
16446 (extern debug (simple_bitmap_def *)): New.
16447 (extern debug_raw (simple_bitmap_def &)): New.
16448 (extern debug_raw (simple_bitmap_def *)): New.
16450 (debug (vinsn_def &)): New.
16451 (debug (vinsn_def *)): New.
16452 (debug_verbose (vinsn_def &)): New.
16453 (debug_verbose (vinsn_def *)): New.
16454 (debug (expr_def &)): New.
16455 (debug (expr_def *)): New.
16456 (debug_verbose (expr_def &)): New.
16457 (debug_verbose (expr_def *)): New.
16458 (debug (vec<rtx> &)): New.
16459 (debug (vec<rtx> *)): New.
16461 (extern debug (vinsn_def &)): New.
16462 (extern debug (vinsn_def *)): New.
16463 (extern debug_verbose (vinsn_def &)): New.
16464 (extern debug_verbose (vinsn_def *)): New.
16465 (extern debug (expr_def &)): New.
16466 (extern debug (expr_def *)): New.
16467 (extern debug_verbose (expr_def &)): New.
16468 (extern debug_verbose (expr_def *)): New.
16469 (extern debug (vec<rtx> &)): New.
16470 (extern debug (vec<rtx> *)): New.
16472 (_list_iter_cond_expr): Make inline instead of static.
16474 (debug (sreal &)): New.
16475 (debug (sreal *)): New.
16477 (extern debug (sreal &)): New.
16478 (extern debug (sreal *)): New.
16480 (extern debug_raw (const tree_node &)): New.
16481 (extern debug_raw (const tree_node *)): New.
16482 (extern debug (const tree_node &)): New.
16483 (extern debug (const tree_node *)): New.
16484 (extern debug_verbose (const tree_node &)): New.
16485 (extern debug_verbose (const tree_node *)): New.
16486 (extern debug_head (const tree_node &)): New.
16487 (extern debug_head (const tree_node *)): New.
16488 (extern debug_body (const tree_node &)): New.
16489 (extern debug_body (const tree_node *)): New.
16490 (extern debug (vec<tree, va_gc> &)): New.
16491 (extern debug (vec<tree, va_gc> *)): New.
16493 (debug (struct loop &)): New.
16494 (debug (struct loop *)): New.
16495 (debug_verbose (struct loop &)): New.
16496 (debug_verbose (struct loop *)): New.
16497 * tree-dump.c: Add header dependence.
16499 (extern debug (struct loop &)): New.
16500 (extern debug (struct loop *)): New.
16501 (extern debug_verbose (struct loop &)): New.
16502 (extern debug_verbose (struct loop *)): New.
16504 (debug (data_reference &)): New.
16505 (debug (data_reference *)): New.
16506 (debug (vec<data_reference_p> &)): New.
16507 (debug (vec<data_reference_p> *)): New.
16508 (debug (vec<ddr_p> &)): New.
16509 (debug (vec<ddr_p> *)): New.
16511 (extern debug (data_reference &)): New.
16512 (extern debug (data_reference *)): New.
16513 (extern debug (vec<data_reference_p> &)): New.
16514 (extern debug (vec<data_reference_p> *)): New.
16515 (extern debug (vec<ddr_p> &)): New.
16516 (extern debug (vec<ddr_p> *)): New.
16518 (debug (pt_solution &)): New.
16519 (debug (pt_solution *)): New.
16521 (extern debug (pt_solution &)): New.
16522 (extern debug (pt_solution *)): New.
16524 (debug (_var_map &)): New.
16525 (debug (_var_map *)): New.
16526 (debug (tree_live_info_d &)): New.
16527 (debug (tree_live_info_d *)): New.
16529 (extern debug (_var_map &)): New.
16530 (extern debug (_var_map *)): New.
16531 (extern debug (tree_live_info_d &)): New.
16532 (extern debug (tree_live_info_d *)): New.
16534 2013-03-28 Jan Hubicka <jh@suse.cz>
16536 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16538 2013-03-28 Ian Bolton <ian.bolton@arm.com>
16540 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16541 record only when desired or required.
16543 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
16545 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16546 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16548 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
16550 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16551 (*andsi3_compare0_uxtw): New pattern.
16552 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16553 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16555 2013-03-28 Jan Hubicka <jh@suse.cz>
16557 * data-streamer-in.c (streamer_read_gcov_count): New function.
16558 * gimple-streamer-out.c: Include value-prof.h.
16559 (output_gimple_stmt): Output histogram.
16560 (output_bb): Use streamer_write_gcov_count.
16561 * value-prof.c: Include data-streamer.h
16562 (dump_histogram_value): Add HIST_TYPE_MAX.
16563 (stream_out_histogram_value): New function.
16564 (stream_in_histogram_value): New function.
16565 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16566 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16567 * data-streamer-out.c (streamer_write_gcov_count): New function.
16568 (streamer_write_gcov_count_stream): New function.
16569 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16570 (lto_output_node): Likewise.
16571 (input_node, input_edge): Likewise.
16572 * lto-streamer-out.c (output_cfg): Update streaming.
16573 * lto-streamer-in.c (input_cfg): Likewise.
16574 * data-streamer.h (streamer_write_gcov_count,
16575 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16576 * gimple-streamer-in.c: Include value-prof.h
16577 (input_gimple_stmt): Input histograms.
16578 (input_bb): Update profile streaming.
16580 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16582 * genmodes.c (emit_max_int): New function.
16583 (emit_insn_modes_h): Added call to emit_max_function.
16584 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16586 * machmode.def: Fixed comment.
16588 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
16590 * combine.c (try_combine): Removed useless assert.
16591 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16593 2013-03-28 Marek Polacek <polacek@redhat.com>
16594 Richard Biener <rguenther@suse.de>
16596 PR tree-optimization/56695
16597 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16598 build signed result of a vector comparison.
16599 * tree-cfg.c (verify_gimple_comparison): Check that a result
16600 of a vector comparison has signed type.
16602 2013-03-28 Richard Biener <rguenther@suse.de>
16604 PR tree-optimization/37021
16605 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16606 do not restrict gaps between groups.
16607 * tree-vect-stmts.c (vectorizable_load): Properly account for
16608 a gap between groups.
16610 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16612 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16613 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16616 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16618 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16619 * doc/extend.texi (Named Address Spaces): Ditto.
16620 (Variable Attributes): Ditto.
16622 2013-03-27 Kai Tietz <ktietz@redhat.com>
16624 * config.build: Add support for cygwin x64 target.
16625 * config.gcc: Likewise.
16626 * config.host: Likewise.
16627 * configure.ac: Likewise
16628 * configure: Regenerated.
16630 2013-03-27 Kai Tietz <ktietz@redhat.com>
16632 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16633 * config/i386/t-cygwin-w64: New file.
16634 * config/i386/cygwin-w64.h: New file.
16635 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16636 and add support for x64-cygwin target.
16637 (CPP_SPEC): Likewise.
16638 (CXX_WRAP_SPEC_LIST): Undefine before define.
16639 (LIBGCJ_SONAME): Use 15 as version.
16641 2013-03-27 Richard Biener <rguenther@suse.de>
16643 PR tree-optimization/56716
16644 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16645 dumping for ref nodes.
16647 2013-03-27 Martin Jambor <mjambor@suse.cz>
16649 PR tree-optimization/55334
16650 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16651 restricted pointers to arrays.
16653 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16655 * Makefile.in (.SUFFIXES): Add .cc.
16656 (.c.o): Apply same recipe for implicit rule .cc.o.
16658 2013-03-27 Richard Biener <rguenther@suse.de>
16660 PR tree-optimization/37021
16661 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16662 REALPART/IMAGPART_EXPRs around the supported refs.
16663 * tree-ssa-structalias.c (find_func_aliases): Assume that
16664 floating-point values are not used to transfer pointers.
16666 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16668 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16669 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16670 * doc/tm.texi: Regenerate.
16671 * targhooks.h (default_has_ifunc_p): New.
16672 * targhooks.c (default_has_ifunc_p): Ditto.
16673 * config/linux-protos.h: New file.
16674 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16675 hook for linux which disables support of indirect functions in android.
16676 * config/linux-android.c: New file.
16677 * config/t-linux-android.c: Ditto.
16678 * config.gcc: Added new object file linux-android.o.
16679 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
16680 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
16681 * varasm.c (do_assemble_alias): Likewise.
16682 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16683 doesn't support indirect functions.
16684 * configure: Regenerate.
16686 2013-03-27 Bin Cheng <bin.cheng@arm.com>
16689 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16690 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16693 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16695 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16697 2013-03-27 Terry Guo <terry.guo@arm.com>
16699 * config/arm/arm-cores.def: Added core cortex-r7.
16700 * config/arm/arm-tune.md: Regenerated.
16701 * config/arm/arm-tables.opt: Regenerated.
16702 * doc/invoke.texi: Added entry for core cortex-r7.
16704 2013-03-27 Walter Lee <walt@tilera.com>
16706 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16707 double-decrement of next_scratch_regno.
16709 2013-03-27 Walter Lee <walt@tilera.com>
16711 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
16713 (insn_v1mulus): Ditto.
16714 (insn_v2muls): Ditto.
16716 2013-03-27 Walter Lee <walt@tilera.com>
16718 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
16719 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16721 2013-03-27 Walter Lee <walt@tilera.com>
16723 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
16724 (*sibcall_value): Ditto.
16726 2013-03-27 Walter Lee <walt@tilera.com>
16728 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16729 (insn_mnz_v8qi): ... this ...
16730 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16732 (insn_v<n>mnz): Replaced by ...
16733 (insn_v1mnz): ... this ...
16734 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16736 (insn_mz_<mode>): Replaced by ...
16737 (insn_mz_v8qi): ... this ...
16738 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16740 (insn_v<n>mz): Replaced by ...
16741 (insn_v1mz): ... this ...
16742 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16745 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16747 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16749 2013-03-26 Roland McGrath <mcgrathr@google.com>
16751 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16752 than fprintf with a non-constant, non-format string.
16754 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16756 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16757 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16758 operand 0 predicate.
16759 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16760 attribute. Use general_x64nomem_operand as operand 1 predicate.
16761 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16762 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16763 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16764 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16765 (mov<mode>_insv_1): Remove expander. Merge insn with
16766 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16767 Use general_x64nomem_operand as operand 1 predicate.
16768 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16769 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16770 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16771 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16772 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16773 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16774 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16775 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16776 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16777 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16778 (general_x64nomem_operand): Ditto.
16780 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
16782 * config/rtems.opt: Add -pthread option.
16784 2013-03-26 Richard Biener <rguenther@suse.de>
16786 * alias.c (find_base_term): Avoid redundant and not used recursion.
16787 (base_alias_check): Get the initial base term from the caller.
16788 (true_dependence_1): Compute and pass base terms to base_alias_check.
16789 (write_dependence_p): Likewise.
16790 (may_alias_p): Likewise.
16792 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16794 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16795 PC-relative load in SI modes and above only.
16797 2013-03-26 Xinyu Qi <xyqi@marvell.com>
16799 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
16800 * config/arm/iwmmxt.md (WCGR0): Update.
16801 (WCGR1, WCGR2, WCGR3): Likewise.
16803 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
16805 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16806 Use x64 and nox64 isa attributes.
16808 2013-03-26 Richard Biener <rguenther@suse.de>
16810 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16811 alignment computations and rely on get_object_alignment_1
16812 for the !TYPE_P case.
16813 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16815 2013-03-26 Walter Lee <walt@tilera.com>
16817 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16818 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16820 2013-03-25 Jeff Law <law@redhat.com>
16822 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16823 check for INTEGRAL_TYPE_P that was missing due to checking in
16824 wrong version of prior patch.
16826 2013-03-25 Walter Lee <walt@tilera.com>
16828 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16829 TILEGX_INSN_SHUFFLEBYTES1.
16830 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16832 (tilegx_builtins): Ditto.
16833 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16835 2013-03-25 Walter Lee <walt@tilera.com>
16837 * config/tilegx/tilegx.md (floatsisf2): New pattern.
16838 (floatunssisf2): New pattern.
16839 (floatsidf2): New pattern.
16840 (floatunssidf2): New pattern.
16842 2013-03-25 Walter Lee <walt@tilera.com>
16844 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16845 tests for constraint J, K, N, P.
16847 2013-03-25 Walter Lee <walt@tilera.com>
16849 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16850 Use indirect/pcrel encoding.
16851 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16854 2013-03-25 Steve Ellcey <sellcey@mips.com>
16856 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16857 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16858 * config/mips/mips.c (mips_option_override): Set IMADD default.
16859 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16860 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16861 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16862 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16863 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16865 2013-03-25 Jeff Law <law@redhat.com>
16867 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16868 slightly to avoid creating and folding useless trees. Simplify
16869 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16871 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
16873 * config/i386/i386.md (*zero_extendsidi2): Merge with
16874 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16875 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16876 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16877 targets. Clarify comment.
16879 2013-03-25 Martin Jambor <mjambor@suse.cz>
16881 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16882 pass-through jump functions differently.
16883 (ipa_read_jump_function): Likewise. Also use setter functions to set
16886 2013-03-25 Martin Jambor <mjambor@suse.cz>
16888 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16889 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16891 (ipa_get_indirect_edge_target): New function.
16892 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16893 ipa_get_indirect_edge_target. Update all callers.
16894 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16895 ipa_get_indirect_edge_target_1 instead of calling
16896 ipa_get_indirect_edge_target.
16897 (create_specialized_node): Pass aggvlas to
16898 ipcp_discover_new_direct_edges.
16900 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16902 * config/arm/arm.md (f_sels, f_seld): New types.
16903 (*cmov<mode>): New pattern.
16904 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16907 2013-03-25 Kai Tietz <ktietz@redhat.com>
16909 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16910 POSIX-printf for mingw-hosted builds.
16912 2013-03-25 Richard Biener <rguenther@suse.de>
16914 PR middle-end/56694
16915 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16916 must-not-throw stmt location.
16918 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16920 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16921 Emit load-acquire versions when acq is true.
16922 (arm_emit_store_exclusive): Add rel parameter.
16923 Emit store-release versions when rel is true.
16924 (arm_split_compare_and_swap): Use acquire-release instructions
16926 of barriers when appropriate.
16927 (arm_split_atomic_op): Likewise.
16928 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16929 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16930 (VUNSPEC_SLX): Likewise.
16931 (VUNSPEC_LDA): Likewise.
16932 (VUNSPEC_STL): Likewise.
16933 * config/arm/sync.md (atomic_load<mode>): New pattern.
16934 (atomic_store<mode>): Likewise.
16935 (arm_load_acquire_exclusive<mode>): Likewise.
16936 (arm_load_acquire_exclusivesi): Likewise.
16937 (arm_load_acquire_exclusivedi): Likewise.
16938 (arm_store_release_exclusive<mode>): Likewise.
16940 2013-03-25 Catherine Moore <clm@codesourcery.com>
16942 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16943 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16944 * config/mip/predicates.md (lwsp_swsp_operand,
16945 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16946 sb16_operand, db4_operand, db7_operand, ib3_operand,
16947 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16948 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16949 andi16_operand): New predicates.
16950 * config/mips/mips.md (compression): New attribute.
16951 (enabled): New attribute.
16952 (length): Consider compression in computing length.
16953 (shift_compression): New code attribute.
16954 (*add<mode>3): New operands. Record compression.
16955 (sub<mode>3): Likewise.
16956 (one_cmpl<mode>2): Likewise.
16957 (*and<mode>3): Likewise.
16958 (*ior<mode>3): Likewise.
16959 (unnamed pattern for xor): Likewise.
16960 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16961 (*<optab><mode>3): Likewise.
16962 (*mov<mode>_internal: Likewise.
16963 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16964 (mips_unsigned_immediate_p): New.
16965 (umips_lwsp_swsp_address_p): New.
16966 (m16_based_address_p): New.
16967 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16968 (mips_unsigned_immediate_p): New prototype.
16969 (lwsp_swsp_address_p): New prototype.
16970 (m16_based_address_p): New prototype.
16971 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16972 (mips_signed_immediate_p): New function.
16973 (m16_based_address_p): New function.
16974 (lwsp_swsp_address_p): New function.
16975 (mips_print_operand_punctuation): Recognize short delay slot insns
16976 for microMIPS.add<mode>3"
16978 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16981 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16982 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16984 2013-03-25 Richard Biener <rguenther@suse.de>
16986 PR tree-optimization/56689
16987 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16990 2013-03-25 Richard Biener <rguenther@suse.de>
16992 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16994 (memory_references): Likewise.
16995 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16996 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16997 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16998 (gather_mem_refs_in_loops): Fold into ...
16999 (analyze_memory_references): ... this. Move initialization
17000 to tree_ssa_lim_initialize.
17001 (fill_always_executed_in): Rename to ...
17002 (fill_always_executed_in_1): ... this.
17003 (fill_always_executed_in): Move contains_call computation to
17004 this new function from ...
17005 (tree_ssa_lim_initialize): ... here.
17006 (tree_ssa_lim): Call fill_always_executed_in.
17008 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17010 * postreload.c (reload_combine): Fix code detecting returns.
17012 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
17014 * function.c (emit_use_return_register_into_block): On cc0 targets,
17015 do not emit the sequence between cc0 setter and user.
17017 2013-03-25 Kai Tietz <ktietz@redhat.com>
17019 * config/i386/predicates.md (local_symbolic_operand): Interpret
17020 dll-imported symbols as none-local.
17022 2013-03-25 Richard Biener <rguenther@suse.de>
17024 * tree-ssa-loop-im.c (struct depend): Remove.
17025 (struct lim_aux_data): Make depends a vec of gimples.
17026 (free_lim_aux_data): Adjust.
17027 (add_dependency): Likewise.
17028 (set_level): Likewise.
17030 2013-03-25 Richard Biener <rguenther@suse.de>
17032 PR middle-end/56434
17033 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
17034 the pointer returned by calls with ECF_MALLOC set.
17036 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17038 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
17040 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
17042 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
17043 using MMXMODE mode iterator.
17044 (*move<mode>_internal): Merge with *movv2sf_internal and
17045 *movv2sf_internal_rex64 using MMXMODE mode iterator.
17047 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17049 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
17050 (record_last_mem_set_info): Likewise.
17052 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
17053 of XNEWVEC followed by memset.
17054 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
17056 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
17058 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
17059 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
17060 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
17061 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
17062 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
17063 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
17064 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
17065 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
17066 BARRIER_P instead of GET_CODE.
17068 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17070 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
17071 inaccuracy in the probing code.
17073 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
17074 (ctrapdi4): Likewise.
17076 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17078 * calls.c (expand_call): Add missing guard to code handling return
17079 of non-BLKmode structures in MSB.
17080 * function.c (expand_function_end): Likewise.
17082 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
17084 * combine.c (try_combine): Adjust comment. Do not add the set of
17085 insn #0 if the destination indirectly is set or dies in insn #2.
17086 Tidy up code to distribute a new note.
17088 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17090 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
17091 also for alternatives 16 and 17.
17093 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
17095 * config/i386/sse.md (*mov<mode>_internal): Merge with
17096 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
17097 Emit insn template depending on type attribute. Use
17098 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17099 movd instead of movq mnemonic for interunit moves. Rewrite mode
17100 attribute calculation. Remove unit attribute calculation.
17101 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
17102 Set prefix_data16 attribute for DImode ssemov types.
17103 Use Ym instead of y for SSE-MMX conversion alternatives.
17104 Reorder operand constraints.
17106 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
17108 * df.h (df_insn_delete): Adjust prototype.
17109 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
17110 and let it decide whether mark the basic block dirty.
17111 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
17112 * df-scan.c (df_insn_info_delete): New helper function, split
17113 off from df_insn_delete.
17114 (df_scan_free_bb_info): Use it.
17115 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
17117 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
17118 that the insn is actually an insn and it has a non-NULL basic block.
17119 Do not mark basic block dirty if only deleting a DEBUG_INSN.
17121 2013-03-22 Richard Biener <rguenther@suse.de>
17123 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
17125 (mem_ref_alloc): Do not allocate them.
17126 (refs_independent_p): Do not query or maintain a cache.
17128 2013-03-22 Richard Biener <rguenther@suse.de>
17130 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
17131 (gather_mem_refs_in_loops): Do not compute it.
17132 (analyze_memory_references): Do not allocate it.
17133 (tree_ssa_lim_finalize): Do not free it.
17134 (for_all_locs_in_loop): Do not query all_refs_in_loop.
17136 2013-03-22 Richard Biener <rguenther@suse.de>
17138 * is-a.h (as_a): Use gcc_checking_assert.
17140 2013-03-22 Ian Bolton <ian.bolton@arm.com>
17142 * config/aarch64/aarch64.c (aarch64_print_operand): New
17143 format specifier for printing a constant in hex.
17144 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
17145 format specifier for printing second operand.
17147 2013-03-22 Richard Biener <rguenther@suse.de>
17149 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
17151 (gather_mem_refs_in_loops): Perform store accumulation here.
17152 (create_vop_ref_mapping_loop): Remove.
17153 (create_vop_ref_mapping): Likewise.
17154 (analyze_memory_references): Initialize refs_stored_in_loop.
17155 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
17156 (record_indep_loop): Remove.
17157 (record_dep_loop): New function.
17158 (ref_indep_loop_p_1): Adjust to only walk over references
17159 in the loop, not its subloops.
17160 (ref_indep_loop_p): Rename to ...
17161 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
17162 maintaining a more fine-grained cache.
17163 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
17164 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
17166 2013-03-22 Richard Biener <rguenther@suse.de>
17168 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
17169 (struct mem_ref): Make accesses_in_loop a vec of a vec of
17170 aggregate mem_ref_loc.
17171 (free_mem_ref_locs): Inline into ...
17172 (memref_free): ... this and adjust.
17173 (mem_ref_alloc): Adjust.
17174 (mem_ref_locs_alloc): Remove.
17175 (record_mem_ref_loc): Adjust.
17176 (get_all_locs_in_loop): Rewrite into ...
17177 (for_all_locs_in_loop): ... this iterator.
17178 (rewrite_mem_ref_loc): New functor.
17179 (rewrite_mem_refs): Use for_all_locs_in_loop.
17180 (sm_set_flag_if_changed): New functor.
17181 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
17182 (ref_always_accessed): New functor.
17183 (ref_always_accessed_p): Use for_all_locs_in_loop.
17185 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17187 * tree-pass.h (PROP_gimple_lvec): New.
17188 * passes.c (dump_properties): Handle PROP_gimple_lvec.
17189 (init_optimization_passes): Move pass_lower_vector.
17190 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
17192 (pass_lower_vector): Provide PROP_gimple_lvec.
17193 (pass_lower_vector_ssa): Likewise.
17194 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
17196 2013-03-21 Mark Wielaard <mjw@redhat.com>
17198 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
17200 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17202 * config/i386/i386.md (*movdi_internal): Disparage slightly
17203 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
17204 conversion alternatives.
17206 2013-03-21 Jakub Jelinek <jakub@redhat.com>
17208 PR middle-end/48087
17209 * diagnostic.def (DK_WERROR): New kind.
17210 * diagnostic.h (werrorcount): Define.
17211 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
17212 promoted to DK_ERROR, increment DK_WERROR counter instead of
17214 * toplev.c (toplev_main): Call print_ignored_options even if
17215 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
17216 even if just werrorcount is non-zero.
17219 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
17221 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
17222 (string_cst_pool_decl): New function.
17223 (optimize_one_addr_into_implicit_ptr): New function.
17224 (resolve_addr_in_expr): Optimize DWARF location expression
17225 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
17226 which doesn't live in memory, but has DW_AT_location or
17227 DW_AT_const_value, or refers to a string literal, into
17228 DW_OP_GNU_implicit_pointer.
17229 (optimize_location_into_implicit_ptr): New function.
17230 (resolve_addr): If removing DW_AT_location of a variable because
17231 it was DW_OP_addr of address of the variable, but the variable doesn't
17232 live in memory, try to emit const value attribute for the initializer.
17234 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17236 * tree.h (VECTOR_TYPE_P): New macro.
17237 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
17238 TYPE_MODE): Use it.
17239 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
17240 VEC_COND_EXPR cannot be lvalues.
17241 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
17243 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
17245 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
17246 Restrict the transformation to equal modes.
17248 2013-03-21 Richard Biener <rguenther@suse.de>
17250 PR tree-optimization/39326
17251 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
17252 (MEM_ANALYZABLE): Adjust.
17253 (record_mem_ref_loc): Move bitmap ops ...
17254 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
17255 unanalyzable refs, do not record locations for it.
17256 (analyze_memory_references): Allocate ref zero as shared
17258 (refs_independent_p): Do not test for unanalyzed mems here.
17259 (ref_indep_loop_p_1): Special-case disambiguation against
17260 the unanalyzed ref.
17261 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
17263 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
17265 * config/arm/arm-protos.h (tune_params): Add
17266 prefer_neon_for_64bits field.
17267 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
17268 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
17269 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
17270 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
17271 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
17272 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
17273 (arm_option_override): Handle -mneon-for-64bits new option.
17274 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
17275 (prefer_neon_for_64bits): Declare new variable.
17276 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
17277 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
17278 (arch_enabled): Handle new arch types. Remove support for onlya8
17280 (one_cmpldi2): Use new arch names.
17281 (zero_extend<mode>di2, extend<mode>di2): Ditto.
17282 * config/arm/arm.opt (mneon-for-64bits): Add option.
17283 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
17284 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
17285 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
17287 * doc/invoke.texi (-mneon-for-64bits): Document.
17289 2013-03-21 Richard Biener <rguenther@suse.de>
17291 PR tree-optimization/39326
17292 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
17293 (sort_bbs_in_loop_postorder_cmp): New function.
17294 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
17296 2013-03-21 Richard Biener <rguenther@suse.de>
17298 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
17299 (vect_insert_into_interleaving_chain): Likewise.
17300 (vect_drs_dependent_in_basic_block): Inline ...
17301 (vect_slp_analyze_data_ref_dependence): ... here. New function,
17303 (vect_analyze_data_ref_dependence): ... here. Simplify.
17304 (vect_check_interleaving): Simplify.
17305 (vect_analyze_data_ref_dependences): Likewise. Split out ...
17306 (vect_slp_analyze_data_ref_dependences): ... this new function.
17307 (dr_group_sort_cmp): New function.
17308 (vect_analyze_data_ref_accesses): Compute data-reference groups
17309 here instead of in vect_analyze_data_ref_dependence. Use
17310 a more efficient algorithm.
17311 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
17312 vect_slp_analyze_data_ref_dependences. Call
17313 vect_analyze_data_ref_accesses earlier.
17314 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
17315 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
17316 (vect_slp_analyze_data_ref_dependences): New prototype.
17318 2013-03-21 Richard Biener <rguenther@suse.de>
17320 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
17321 ref is stored in the loop.
17322 (find_refs_for_sm): Walk only over all stores.
17323 (store_motion_loop): Allocate from lim_bitmap_obstack.
17324 (store_motion): Likewise.
17326 2013-03-21 Richard Biener <rguenther@suse.de>
17328 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
17329 Update virtual SSA form.
17331 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17333 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
17334 * configure: Regenerate.
17335 * config.in: Regenerate.
17336 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
17337 if !HAVE_LD_EH_FRAME_CIEV3.
17339 2013-03-21 Richard Biener <rguenther@suse.de>
17341 * tree-cfg.c (verify_expr_no_block): New function.
17342 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
17343 nor DECL_VALUE_EXPR have locations with associated blocks.
17344 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
17345 (clear_unused_block_pointer): Remove code dealing with
17346 blocks in DECL_DEBUG_EXPR locations.
17348 2013-03-21 Richard Biener <rguenther@suse.de>
17350 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
17351 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
17352 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
17353 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
17354 instead of DECL_DEBUG_EXPR_IS_FROM.
17355 * gimplify.c (gimplify_modify_expr): Likewise.
17356 * tree-cfg.c (verify_expr_location_1): Likewise.
17357 * tree-complex.c (create_one_component_var): Likewise.
17358 * tree-sra.c (create_access_replacement): Likewise.
17359 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
17360 (clear_unused_block_pointer): Likewise.
17361 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
17362 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
17363 * var-tracking.c (var_debug_decl): Likewise.
17364 (track_expr_p): Likewise.
17365 * tree-inline.c (add_local_variables): Likewise. Set
17366 DECL_HAS_DEBUG_EXPR_P after copying it.
17367 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
17368 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
17370 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
17373 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
17374 * configure: Regenerate.
17375 * config.in: Regenerate.
17376 * config/i386/i386.md (*movdf_internal): Use
17377 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17378 movd instead of movq mnemonic for interunit moves.
17379 (*movdi_internal): Ditto.
17381 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17383 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
17384 (abd<mode>_3): New pattern.
17385 (aba<mode>_3): New pattern.
17386 (fabd<mode>_3): New pattern.
17388 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17390 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
17391 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
17392 occurrence of REGISTER_PREFIX as its empty string.
17394 2013-03-20 Jeff Law <law@redhat.com>
17396 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
17397 addititional equivalences for equality comparisons between an SSA_NAME
17398 and a constant where the SSA_NAME was set from a widening conversion.
17400 2013-03-20 Walter Lee <walt@tilera.com>
17402 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
17404 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17406 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
17407 depending on type attribute.
17408 (*movti_internal): Ditto.
17409 (*movtf_internal): Ditto.
17410 (*movxf_internal): Ditto.
17411 (*movdf_internal): Ditto.
17412 (*movsf_internal): Ditto.
17414 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17416 * config/i386/i386.md (*movti_internal): Set prefix attribute to
17417 maybe_vex for sselog1 and ssemov types.
17418 (*movdi_internal): Reorder operand constraints.
17419 (*movsi_internal): Ditto. Set prefix attribute to
17420 maybe_vex for sselog1 and ssemov types.
17421 (*movtf_internal): Set prefix attribute to maybe_vex
17422 for sselog1 and ssemov types.
17423 (*movdf_internal): Ditto. Set prefix_data16 attribute for
17424 DImode ssemov types. Reorder operand constraints.
17425 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
17426 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
17427 attribute for SImode ssemov types. Reorder operand constraints.
17429 2013-03-20 Martin Jambor <mjambor@suse.cz>
17431 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
17432 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
17434 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
17436 * config/rs6000/predicates.md (indexed_address, update_address_mem
17437 update_indexed_address_mem): New predicates.
17438 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
17439 attribute for load/store instructions.
17440 * config/rs6000/dfp.md (movsd_store): Likewise.
17441 (movsd_load): Likewise.
17442 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
17443 (unnamed HI->DI extend define_insn): Likewise.
17444 (unnamed SI->DI extend define_insn): Likewise.
17445 (unnamed QI->SI extend define_insn): Likewise.
17446 (unnamed QI->HI extend define_insn): Likewise.
17447 (unnamed HI->SI extend define_insn): Likewise.
17448 (unnamed HI->SI extend define_insn): Likewise.
17449 (extendsfdf2_fpr): Likewise.
17450 (movsi_internal1): Likewise.
17451 (movsi_internal1_single): Likewise.
17452 (movhi_internal): Likewise.
17453 (movqi_internal): Likewise.
17454 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
17455 attribute for load/store instructions.
17456 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17458 (mov<mode>_softfloat): Likewise.
17459 (mov<mode>_hardfloat32): Likewise.
17460 (mov<mode>_hardfloat64): Likewise.
17461 (mov<mode>_softfloat64): Likewise.
17462 (movdi_internal32): Likewise.
17463 (movdi_internal64): Likewise.
17464 (probe_stack_<mode>): Likewise.
17466 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17468 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17469 floating point, and decimal floating point to reload iterator.
17471 * config/rs6000/constraints.md (wl constraint): New constraints to
17472 return FLOAT_REGS if certain options are used to reduce the number
17473 of separate patterns that exist in the file.
17474 (wx constraint): Likewise.
17475 (wz constraint): Likewise.
17477 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17478 -mdebug=reg, print wg, wl, wx, and wz constraints.
17479 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17480 Initialize the reload functions for 64-bit binary/decimal floating
17482 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17483 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17484 create the buffer on the stack to overcome not having a 32-bit
17486 (rs6000_emit_move): Likewise.
17487 (rs6000_secondary_memory_needed_rtx): Likewise.
17488 (rs6000_alloc_sdmode_stack_slot): Likewise.
17489 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17490 via xxlxor, just like DFmode 0.0.
17492 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17493 define as 1 if we are running on a power7 or newer.
17494 (enum r6000_reg_class_enum): Add new constraints.
17496 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17497 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17498 with other moves by using conditional constraits (wg). Use LFIWZX
17499 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
17500 (movsd splitter): Likewise.
17501 (movsd_hardfloat): Likewise.
17502 (movsd_softfloat): Likewise.
17504 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17505 binary and decimal floating point moves.
17506 (fmove_ok): New attributes to combine binary and decimal floating
17507 point moves, and to combine power6x (mfpgpr) moves along normal
17509 (real_value_to_target): Likewise.
17510 (f32_lr): Likewise.
17511 (f32_lm): Likewise.
17512 (f32_li): Likewise.
17513 (f32_sr): Likewise.
17514 (f32_sm): Likewise.
17515 (f32_si): Likewise.
17516 (movsf): Combine binary and decimal floating point moves. Combine
17517 power6x (mfpgpr) moves with other moves by using conditional
17518 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
17519 (mov<mode> for SFmode/SDmode); Likewise.
17520 (SFmode/SDmode splitters): Likewise.
17521 (movsf_hardfloat): Likewise.
17522 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17523 (movsf_softfloat): Likewise.
17524 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17526 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17527 wx and wz constraints.
17529 * config/rs6000/constraints.md (wg constraint): New constraint to
17530 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17532 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17535 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17536 -mdebug=reg, print wg, wl, wx, and wz constraints.
17537 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17538 Initialize the reload functions for 64-bit binary/decimal floating
17540 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17541 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17542 create the buffer on the stack to overcome not having a 32-bit
17544 (rs6000_emit_move): Likewise.
17545 (rs6000_secondary_memory_needed_rtx): Likewise.
17546 (rs6000_alloc_sdmode_stack_slot): Likewise.
17547 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17548 via xxlxor, just like DFmode 0.0.
17550 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17551 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17552 with other moves by using conditional constraits (wg). Use LFIWZX
17553 and STFIWX for loading SDmode on power7.
17554 (movdd splitters): Likewise.
17555 (movdd_hardfloat32): Likewise.
17556 (movdd_softfloat32): Likewise.
17557 (movdd_hardfloat64_mfpgpr): Likewise.
17558 (movdd_hardfloat64): Likewise.
17559 (movdd_softfloat64): Likewise.
17561 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17562 64-bit binary and decimal floating point moves.
17563 (FMOVE64X): Likewise.
17564 (movdf): Combine 64-bit binary and decimal floating point moves.
17565 Combine power6x (mfpgpr) moves with other moves by using
17566 conditional constraits (wg).
17567 (mov<mode> for DFmode/DDmode): Likewise.
17568 (DFmode/DDmode splitters): Likewise.
17569 (movdf_hardfloat32): Likewise.
17570 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17571 (movdf_softfloat32): Likewise.
17572 (movdf_hardfloat64_mfpgpr): Likewise.
17573 (movdf_hardfloat64): Likewise.
17574 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17575 (movdf_softfloat64): Likewise.
17576 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17577 (reload_<mode>_load): Move to later in the file so they aren't in
17578 the middle of the floating point move insns.
17579 (reload_<mode>_store): Likewise.
17581 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17584 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17585 constraint if -mdebug=reg.
17586 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17587 Enable using dd reload support if needed.
17589 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17590 binary and decimal floating point moves in rs6000.md.
17591 (movtd_internal): Likewise.
17593 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17594 decimal floating point moves.
17596 (movtf_internal): Likewise.
17597 (mov<mode>_internal, TDmode/TFmode): Likewise.
17598 (movtf_softfloat): Likewise.
17599 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17601 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17602 movdi_internal64, using wg constraint for move direct operations.
17603 (movdi_internal64): Likewise.
17605 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17606 MODES_TIEABLE_P for selected modes. Print the numerical value of
17607 the various virtual registers. Use GPR/FPR first/last values,
17608 instead of hard coding the register numbers. Print which modes
17609 have reload functions registered.
17610 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17611 settings before/after setting cpu, target and subtarget settings.
17612 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17613 and for secondary reload failures in rs6000_secondary_reload_inner.
17614 (rs6000_secondary_reload_fail): Likewise.
17615 (rs6000_secondary_reload_inner): Likewise.
17617 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17618 macros for first/last GPR and FPR registers.
17619 (LAST_GPR_REGNO): Likewise.
17620 (FIRST_FPR_REGNO): Likewise.
17621 (LAST_FPR_REGNO): Likewise.
17623 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17624 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17625 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17626 (vcond<mode><mode>): Likewise.
17627 (vcondu<mode><mode>): Likewise.
17628 (vector_gtu<mode>): Likewise.
17629 (vector_gte<mode>): Likewise.
17630 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
17631 to prevent the compiler from converting DImode operations to TImode.
17632 (ior<mode>3): Likewise.
17633 (and<mode>3): Likewise.
17634 (one_cmpl<mode>2): Likewise.
17635 (nor<mode>3): Likewise.
17636 (andc<mode>3): Likewise.
17638 * config/rs6000/constraints.md (wt constraint): New constraint
17639 that returns VSX_REGS if TImode is allowed in VSX registers.
17641 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17642 constant under VSX.
17644 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17645 similar to TImode, but it is restricted to being in the GPRs.
17647 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17648 TImode to occupy a single VSX register.
17650 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17651 -mvsx-timode for power7/power8.
17652 (power7 cpu): Likewise.
17653 (power8 cpu): Likewise.
17655 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17656 sure that TFmode/TDmode take up two registers if they are ever
17657 allowed in the upper VSX registers.
17658 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17660 (rs6000_init_hard_regno_mode_ok): Likewise.
17661 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17662 constraint. Print if LRA is turned on.
17663 (rs6000_option_override_internal): Give an error if -mvsx-timode
17664 and VSX is not enabled.
17665 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17666 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17667 to reg+offset addressing. Use PTImode when checking offset
17668 addresses for validity.
17669 (reg_offset_addressing_ok_p): Likewise.
17670 (rs6000_legitimate_offset_address_p): Likewise.
17671 (rs6000_legitimize_address): Likewise.
17672 (rs6000_legitimize_reload_address): Likewise.
17673 (rs6000_legitimate_address_p): Likewise.
17674 (rs6000_eliminate_indexed_memrefs): Likewise.
17675 (rs6000_emit_move): Likewise.
17676 (rs6000_secondary_reload): Likewise.
17677 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17678 reloads to fpr registers to continue to use reg+offset addressing,
17679 but 64-bit reloads to altivec registers need reg+reg addressing.
17680 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17681 it. Treat LO_SUM like a PLUS operation.
17682 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
17683 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
17684 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17685 registers to share a register with a smaller sized type, since VSX
17686 puts scalars in the upper 64-bits.
17687 (print_operand): Add support for PTImode.
17688 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17689 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17690 registers, but don't have arithmetic support.
17691 (rs6000_memory_move_cost): Add test for VSX.
17692 (rs6000_opt_masks): Add -mvsx-timode.
17694 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17697 (VSr): Use wt constraint for TImode.
17698 (VSv): Drop TImode support.
17699 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17700 (vsx_movti_64bit): Likewise.
17701 (vsx_movti_32bit): Likewise.
17702 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17703 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17704 one '?' on the appropriate output constraint. Do not allow TImode
17705 logical operations on 32-bit systems.
17706 (vsx_ior<mode>3): Likewise.
17707 (vsx_xor<mode>3): Likewise.
17708 (vsx_one_cmpl<mode>2): Likewise.
17709 (vsx_nor<mode>3): Likewise.
17710 (vsx_andc<mode>3): Likewise.
17711 (vsx_concat_<mode>): Likewise.
17712 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17714 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17715 OPTION_MASK_VSX_TIMODE.
17716 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17717 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17719 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17720 (TI2 iterator): New iterator for TImode, PTImode.
17721 (wd mode attribute): Add values for vector types.
17722 (movti_string): Replace TI move operations with operations for TImode
17723 and PTImode. Add support for TImode being allowed in VSX registers.
17724 (mov<mode>_string, TImode/PTImode): Likewise.
17725 (movti_ppc64): Likewise.
17726 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17727 (TI mode splitters): Likewise.
17729 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17732 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
17734 PR tree-optimization/56355
17735 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17736 Also handle integers with undefined overflow.
17738 2013-03-20 Catherine Moore <clm@codesourcery.com>
17739 Maciej W. Rozycki <macro@codesourcery.com>
17740 Tom de Vries <tom@codesourcery.com>
17741 Nathan Sidwell <nathan@codesourcery.com>
17742 Iain Sandoe <iain@codesourcery.com>
17743 Nathan Froyd <froydnj@codesourcery.com>
17744 Chao-ying Fu <fu@mips.com>
17746 * doc/extend.texi: (micromips, nomicromips, nocompression):
17747 Document new function attributes.
17748 * doc/invoke.texi (minterlink-compressed, mmicromips,
17749 m14k, m14ke, m14kec): Document new options.
17750 (minterlink-mips16): Update documentation.
17751 * doc/md.texi (ZC, ZD): Document new constraints.
17752 * configure.ac (gcc_cv_as_micromips): Check if linker
17753 supports the .set micromips directive.
17754 * configure: Regenerate.
17755 * config.in: Regenerate.
17756 * config/mips/mips-tables.opt: Regenerate.
17757 * config/mips/micromips.md: New file.
17758 * constraints.md (ZC, ZD): New constraints.
17759 * config/mips/predicates.md (movep_src_register): New predicate.
17760 (movep_src_operand): New predicate.
17761 (non_volatile_mem_operand): New predicate.
17762 * config/mips/mips.md (multimem): New type.
17763 (length): Differentiate between 17-bit and 18-bit branch offsets.
17764 (MOVEP1, MOVEP2): New mode iterator.
17765 (mov_<load>l): Use ZC constraint.
17766 (mov_<load>r): Likewise.
17767 (mov_<store>l): Likewise.
17768 (mov_<store>r): Likewise.
17769 (*branch_equality<mode>_inverted): Add microMIPS support.
17770 (*branch_equality<mode>): Likewise.
17771 (*jump_absolute): Likewise.
17772 (indirect_jump_<mode>): Likewise.
17773 (tablejump_<mode>): Likewise.
17774 (<optab>_internal): Likewise.
17775 (sibcall_internal): Likewise.
17776 (sibcall_value_internal): Likewise.
17777 (prefetch): Use constraint ZD.
17778 * config/mips/mips.opt (minterlink-compressed): New option.
17779 (minterlink-mips16): Now an alias for minterlink-compressed.
17780 (mmicromips): New option.
17781 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17782 (compare_and_swap_12): Likewise.
17783 (sync_add<mode>): Likewise.
17784 (sync_<optab>_12): Likewise.
17785 (sync_old_<optab>_12): Likewise.
17786 (sync_new_<optab>_12): Likewise.
17787 (sync_nand_12): Likewise.
17788 (sync_old_nand_12): Likewise.
17789 (sync_new_nand_12): Likewise.
17790 (sync_sub<mode>): Likewise.
17791 (sync_old_add<mode>): Likewise.
17792 (sync_old_sub<mode>): Likewise.
17793 (sync_new_add<mode>): Likewise.
17794 (sync_new_sub<mode>): Likewise.
17795 (sync_<optab><mode>): Likewise.
17796 (sync_old_<optab><mode>): Likewise.
17797 (sync_new_<optab><mode>): Likewise.
17798 (sync_nand<mode>): Likewise.
17799 (sync_old_nand<mode>): Likewise.
17800 (sync_new_nand<mode>): Likewise.
17801 (sync_lock_test_and_set<mode>): Likewise.
17802 (test_and_set_12): Likewise.
17803 (atomic_compare_and_swap<mode>): Likewise.
17804 (atomic_exchange<mode>_llsc): Likewise.
17805 (atomic_fetch_add<mode>_llsc): Likewise.
17806 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17807 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17808 (umips_save_restore_pattern_p): Likewise.
17809 (umips_load_store_pair_p): Likewise.
17810 (umips_output_load_store_pair): Likewise.
17811 (umips_movep_target_p): Likewise.
17812 (umips_12bit_offset_address_p): Likewise.
17813 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17814 (mips_base_mips16): Rename this...
17815 (mips_base_compression_flags): ...to this. Update all uses.
17816 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17817 (mips_mips16_decl_p): Delete.
17818 (mips_nomips16_decl_p): Delete.
17819 (mips_get_compress_on_flags): New function.
17820 (mips_get_compress_off_flags): New function.
17821 (mips_get_compress_mode): New function.
17822 (mips_get_compress_on_name): New function.
17823 (mips_get_compress_off_name): New function.
17824 (mips_insert_attributes): Support multiple compression types.
17825 (mips_merge_decl_attributes): Likewise.
17826 (umips_12bit_offset_address_p): New function.
17827 (mips_start_function_definition): Emit .set micromips directive.
17828 (mips_call_may_need_jalx_p): New function.
17829 (mips_function_ok_for_sibcall): Add microMIPS support.
17830 (mips_print_operand_punctuation): Support short delay slots and
17832 (umips_swm_mask, umips_swm_encoding): New.
17833 (umips_build_save_restore): New function.
17834 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17835 (was_mips16_p): Remove.
17836 (old_compression_mode): New.
17837 (mips_set_compression_mode): New function.
17838 (mips_set_current_function): Add microMIPS support.
17839 (mips_option_override): Likewise.
17840 (umips_save_restore_pattern_p): New function.
17841 (umips_output_save_restore): New function.
17842 (umips_load_store_pair_p_1): New function.
17843 (umips_load_store_pair_p): New function.
17844 (umips_output_load_store_pair_1): New function.
17845 (umips_output_load_store_pair): New function.
17846 (umips_movep_target_p) New function.
17847 (mips_prepare_pch_save): Add microMIPS support.
17848 * config/mips/mips.h (TARGET_COMPRESSION): New.
17849 (TARGET_CPU_CPP_BUILTINS): Update macro
17850 to use new compression flags and to support microMIPS.
17851 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17852 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17853 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17854 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17855 (ASM_SPEC): Support mmicromips and mno-micromips.
17856 (M16STORE_REG_P): New macro.
17857 (MIPS_CALL): Support TARGET_MICROMIPS.
17858 (MICROMIPS_J): New macro.
17859 (mips_base_mips16): Rename this...
17860 (mips_base_compression_flags): ...to this.
17861 (UMIPS_12BIT_OFFSET_P): New macro.
17862 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17863 (MULTILIB_DIRNAMES): Likewise.
17864 2013-03-20 Richard Biener <rguenther@suse.de>
17866 PR tree-optimization/56661
17867 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17868 the result does not have to be distinct.
17870 2013-03-20 Richard Biener <rguenther@suse.de>
17872 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17875 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17876 Steven Bosscher <steven@gcc.gnu.org>
17878 PR rtl-optimization/56605
17879 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17881 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
17884 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17885 that require movd instead of movq.
17887 2013-03-20 Richard Biener <rguenther@suse.de>
17889 * tree-ssa-structalias.c (struct variable_info): Add pointer
17890 to the first field of an aggregate with sub-vars. Make
17891 this and the pointer to the next subfield its ID.
17892 (vi_next): New function.
17893 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17894 storedanything_id, integer_id): Increment by one.
17895 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17896 get_call_clobber_vi): Adjust.
17897 (solution_set_expand): Simplify and speedup.
17898 (solution_set_add): Inline into ...
17899 (set_union_with_increment): ... this. Adjust accordingly.
17900 (do_sd_constraint): Likewise.
17901 (do_ds_constraint): Likewise.
17902 (do_complex_constraint): Simplify.
17903 (build_pred_graph): Adjust.
17904 (solve_graph): Likewise. Simplify and speedup.
17905 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17906 get_constraint_for_component_ref, get_constraint_for_1,
17907 first_vi_for_offset, first_or_preceding_vi_for_offset,
17908 create_function_info_for, create_variable_info_for_1,
17909 create_variable_info_for, intra_create_variable_infos): Adjust.
17910 (init_base_vars): Push NULL for ID zero.
17911 (compute_points_to_sets): Adjust.
17913 2013-03-20 Richard Biener <rguenther@suse.de>
17915 * cfgloop.c (verify_loop_structure): Streamline and avoid
17916 ICEing on corrupt loop tree.
17917 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17920 2013-03-20 Richard Biener <rguenther@suse.de>
17922 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17923 check whether an SSA update is needed.
17925 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17927 * config/mips/constraints.md (T): Rename to...
17931 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17932 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17934 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17936 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17937 (*subsi3_carryin_uxtw): Likewise.
17939 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17941 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17942 (*rorsi3_insn_uxtw): Likewise.
17944 2013-03-19 Ian Bolton <ian.bolton@arm.com>
17946 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17947 (*extrsi5_insn_uxtw): Likewise.
17949 2013-03-19 Richard Biener <rguenther@suse.de>
17951 PR tree-optimization/56273
17952 * passes.c (init_optimization_passes): Move second VRP after DOM.
17954 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
17956 * config/i386/i386.md (*movti_internal): Merge from
17957 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17958 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17959 nox64 isa attributes.
17961 2013-03-18 Richard Biener <rguenther@suse.de>
17963 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17965 (merge_node_constraints): Likewise.
17966 (build_succ_graph): Likewise.
17967 (valid_graph_edge): Inline into single caller.
17968 (unify_nodes): Likewise. Use bitmap_set_bit return value
17970 (scc_visit): Fix formatting and variable use.
17971 (do_sd_constraint): Use gcc_checking_assert.
17972 (do_ds_constraint): Likewise.
17973 (do_complex_constraint): Likewise.
17974 (condense_visit): Likewise. Cleanup.
17975 (dump_pred_graph): New function.
17976 (perform_var_substitution): Dump the pred-graph before
17977 variable substitution.
17978 (find_equivalent_node): Use gcc_checking_assert.
17979 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17981 2013-03-18 Richard Biener <rguenther@suse.de>
17983 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17984 Remove cond_expr_stmt_list argument and do not gimplify the
17986 (vect_loop_versioning): Adjust.
17987 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17988 Cleanup to use less temporaries.
17989 (vect_create_data_ref_ptr): Cleanup.
17991 2013-03-18 Jakub Jelinek <jakub@redhat.com>
17993 PR tree-optimization/56635
17994 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17995 require types_compatible_p types.
17997 2013-03-18 Nick Clifton <nickc@redhat.com>
17999 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
18000 spurious backslash.
18002 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
18003 Add missing line to comment describing function.
18005 2013-03-18 Richard Biener <rguenther@suse.de>
18007 PR tree-optimization/56210
18008 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
18009 Handle string / character search functions.
18010 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
18012 2013-03-18 Richard Biener <rguenther@suse.de>
18014 PR middle-end/56483
18015 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
18016 and implement properly.
18017 * gimple.h (gimple_cond_single_var_p): Remove.
18019 2013-03-18 Richard Biener <rguenther@suse.de>
18021 * tree-data-ref.h (find_data_references_in_loop): Declare.
18022 * tree-data-ref.c (get_references_in_stmt): Use a stack
18023 vector pre-allocated in the callers.
18024 (find_data_references_in_stmt): Adjust.
18025 (graphite_find_data_references_in_stmt): Likewise.
18026 (create_rdg_vertices): Likewise.
18027 (find_data_references_in_loop): Export.
18028 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
18029 Compute dependences here...
18030 (vect_analyze_data_refs): ...not here. When we encounter
18031 a non-vectorizable data reference in basic-block vectorization
18032 truncate the data reference vector. Do not bother to
18033 fixup data-dependence information for gather loads.
18034 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
18035 of data references, as reported.
18037 2013-03-18 Richard Biener <rguenther@suse.de>
18039 PR tree-optimization/3713
18040 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
18041 has_constants and expr.
18042 (stmt_has_constants): Properly valueize SSA names when deciding
18043 whether the stmt has constants.
18045 2013-03-18 Richard Biener <rguenther@suse.de>
18047 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
18048 whole function when there is nothing to do.
18049 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
18050 * tree-vectorizer.c (vectorize_loops): Update virtual and
18051 loop-closed SSA once.
18052 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
18054 2013-03-18 Richard Biener <rguenther@suse.de>
18056 PR middle-end/56113
18057 * domwalk.c (bb_postorder): New global static.
18058 (cmp_bb_postorder): New function.
18059 (walk_dominator_tree): Replace scheme imposing an order for
18060 visiting dominator sons by one sorting them at the time they
18061 are pushed on the stack.
18063 2013-03-18 Richard Biener <rguenther@suse.de>
18065 PR tree-optimization/39326
18066 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
18067 (struct mem_ref): Replace mem member with ao_ref typed member.
18068 (MEM_ANALYZABLE): Adjust.
18069 (memref_eq): Likewise.
18070 (mem_ref_alloc): Likewise.
18071 (gather_mem_refs_stmt): Likewise.
18072 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
18073 (execute_sm_if_changed_flag_set): Adjust.
18074 (execute_sm): Likewise.
18075 (ref_always_accessed_p): Likewise.
18076 (refs_independent_p): Likewise.
18077 (can_sm_ref_p): Likewise.
18079 2013-03-18 Jakub Jelinek <jakub@redhat.com>
18082 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
18083 return 1 even for !unsignedp.
18085 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18087 * config/i386/i386.md (isa): Add x64 and nox64.
18088 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
18089 (*pushtf): Enable *roF alternative for x64 isa only.
18090 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
18091 mode attribute of integer alternatives to DImode for TARGET_64BIT.
18092 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
18093 (*movtf_internal): Merge from *movtf_internal_rex64 and
18094 *movtf_internal_sse. Use x64 and nox64 isa attributes.
18095 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
18096 nox64 isa attributes.
18097 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
18098 nox64 isa attributes.
18099 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
18101 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
18103 * config/alpha/alpha.c (TARGET_LRA_P): New define.
18105 2013-03-17 Jakub Jelinek <jakub@redhat.com>
18108 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
18109 class names. Remove trailing comma after "ALL_REGS".
18111 2013-03-16 Jan Hubicka <jh@suse.cz>
18113 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
18114 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
18115 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
18116 of cgraph_get_create_node.
18117 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
18119 2013-03-16 Jason Merrill <jason@redhat.com>
18122 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
18123 with DW_AT_default_value.
18125 2013-03-16 Jakub Jelinek <jakub@redhat.com>
18127 * BASE-VER: Set to 4.9.0.
18129 2013-03-14 Andi Kleen <ak@linux.intel.com>
18132 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
18133 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
18134 Document _x* TSX intrinsics.
18136 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
18137 David Holsgrove <david.holsgrove@xilinx.com>
18139 * configure.ac: Add MicroBlaze TLS support detection.
18140 * configure: Regenerate.
18141 * config/microblaze/microblaze-protos.h
18142 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
18143 symbol_mentioned_p, label_mentioned_p): Add prototypes.
18144 * config/microblaze/microblaze.c (microblaze_address_type): Add
18145 ADDRESS_TLS and tls_reloc address types.
18146 (microblaze_address_info): Add tls_reloc.
18147 (TARGET_HAVE_TLS): Define.
18148 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
18149 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
18150 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
18151 load_tls_operand, microblaze_call_tls_get_addr,
18152 microblaze_legitimize_tls_address): New functions.
18153 (microblaze_classify_unspec): Handle UNSPEC_TLS.
18154 (get_base_reg): Use microblaze_tls_symbol_p.
18155 (microblaze_classify_address): Handle TLS.
18156 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
18157 label_mentioned_p and microblaze_tls_referenced_p.
18158 (microblaze_legitimize_address): Handle TLS.
18159 (microblaze_address_insns): Handle ADDRESS_TLS.
18160 (pic_address_needs_scratch): Handle TLS.
18161 (print_operand_address): Handle TLS.
18162 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
18163 (microblaze_expand_move): Handle TLS.
18164 (microblaze_legitimate_constant_p): Check
18165 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
18166 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
18167 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
18168 (PIC_OFFSET_TABLE_REGNUM): Set.
18169 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
18170 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
18171 (addsi3, movsi_internal2, movdf_internal): Update constraints
18172 * config/microblaze/predicates.md (arith_plus_operand): Define
18173 (move_operand): Redefine as move_src_operand,
18174 check microblaze_tls_referenced_p.
18176 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18178 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
18179 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
18181 2013-03-14 Ian Bolton <ian.bolton@arm.com>
18183 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
18186 2013-03-14 Jakub Jelinek <jakub@redhat.com>
18188 PR tree-optimization/53265
18189 * common.opt (Waggressive-loop-optimizations): New option.
18190 * tree-ssa-loop-niter.c: Include tree-pass.h.
18191 (do_warn_aggressive_loop_optimizations): New function.
18192 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
18193 if number_of_latch_executions returned constant.
18194 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
18195 early. If number_of_latch_executions returned constant, set
18196 nb_iterations_upper_bound back to it.
18197 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
18199 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
18200 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
18202 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
18203 (MULTILIB_OSDIRNAMES): Set.
18204 * genmultilib: If defaultosdirname doesn't start with :: , set
18205 defaultosdirname2 instead, clear it and emit two . multilib_raw
18206 entries instead of just one.
18208 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
18210 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18211 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
18212 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
18213 (SUBTARGET_OVERRIDE_OPTIONS): New.
18215 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
18218 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
18219 (musermode): Convert to Var(TARGET_USERMODE).
18220 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
18221 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
18222 * config/sh/sh.c (sh_option_override): Use
18223 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
18224 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
18226 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
18228 (udivsi3_i4_single, divsi3_i4_single): Use
18229 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
18231 2013-03-13 Dave Korn <dave.korn.cygwin@....>
18233 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
18236 2013-03-13 Richard Biener <rguenther@suse.de>
18238 PR tree-optimization/56608
18239 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
18240 calls when vectorizing basic-blocks.
18242 2013-03-13 Jakub Jelinek <jakub@redhat.com>
18245 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
18248 2013-03-12 Jakub Jelinek <jakub@redhat.com>
18250 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
18252 2013-03-11 Jan Hubicka <jh@suse.cz>
18255 * lto-streamer-out.c (output_symbol_p): Skip references from
18256 constructors of external variables.
18258 2013-03-11 Jan Hubicka <jh@suse.cz>
18260 PR middle-end/56571
18261 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
18263 * emit-rtl.c (verify_rtx_sharing): Likewise.
18264 (copy_insn_1): Likewise.
18265 * rtl.c (copy_rtx): Likewise.
18267 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
18270 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
18271 output_operand_lossage message.
18273 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
18276 * arm.c (shift_op): Validate RTL pattern on the fly.
18277 (arm_print_operand, case 'S'): Don't use shift_operator to validate
18280 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18283 * config/pa/pa.md (call_value): Check for calls to powf and direct to
18284 new call patterns that clobber %fr12.
18285 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
18286 split and postreload patterns.
18287 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
18288 registers %fr12 and %fr12R as call used.
18290 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
18292 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
18293 (canon_address, record_store, replace_read, check_mem_read_rtx,
18294 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
18295 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
18296 rest_of_handle_dse): Likewise.
18298 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
18300 PR middle-end/56524
18301 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
18303 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
18304 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
18305 (save_optabs_if_changed): Replace with...
18306 (init_tree_optimization_optabs): ...this.
18307 * optabs.c (save_optabs_if_changed): Rename to...
18308 (init_tree_optimization_optabs): ...this. Take the optimization node
18309 as argument. Do nothing if the base optabs are already correct.
18310 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
18311 to recompute optabs.
18312 * function.h (function): Remove optabs field.
18313 * function.c (invoke_set_current_function_hook): Call
18314 init_tree_optimization_optabs. Use the result to initialize
18317 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
18319 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
18320 if GTMA_HAS_NO_INSTRUMENTATION.
18321 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
18322 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
18323 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
18324 * gimple-pretty-print.c (dump_gimple_transaction): Handle
18325 GTMA_HAS_NO_INSTRUMENTATION.
18327 2013-03-08 Jakub Jelinek <jakub@redhat.com>
18329 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
18332 2013-03-08 Marek Polacek <polacek@redhat.com>
18333 Jakub Jelinek <jakub@redhat.com>
18335 PR tree-optimization/56478
18336 * predict.c (is_comparison_with_loop_invariant_p): Change the
18337 type of loop_step to tree.
18338 (predict_loops): Adjust.
18339 (predict_iv_comparison): Perform the computations on double_ints.
18341 2013-03-08 Richard Biener <rguenther@suse.de>
18343 PR tree-optimization/56570
18344 * tree-cfg.c (verify_expr_location_1): Verify locations for
18346 * tree-sra.c (create_access_replacement): Strip locations
18347 from DECL_DEBUG_EXPRs.
18349 2013-03-08 Richard Biener <rguenther@suse.de>
18351 * tree-inline.c (expand_call_inline): Do not associate
18352 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
18353 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
18355 2013-03-08 Richard Biener <rguenther@suse.de>
18357 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
18358 or block changes with -Og. Fix for location / block encoding
18359 changes and PHI arguments with locations.
18361 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
18363 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
18365 (struct output_info): Likewise.
18366 (register_overhead): Remove bad gcc_assert.
18367 (bitmap_find_bit): If there is only a single bitmap element, do not
18368 count a miss as a search.
18369 (print_statistics): Update for counter type changes.
18370 (dump_bitmap_statistics): Likewise. Print headers such that they
18371 are properly lined up with the printed counters.
18373 2013-03-07 Jakub Jelinek <jakub@redhat.com>
18375 PR tree-optimization/56559
18376 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
18377 check that it has only a single use.
18379 2013-03-07 Richard Biener <rguenther@suse.de>
18381 * doc/invoke.texi (fwhole-program): Discourage use in combination
18384 2013-03-06 Jakub Jelinek <jakub@redhat.com>
18386 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
18388 PR tree-optimization/56539
18389 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
18390 instead of GSI_CONTINUE_LINKING as last argument to
18391 force_gimple_operand_gsi. Adjust function comment.
18393 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
18396 PR middle-end/56548
18397 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
18398 promoted mode, convert the result back to the original mode.
18400 2013-03-06 Richard Biener <rguenther@suse.de>
18402 PR middle-end/56294
18403 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
18404 (insert_updated_phi_nodes_compare_uids): New function.
18405 (update_ssa): Sort symbols_to_rename after UID before
18406 traversing it to insert PHI nodes.
18408 2013-03-06 Richard Biener <rguenther@suse.de>
18410 PR middle-end/50494
18411 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
18412 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
18415 2013-02-13 Richard Biener <rguenther@suse.de>
18418 * varasm.c (output_constant_def_1): Get the decl representing
18419 the constant as argument.
18420 (output_constant_def): Wrap output_constant_def_1.
18421 (make_decl_rtl): Use output_constant_def_1 with the decl
18422 representing the constant.
18423 (build_constant_desc): Optionally re-use a decl already
18424 representing the constant.
18425 (tree_output_constant_def): Adjust.
18427 2013-03-06 Joey Ye <joey.ye@arm.com>
18430 * gcc.c (convert_white_space): New function.
18431 (main): Handles white space in function name.
18433 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
18436 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
18437 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
18438 to SH_DIV_CALL_TABLE for TARGET_SH2.
18439 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
18441 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
18442 call-table options.
18444 2013-03-05 Sterling Augustine <saugustine@google.com>
18445 Cary Coutant <ccoutant@google.com>
18448 * dwarf2out.c (resolve_addr): Don't call
18449 remove_loc_list_addr_table_entries a second time for the same
18452 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18455 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
18456 (avoid_complex_debug_insns): New function.
18457 (expand_debug_locations): Call it.
18459 PR rtl-optimization/56484
18460 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18461 lifetimes of hard registers on small register class machines.
18463 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18465 * config/microblaze/microblaze-protos.h: Rename
18466 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
18467 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
18469 (microblaze_fast_interrupt_function_p): New function.
18470 (microblaze_is_interrupt_handler): Rename to
18471 microblaze_is_interrupt_variant and add fast_interrupt check.
18472 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18473 (save_restore_insns): Likewise.
18474 (compute_frame_size): Likewise.
18475 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18476 (microblaze_globalize_label): Likewise.
18477 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18478 * config/microblaze/microblaze.md: Use wrapper
18479 microblaze_is_interrupt_variant.
18481 2013-03-05 Kai Tietz <ktietz@redhat.com>
18483 * sdbout.c (sdbout_one_type): Switch to current function's section
18484 supporting cold/hot.
18486 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18488 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18491 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18493 PR middle-end/56461
18494 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18495 if VALGRIND_GET_VBITS is defined, temporarily make object
18496 memory all defined, and restore previous valgrind addressability
18497 and definability afterwards. Free this_object at the end.
18499 PR middle-end/56461
18500 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18501 right before calling lra_create_live_ranges, also call it
18502 when clearing live_p. Only call lra_clear_live_ranges
18503 at the end if live_p.
18505 PR middle-end/56461
18506 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18508 2013-03-05 Richard Biener <rguenther@suse.de>
18510 PR tree-optimization/56521
18511 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18514 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18517 * except.h (remove_unreachable_eh_regions): New prototype.
18518 * except.c (remove_eh_handler_splicer): New function, split out
18519 of remove_eh_handler.
18520 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18521 warning about running it on many EH regions one at a time.
18522 (remove_unreachable_eh_regions_worker): New function, walk the
18523 EH tree in depth-first order and remove non-marked regions.
18524 (remove_unreachable_eh_regions): New function.
18525 * tree-eh.c (mark_reachable_handlers): New function, split out
18526 from remove_unreachable_handlers.
18527 (remove_unreachable_handlers): Use mark_reachable_handlers and
18528 remove_unreachable_eh_regions.
18529 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18530 and remove_unreachable_eh_regions.
18532 2013-03-05 Richard Biener <rguenther@suse.de>
18534 PR middle-end/56525
18535 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18536 not freeing them until the end.
18538 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18540 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18542 2013-03-05 Richard Biener <rguenther@suse.de>
18544 PR tree-optimization/56270
18545 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18546 of loads after scheduling an SLP instance.
18548 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18550 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18552 (check_gcc_parallelize): Run guality.exp as a separate job from
18553 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18554 struct-layout-1.exp with stackalign.exp.
18556 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18558 PR middle-end/56461
18559 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18560 load_index sbitmap even if some bit in it isn't set.
18562 PR middle-end/56461
18563 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18564 (discover_iteration_bound_by_body_walk): Change queues to
18565 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18566 spelling in comment. Call safe_push on queues[bound_index] directly.
18567 Release queues[queue_index] in every iteration unconditionally.
18568 Release bounds vector.
18570 PR middle-end/56461
18571 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18572 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18573 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18576 2013-03-05 Richard Biener <rguenther@suse.de>
18579 * tree-inline.c (remap_blocks_to_null): New function.
18580 (expand_call_inline): When expanding a call stmt without
18581 an associated block inline remap all callee blocks to NULL.
18583 2013-03-05 Jakub Jelinek <jakub@redhat.com>
18585 PR rtl-optimization/56494
18586 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18587 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18588 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18590 PR middle-end/56461
18591 * sel-sched-ir.c (free_sched_pools): Release
18592 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18593 if succs_info_pool.max_top isn't -1.
18596 * opts.c (opts_obstack, opts_concat): Moved to...
18597 * opts-common.c (opts_obstack, opts_concat): ... here.
18599 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18601 PR middle-end/56461
18602 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18604 2013-03-04 Martin Jambor <mjambor@suse.cz>
18606 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18607 all appropriate places.
18609 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18611 PR tree-optimization/56424
18612 * ipa-split.c (split_function): Do not set the RSO flag if result is
18613 not by reference and its type is a register type.
18615 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
18617 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
18618 (microblaze_legitimate_pic_operand): Likewise
18619 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
18620 new function microblaze_legitimate_pic_operand
18621 * config/microblaze/microblaze-protos.h
18622 (microblaze_legitimate_pic_operand): Declare.
18624 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18626 * config/microblaze/predicates.md (call_insn_simple_operand):
18627 New predicate for supported rtx code types.
18628 * config/microblaze/microblaze.md (call_internal1): Use
18629 call_insn_simple_operand predicate.
18631 2013-03-04 Jakub Jelinek <jakub@redhat.com>
18633 PR middle-end/56461
18634 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18635 partitions.ordered_remove.
18637 PR middle-end/56461
18638 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18639 vec_oprnds0.create (1) for modifier == NONE.
18641 PR middle-end/56461
18642 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18643 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18644 vec_oprnds1 right before pushing anything to it for
18647 PR middle-end/56461
18648 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18649 set nbbs to 0 instead of having separate code path.
18650 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18651 instead of false as last argument if returning NULL.
18653 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18655 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18656 the attribute is now called "target" instead of "option".
18657 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18658 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18659 attribute/pragma name for TARGET_OPTION_VALID_P and
18660 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18661 * doc/tm.texi: Regenerated.
18663 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
18665 * config/microblaze/microblaze.c:
18666 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
18667 * config/microblaze/microblaze.h: Add -mxl-reorder to
18669 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18670 instructions emitted if TARGET_REORDER.
18671 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18672 or 0 for -m/-mno case, but initialises as 2 to detect default use case
18675 2013-03-01 Xinliang David Li <davidxl@google.com>
18677 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18680 2013-03-01 Jakub Jelinek <jakub@redhat.com>
18682 PR middle-end/56461
18683 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18684 vector even when returning true. Fix up function comment formatting.
18686 PR middle-end/56461
18687 * ira-build.c (ira_loop_nodes_count): New variable.
18688 (create_loop_tree_nodes): Initialize it.
18689 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18691 PR middle-end/56461
18692 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18693 method on dr_chain and result_chain.
18694 * tree-vect-stmts.c (vectorizable_store): Only call
18695 result_chain.create if j == 0.
18697 PR middle-end/56461
18698 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18699 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18700 before overwriting it.
18702 2013-03-01 Tobias Burnus <burnus@net-b.de>
18704 * doc/extended.texi (C Extensions): Change order in @menu
18706 (Other MIPS Built-in Functions): Move last MIPS entry before
18707 "picoChip Built-in Functions".
18708 (SH Built-in Functions): Move after RX Built-in Functions.
18709 * doc/gcc.texi (Introduction): Change order in @menu
18711 * doc/md.texi (Constraints): Ditto.
18712 * gty.texi (Type Information): Ditto.
18713 (User-provided marking routines for template types): Make
18715 * doc/invoke.texi (AArch64 Options): Move before
18716 "Adapteva Epiphany Options".
18718 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18719 Jakub Jelinek <jakub@redhat.com>
18722 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18723 no_address_safety_analysis attribute.
18724 * doc/extend.texi (no_address_safety_attribute): Rename to
18725 no_sanitize_address attribute, mention no_address_safety_analysis
18726 attribute as deprecated alias.
18728 2013-02-28 Jakub Jelinek <jakub@redhat.com>
18730 PR middle-end/56461
18731 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18732 type to vec<vec<tree> > *.
18733 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18734 to be vec<tree> instead of vec<tree> *, set vec_defs
18735 to vNULL and call vec_defs.create (number_of_vects), adjust other
18737 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18738 vectorizable_condition): Adjust vect_get_slp_defs callers.
18740 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18742 * config/aarch64/aarch64.c
18743 (aarch64_float_const_representable): Remove unused variable.
18745 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18747 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18749 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18751 * config/aarch64/aarch64-builtins.c
18752 (aarch64_init_simd_builtins): Make static.
18754 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18756 * config/aarch64/aarch64.c
18757 (aarch64_simd_make_constant): Make static.
18759 2013-02-28 Martin Jambor <mjambor@suse.cz>
18761 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18762 with no initialization to the RHS of debug statements.
18764 2013-02-28 Martin Jambor <mjambor@suse.cz>
18766 PR tree-optimization/56294
18767 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18769 (get_access_replacement): Do not call create_access_replacement.
18770 Assert a replacement exists.
18771 (get_repl_default_def_ssa_name): Create the replacement declaration
18774 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18776 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18777 final_end_function.
18779 2013-02-28 Marek Polacek <polacek@redhat.com>
18781 PR rtl-optimization/56466
18782 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18783 if we're changing a loop.
18784 (peel_loops_completely): Likewise.
18786 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18789 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18791 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18794 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18795 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18796 INTX_FTYPE_FX, FX_FTYPE_INTX.
18797 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18799 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18801 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18802 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18803 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18804 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18805 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18806 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18807 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18808 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18809 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18810 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18811 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18812 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18813 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18814 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18815 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18816 (avrxmega6): Increase max flash segments from 5 to 6.
18817 * config/avr/t-multilib: Regenerate.
18818 * config/avr/avr-tables.opt: Regenerate.
18819 * doc/avr-mmcu.texi: Regenerate.
18821 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
18823 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18824 (avr_device_to_arch): Rename to avr_device_to_ld.
18825 (avr_device_to_as): New prototype.
18826 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18827 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18828 * config/avr/driver-avr.c (avr_device_to_as): New.
18829 (avr_device_to_arch): Rename to avr_device_to_ld.
18831 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18833 PR middle-end/56461
18834 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18835 method on dr_chain and result_chain.
18837 PR middle-end/56461
18838 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18839 pointer_set_destroy on not_executed_last_iteration.
18841 PR middle-end/56461
18842 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
18844 PR middle-end/56461
18845 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18846 FOR_EACH_DEFINED_FUNCTION when freeing state.
18848 PR middle-end/56461
18849 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18851 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18854 PR middle-end/56461
18855 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18856 known_aggs[i].items and release known_aggs vector.
18858 PR middle-end/56461
18859 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18861 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
18863 * config/microblaze/microblaze.c (microblaze_emit_compare):
18864 Use xor for EQ/NE comparisions.
18865 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18866 (cbranchsf4): Adjust operator to comparison_operator.
18868 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18870 PR middle-end/56461
18871 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18873 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18874 vec_safe_push, always update *slot.
18875 (redirect_edge_var_map_clear): Use vec_free.
18876 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18877 (free_var_map_entry): Use vec_free.
18878 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18879 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18881 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
18883 PR middle-end/45472
18884 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18885 when the may_trap_p bit of the exprs being merged differs.
18886 Reorder tests for speculativeness in the logical and operator.
18888 2013-02-27 Jakub Jelinek <jakub@redhat.com>
18890 * incpath.c (add_standard_paths): Use reconcat instead of concat
18891 where appropriate and avoid leaking memory.
18893 * opts.h: Include obstack.h.
18894 (opts_concat): New prototype.
18895 (opts_obstack): New declaration.
18896 * opts.c (opts_concat): New function.
18897 (opts_obstack): New variable.
18898 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18899 (finish_options): Use opts_concat instead of concat
18900 and XOBNEWVEC instead of XNEWVEC.
18901 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18902 generate_option): Likewise.
18903 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18904 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18907 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18908 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18910 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18912 PR middle-end/56461
18913 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18915 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18917 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18918 (arm_block_move_unaligned_straight): Likewise.
18919 (arm_adjust_block_mem): Likewise.
18921 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18924 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18925 temp, cond and label.
18926 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18929 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18930 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18933 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18934 of prologue/epilogue functions.
18937 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18938 Remove unused variable cfa_offset.
18939 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18942 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18943 type promotion to unsigned.
18946 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18947 for HOST_WIDE_INT of 32 bit / same size as int.
18948 (arm_block_move_unaligned_straight): Likewise.
18949 (arm_adjust_block_mem): Likewise.
18952 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18955 2013-02-26 Marek Polacek <polacek@redhat.com>
18957 PR tree-optimization/56426
18958 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
18960 2013-02-26 Richard Biener <rguenther@suse.de>
18963 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18964 unused variable loops.
18966 2013-02-26 Jakub Jelinek <jakub@redhat.com>
18968 PR tree-optimization/56448
18969 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18970 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18971 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18972 later operands of the references, or even first operand for
18973 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18975 PR tree-optimization/56443
18976 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18977 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18978 to type_for_mode langhook.
18980 2013-02-25 Matt Turner <mattst88@gmail.com>
18982 * doc/invoke.texi: Document r4700.
18984 2013-02-25 Richard Biener <rguenther@suse.de>
18986 PR tree-optimization/56175
18987 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18989 (simplify_bitwise_binary): ... here. Also guard the conversion
18990 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18992 2013-02-25 Catherine Moore <clm@codesourcery.com>
18995 2013-02-24 Catherine Moore <clm@codesourcery.com>
18996 Maciej W. Rozycki <macro@codesourcery.com>
18997 Tom de Vries <tom@codesourcery.com>
18998 Nathan Sidwell <nathan@codesourcery.com>
18999 Iain Sandoe <iain@codesourcery.com>
19000 Nathan Froyd <froydnj@codesourcery.com>
19001 Chao-ying Fu <fu@mips.com>
19003 * doc/extend.texi: (micromips, nomicromips, nocompression):
19004 Document new function attributes.
19005 * doc/invoke.texi (minterlink-compressed, mmicromips,
19006 m14k, m14ke, m14kec): Document new options.
19007 (minterlink-mips16): Update documentation.
19008 * doc/md.texi (ZC, ZD): Document new constraints.
19009 * configure.ac (gcc_cv_as_micromips): Check if linker
19010 supports the .set micromips directive.
19011 * configure: Regenerate.
19012 * config.in: Regenerate.
19013 * config/mips/mips-tables.opt: Regenerate.
19014 * config/mips/micromips.md: New file.
19015 * constraints.md (ZC, AD): New constraints.
19016 * config/mips/predicates.md (movep_src_register): New predicate.
19017 (movep_src_operand): New predicate.
19018 (non_volatile_mem_operand): New predicate.
19019 * config/mips/mips.md (multimem): New type.
19020 (length): Differentiate between 17-bit and 18-bit branch offsets.
19021 (MOVEP1, MOVEP2): New mode iterator.
19022 (mov_<load>l): Use ZC constraint.
19023 (mov_<load>r): Likewise.
19024 (mov_<store>l): Likewise.
19025 (mov_<store>r): Likewise.
19026 (*branch_equality<mode>_inverted): Add microMIPS support.
19027 (*branch_equality<mode>): Likewise.
19028 (*jump_absolute): Likewise.
19029 (indirect_jump_<mode>): Likewise.
19030 (tablejump_<mode>): Likewise.
19031 (<optab>_internal): Likewise.
19032 (sibcall_internal): Likewise.
19033 (sibcall_value_internal): Likewise.
19034 (prefetch): Use constraint ZD.
19035 * config/mips/mips.opt (minterlink-compressed): New option.
19036 (minterlink-mips16): Now an alias for minterlink-compressed.
19037 (mmicromips): New option.
19038 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19039 (compare_and_swap_12): Likewise.
19040 (sync_add<mode>): Likewise.
19041 (sync_<optab>_12): Likewise.
19042 (sync_old_<optab>_12): Likewise.
19043 (sync_new_<optab>_12): Likewise.
19044 (sync_nand_12): Likewise.
19045 (sync_old_nand_12): Likewise.
19046 (sync_new_nand_12): Likewise.
19047 (sync_sub<mode>): Likewise.
19048 (sync_old_add<mode>): Likewise.
19049 (sync_old_sub<mode>): Likewise.
19050 (sync_new_add<mode>): Likewise.
19051 (sync_new_sub<mode>): Likewise.
19052 (sync_<optab><mode>): Likewise.
19053 (sync_old_<optab><mode>): Likewise.
19054 (sync_new_<optab><mode>): Likewise.
19055 (sync_nand<mode>): Likewise.
19056 (sync_old_nand<mode>): Likewise.
19057 (sync_new_nand<mode>): Likewise.
19058 (sync_lock_test_and_set<mode>): Likewise.
19059 (test_and_set_12): Likewise.
19060 (atomic_compare_and_swap<mode>): Likewise.
19061 (atomic_exchange<mode>_llsc): Likewise.
19062 (atomic_fetch_add<mode>_llsc): Likewise.
19063 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19064 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19065 (umips_save_restore_pattern_p): Likewise.
19066 (umips_load_store_pair_p): Likewise.
19067 (umips_output_load_store_pair): Likewise.
19068 (umips_movep_target_p): Likewise.
19069 (umips_12bit_offset_address_p): Likewise.
19070 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19071 (mips_base_mips16): Rename this...
19072 (mips_base_compression_flags): ...to this. Update all uses.
19073 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19074 (mips_mips16_decl_p): Delete.
19075 (mips_nomips16_decl_p): Delete.
19076 (mips_get_compress_on_flags): New function.
19077 (mips_get_compress_off_flags): New function.
19078 (mips_get_compress_mode): New function.
19079 (mips_get_compress_on_name): New function.
19080 (mips_get_compress_off_name): New function.
19081 (mips_insert_attributes): Support multiple compression types.
19082 (mips_merge_decl_attributes): Likewise.
19083 (umips_12bit_offset_address_p): New function.
19084 (mips_start_function_definition): Emit .set micromips directive.
19085 (mips_call_may_need_jalx_p): New function.
19086 (mips_function_ok_for_sibcall): Add microMIPS support.
19087 (mips_print_operand_punctuation): Support short delay slots and
19089 (umips_swm_mask, umips_swm_encoding): New.
19090 (umips_build_save_restore): New function.
19091 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19092 (was_mips16_p): Remove.
19093 (old_compression_mode): New.
19094 (mips_set_compression_mode): New function.
19095 (mips_set_current_function): Add microMIPS support.
19096 (mips_option_override): Likewise.
19097 (umips_save_restore_pattern_p): New function.
19098 (umips_output_save_restore): New function.
19099 (umips_load_store_pair_p_1): New function.
19100 (umips_load_store_pair_p): New function.
19101 (umips_output_load_store_pair_1): New function.
19102 (umips_output_load_store_pair): New function.
19103 (umips_movep_target_p) New function.
19104 (mips_prepare_pch_save): Add microMIPS support.
19105 * config/mips/mips.h (TARGET_COMPRESSION): New.
19106 (TARGET_CPU_CPP_BUILTINS): Update macro
19107 to use new compression flags and to support microMIPS.
19108 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19109 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19110 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19111 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19112 (ASM_SPEC): Support mmicromips and mno-micromips.
19113 (M16STORE_REG_P): New macro.
19114 (MIPS_CALL): Support TARGET_MICROMIPS.
19115 (MICROMIPS_J): New macro.
19116 (mips_base_mips16): Rename this...
19117 (mips_base_compression_flags): ...to this.
19118 (UMIPS_12BIT_OFFSET_P): New macro.
19119 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19120 (MULTILIB_DIRNAMES): Likewise.
19122 2013-02-25 Tom de Vries <tom@codesourcery.com>
19124 PR rtl-optimization/56131
19125 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
19126 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
19127 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
19129 2013-02-25 Tobias Burnus <burnus@net-b.de>
19131 * doc/invoke.texi (-fsanitize=): Move from optimization
19132 to debugging options.
19134 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19136 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
19138 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
19139 Alexander Monakov <amonakov@ispras.ru>
19141 PR middle-end/56077
19142 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
19143 flush pending lists also on non-jumps. Adjust comment.
19145 2013-02-24 Catherine Moore <clm@codesourcery.com>
19146 Maciej W. Rozycki <macro@codesourcery.com>
19147 Tom de Vries <tom@codesourcery.com>
19148 Nathan Sidwell <nathan@codesourcery.com>
19149 Iain Sandoe <iain@codesourcery.com>
19150 Nathan Froyd <froydnj@codesourcery.com>
19151 Chao-ying Fu <fu@mips.com>
19153 * doc/extend.texi: (micromips, nomicromips, nocompression):
19154 Document new function attributes.
19155 * doc/invoke.texi (minterlink-compressed, mmicromips,
19156 m14k, m14ke, m14kec): Document new options.
19157 (minterlink-mips16): Update documentation.
19158 * doc/md.texi (ZC, ZD): Document new constraints.
19159 * configure.ac (gcc_cv_as_micromips): Check if linker
19160 supports the .set micromips directive.
19161 * configure: Regenerate.
19162 * config.in: Regenerate.
19163 * config/mips/mips-tables.opt: Regenerate.
19164 * config/mips/micromips.md: New file.
19165 * constraints.md (ZC, AD): New constraints.
19166 * config/mips/predicates.md (movep_src_register): New predicate.
19167 (movep_src_operand): New predicate.
19168 (non_volatile_mem_operand): New predicate.
19169 * config/mips/mips.md (multimem): New type.
19170 (length): Differentiate between 17-bit and 18-bit branch offsets.
19171 (MOVEP1, MOVEP2): New mode iterator.
19172 (mov_<load>l): Use ZC constraint.
19173 (mov_<load>r): Likewise.
19174 (mov_<store>l): Likewise.
19175 (mov_<store>r): Likewise.
19176 (*branch_equality<mode>_inverted): Add microMIPS support.
19177 (*branch_equality<mode>): Likewise.
19178 (*jump_absolute): Likewise.
19179 (indirect_jump_<mode>): Likewise.
19180 (tablejump_<mode>): Likewise.
19181 (<optab>_internal): Likewise.
19182 (sibcall_internal): Likewise.
19183 (sibcall_value_internal): Likewise.
19184 (prefetch): Use constraint ZD.
19185 * config/mips/mips.opt (minterlink-compressed): New option.
19186 (minterlink-mips16): Now an alias for minterlink-compressed.
19187 (mmicromips): New option.
19188 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
19189 (compare_and_swap_12): Likewise.
19190 (sync_add<mode>): Likewise.
19191 (sync_<optab>_12): Likewise.
19192 (sync_old_<optab>_12): Likewise.
19193 (sync_new_<optab>_12): Likewise.
19194 (sync_nand_12): Likewise.
19195 (sync_old_nand_12): Likewise.
19196 (sync_new_nand_12): Likewise.
19197 (sync_sub<mode>): Likewise.
19198 (sync_old_add<mode>): Likewise.
19199 (sync_old_sub<mode>): Likewise.
19200 (sync_new_add<mode>): Likewise.
19201 (sync_new_sub<mode>): Likewise.
19202 (sync_<optab><mode>): Likewise.
19203 (sync_old_<optab><mode>): Likewise.
19204 (sync_new_<optab><mode>): Likewise.
19205 (sync_nand<mode>): Likewise.
19206 (sync_old_nand<mode>): Likewise.
19207 (sync_new_nand<mode>): Likewise.
19208 (sync_lock_test_and_set<mode>): Likewise.
19209 (test_and_set_12): Likewise.
19210 (atomic_compare_and_swap<mode>): Likewise.
19211 (atomic_exchange<mode>_llsc): Likewise.
19212 (atomic_fetch_add<mode>_llsc): Likewise.
19213 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
19214 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
19215 (umips_save_restore_pattern_p): Likewise.
19216 (umips_load_store_pair_p): Likewise.
19217 (umips_output_load_store_pair): Likewise.
19218 (umips_movep_target_p): Likewise.
19219 (umips_12bit_offset_address_p): Likewise.
19220 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
19221 (mips_base_mips16): Rename this...
19222 (mips_base_compression_flags): ...to this. Update all uses.
19223 (mips_attribute_table): Add micromips, nomicromips and nocompression.
19224 (mips_mips16_decl_p): Delete.
19225 (mips_nomips16_decl_p): Delete.
19226 (mips_get_compress_on_flags): New function.
19227 (mips_get_compress_off_flags): New function.
19228 (mips_get_compress_mode): New function.
19229 (mips_get_compress_on_name): New function.
19230 (mips_get_compress_off_name): New function.
19231 (mips_insert_attributes): Support multiple compression types.
19232 (mips_merge_decl_attributes): Likewise.
19233 (umips_12bit_offset_address_p): New function.
19234 (mips_start_function_definition): Emit .set micromips directive.
19235 (mips_call_may_need_jalx_p): New function.
19236 (mips_function_ok_for_sibcall): Add microMIPS support.
19237 (mips_print_operand_punctuation): Support short delay slots and
19239 (umips_swm_mask, umips_swm_encoding): New.
19240 (umips_build_save_restore): New function.
19241 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
19242 (was_mips16_p): Remove.
19243 (old_compression_mode): New.
19244 (mips_set_compression_mode): New function.
19245 (mips_set_current_function): Add microMIPS support.
19246 (mips_option_override): Likewise.
19247 (umips_save_restore_pattern_p): New function.
19248 (umips_output_save_restore): New function.
19249 (umips_load_store_pair_p_1): New function.
19250 (umips_load_store_pair_p): New function.
19251 (umips_output_load_store_pair_1): New function.
19252 (umips_output_load_store_pair): New function.
19253 (umips_movep_target_p) New function.
19254 (mips_prepare_pch_save): Add microMIPS support.
19255 * config/mips/mips.h (TARGET_COMPRESSION): New.
19256 (TARGET_CPU_CPP_BUILTINS): Update macro
19257 to use new compression flags and to support microMIPS.
19258 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
19259 (MIPS_ARCH_FLOAT_SPEC): Likewise.
19260 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19261 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19262 (ASM_SPEC): Support mmicromips and mno-micromips.
19263 (M16STORE_REG_P): New macro.
19264 (MIPS_CALL): Support TARGET_MICROMIPS.
19265 (MICROMIPS_J): New macro.
19266 (mips_base_mips16): Rename this...
19267 (mips_base_compression_flags): ...to this.
19268 (UMIPS_12BIT_OFFSET_P): New macro.
19269 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19270 (MULTILIB_DIRNAMES): Likewise.
19272 2013-02-24 Jakub Jelinek <jakub@redhat.com>
19275 * target-globals.c (save_target_globals): For init_reg_sets and
19276 target_reinit remporarily set this_fn_optabs to this_target_optabs.
19278 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
19280 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
19281 * config/aarch64/t-aarch64
19282 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
19284 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
19286 PR inline-asm/56148
19287 * lra-constraints.c (process_alt_operands): Reload operand
19288 conflicting with earlier clobber only if no more other conflicting
19291 2013-02-22 Jakub Jelinek <jakub@redhat.com>
19294 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
19295 if not linking a shared library.
19297 2013-02-22 Seth LaForge <sethml@google.com>
19299 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
19301 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
19303 * config/arm/arm.md (split for extendsidi): Update condition.
19304 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
19305 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
19306 (qhs_zextenddi_cstr): Likewise.
19308 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19310 PR middle-end/56420
19311 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
19312 avoid signed wrapping.
19313 (expand_mult): Handle properly multiplication by
19314 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
19315 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
19316 in the compiler if coeff is HOST_WIDE_INT_MIN.
19317 (expand_divmod): Don't make ext_op1 static, change it's type to
19318 uhwi. Avoid undefined behavior in -INTVAL (op1).
19320 PR rtl-optimization/50339
19321 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
19323 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
19324 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
19325 into splitting_ashiftrt field.
19326 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
19328 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
19331 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
19333 PR middle-end/56108
19334 * trans-mem.c (execute_tm_mark): Do not expand transactions that
19335 are sure to go irrevocable.
19337 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
19339 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
19340 scalars are valid operands.
19342 2013-02-21 Martin Jambor <mjambor@suse.cz>
19344 PR tree-optimization/56310
19345 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
19346 only matching indices and non-negative final offsets.
19347 (intersect_aggregates_with_edge): Pass src_idx to
19348 agg_replacements_to_vector. Pass src_idx insstead of index to
19349 intersect_with_agg_replacements.
19351 2013-02-21 Martin Jambor <mjambor@suse.cz>
19353 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
19354 instead of hard-wired defaults.
19356 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
19358 * doc/invoke.texi (MIPS Options): Update documentation of the
19359 floating-point multiply-accumulate instruction restrictions.
19361 2013-02-21 Kostya Serebryany <kcc@google.com>
19363 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
19364 asan_shadow_offset on x86_64 linux.
19366 2013-02-21 Richard Biener <rguenther@suse.de>
19368 PR tree-optimization/56415
19370 2013-02-11 Richard Biener <rguenther@suse.de>
19372 PR tree-optimization/56273
19373 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19376 2013-02-21 Jakub Jelinek <jakub@redhat.com>
19379 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
19382 PR inline-asm/56405
19383 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
19384 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
19386 2013-02-20 Jan Hubicka <jh@suse.cz>
19388 PR tree-optimization/56265
19389 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
19390 when target is referenced for first time.
19392 2013-02-20 Richard Biener <rguenther@suse.de>
19394 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
19395 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
19396 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
19397 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
19398 not return anything.
19399 (rename_ssa_copies): Do not remove unused locals.
19400 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
19401 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
19402 * passes.c (execute_function_todo): Do not schedule unused locals
19403 removal if cleanup_tree_cfg did something.
19404 * tree-ssa-live.c (remove_unused_locals): Dump statistics
19405 about the number of removed locals.
19407 2013-02-20 Richard Biener <rguenther@suse.de>
19409 PR tree-optimization/56398
19410 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
19412 2013-02-20 Martin Jambor <mjambor@suse.cz>
19414 PR tree-optimization/55334
19415 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
19416 restricted pointers to arrays.
19418 2013-02-20 Richard Biener <rguenther@suse.de>
19419 Jakub Jelinek <jakub@redhat.com>
19421 PR tree-optimization/56396
19422 * tree-ssa-ccp.c (n_const_val): New static variable.
19423 (get_value): Return NULL for SSA names we don't have a lattice
19425 (ccp_initialize): Initialize n_const_val.
19426 * tree-ssa-copy.c (n_copy_of): New static variable.
19427 (init_copy_prop): Initialize n_copy_of.
19428 (get_value): Return NULL_TREE for SSA names we don't have a
19431 2013-02-20 Martin Jambor <mjambor@suse.cz>
19433 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
19435 2013-02-20 Richard Biener <rguenther@suse.de>
19437 * genpreds.c (write_lookup_constraint): Do not compare first
19438 letter of the constraint again.
19440 2013-02-20 Richard Biener <rguenther@suse.de>
19442 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
19444 (get_use_iv_cost): Terminate hashtable walk when coming across
19447 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
19449 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
19450 reassociation for avx2 targets.
19452 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19454 * config/microblaze/microblaze.c: microblaze_has_clz = 0
19455 Add version check for v8.10.a to enable microblaze_has_clz
19456 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
19457 version and TARGET_PATTERN_COMPARE check
19458 * config/microblaze/microblaze.md: New clzsi2 instruction
19460 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19462 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
19463 function before branching.
19465 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19467 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19469 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19471 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
19473 PR middle-end/55889
19474 * sel-sched.c: Include ira.h.
19475 (implicit_clobber_conflict_p): New function.
19476 (moveup_expr): Use it.
19477 * Makefile.in (sel-sched.o): Depend on ira.h.
19479 2013-02-19 Richard Biener <rguenther@suse.de>
19481 PR tree-optimization/56384
19482 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19483 (vn_hash_type): Split out from ...
19484 (vn_hash_constant_with_type): ... here.
19485 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19486 (vn_phi_eq): Compare types from vn_phi_s structure.
19487 (vn_phi_lookup): Populate vn_phi_s type.
19488 (vn_phi_insert): Likewise.
19490 2013-02-19 Jakub Jelinek <jakub@redhat.com>
19492 PR tree-optimization/56350
19493 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19494 if haven't found reduction or nested cycle operand, rather than
19495 asserting we must find it.
19497 PR tree-optimization/56381
19498 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19501 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
19502 Jakub Jelinek <jakub@redhat.com>
19505 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19506 (swap_optab_enable): Same.
19507 (init_all_optabs): Use argument instead of global.
19508 * tree.h (struct tree_optimization_option): New field target_optabs.
19509 * expr.h (init_all_optabs): Add argument to prototype.
19510 (TREE_OPTIMIZATION_OPTABS): New.
19511 (save_optabs_if_changed): Protoize.
19512 * optabs.h: Declare this_fn_optabs.
19513 * optabs.c (save_optabs_if_changed): New.
19514 Declare this_fn_optabs.
19515 (init_optabs): Add argument to init_all_optabs() call.
19516 * function.c (invoke_set_current_function_hook): Handle per
19518 * function.h (struct function): New field optabs.
19519 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19520 optimization_current_node has changed.
19521 * target-globals.h (save_target_globals_default_opts): Protoize.
19522 * target-globals.c (save_target_globals_default_opts): New.
19524 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19527 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19528 registers %fr12 and %fr12R as call used.
19531 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
19532 and HImode, require all displacements to be an integer multiple of
19534 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19535 only allow QImode and HImode when reload is in progress and strict is
19536 true. Likewise for symbolic addresses. Use base14_operand to check
19537 displacements in REG+BASE addresses.
19539 2013-02-18 Richard Biener <rguenther@suse.de>
19541 PR tree-optimization/56366
19542 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19543 handle sign-conversion of outer-loop initial induction value.
19545 2013-02-18 Richard Biener <rguenther@suse.de>
19547 PR middle-end/56349
19548 * cfghooks.c (merge_blocks): If we merge a latch into another
19549 block adjust references to it.
19550 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19551 (verify_loop_structure): Verify that a recorded latch is in fact
19554 2013-02-18 Richard Biener <rguenther@suse.de>
19556 PR tree-optimization/56321
19557 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19558 order SSA name release and virtual operand unlinking.
19560 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19562 * config/microblaze/microblaze.md (save_stack_block): Define.
19563 (restore_stack_block): Likewise.
19565 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19567 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19568 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19569 * config/microblaze/microblaze.c (microblaze_option_override):
19570 Bail out early for PIC modes when target does not support PIC.
19572 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19574 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
19575 Replace with a microblaze version.
19576 (microblaze_trampoline_init): Adapt for microblaze.
19577 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
19580 2013-02-16 Jakub Jelinek <jakub@redhat.com>
19581 Dodji Seketeli <dodji@redhat.com>
19584 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
19585 (instrument_mem_region_access): Do not forget to always put
19586 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19587 0) statement, even for cases where either 'base' or 'base + len'
19588 are not instrumented -- because they have been previously
19589 instrumented. Simplify the logic by putting all the statements
19590 instrument 'base + len' inside a sequence, and then insert that
19591 sequence right before the current insertion point. Then, to
19592 instrument 'base + len', just get an iterator on that statement.
19593 And do not forget to update the pointer to iterator the function
19594 received as argument.
19596 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19598 PR rtl-optimization/56348
19599 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19601 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19603 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19604 (clean_graph_dump_file): Pass base to start_graph_dump.
19606 2013-02-14 Richard Henderson <rth@redhat.com>
19609 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19611 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19613 * collect2-aix.h: Define F_LOADONLY.
19615 2013-02-14 Richard Biener <rguenther@suse.de>
19618 * varasm.c (output_constant_def_1): Get the decl representing
19619 the constant as argument.
19620 (output_constant_def): Wrap output_constant_def_1.
19621 (make_decl_rtl): Use output_constant_def_1 with the decl
19622 representing the constant.
19623 (build_constant_desc): Optionally re-use a decl already
19624 representing the constant.
19625 (tree_output_constant_def): Adjust.
19627 2013-02-14 Dodji Seketeli <dodji@redhat.com>
19630 * asan.c (instrument_builtin_call): Really put the length of the
19631 second source argument into src1_len.
19633 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19635 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19636 argument. If it is false, don't create edge from then_bb to
19638 (insert_if_then_before_iter): Pass true to it.
19639 (build_check_stmt): Pass false to it.
19640 (transform_statements): Flush hash table only on extended basic
19641 block boundaries, rather than at the beginning of every bb.
19642 Don't flush hash table on nonfreeing_call_p calls.
19643 * tree-flow.h (nonfreeing_call_p): New prototype.
19644 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19646 2013-02-13 David S. Miller <davem@davemloft.net>
19648 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19650 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19653 * ira.c (max_regno_before_ira): Move from ...
19655 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19658 2013-02-13 Jakub Jelinek <jakub@redhat.com>
19660 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19662 2013-02-13 Richard Biener <rguenther@suse.de>
19665 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19666 globals in MEM_REFs.
19668 2013-02-13 Richard Biener <rguenther@suse.de>
19670 * loop-init.c (loop_optimizer_init): Clear loop state when
19671 re-initializing preserved loops.
19672 * loop-unswitch.c (unswitch_single_loop): Return whether
19673 we unswitched the loop. Do not verify loop state here.
19674 (unswitch_loops): When we unswitched a loop discover new loops.
19676 2013-02-13 Kostya Serebryany <kcc@google.com>
19678 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19680 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19682 2013-02-12 Dodji Seketeli <dodji@redhat.com>
19684 Avoid instrumenting duplicated memory access in the same basic block
19685 * Makefile.in (asan.o): Add new dependency on hash-table.h
19686 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19687 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19688 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19689 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19690 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19691 (get_mem_ref_of_assignment): New functions.
19692 (get_mem_refs_of_builtin_call): Extract from
19693 instrument_builtin_call and tweak a little bit to make it fit with
19695 (instrument_builtin_call): Use the new
19696 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19697 of is_gimple_builtin_call.
19698 (instrument_derefs, instrument_mem_region_access): Insert the
19699 instrumented memory reference into the hash table.
19700 (maybe_instrument_assignment): Renamed instrument_assignment into
19701 this, and change it to advance the iterator when instrumentation
19702 actually happened and return true in that case. This makes it
19703 homogeneous with maybe_instrument_assignment, and thus give a
19704 chance to callers to be more 'regular'.
19705 (transform_statements): Clear the memory reference hash table
19706 whenever we enter a new BB, when we cross a function call, or when
19707 we are done transforming statements. Use
19708 maybe_instrument_assignment instead of instrumentation. No more
19709 need to special case maybe_instrument_assignment and advance the
19710 iterator after calling it; it's now handled just like
19711 maybe_instrument_call. Update comment.
19713 2013-02-13 Richard Biener <rguenther@suse.de>
19715 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19716 Fix loop discovery code.
19718 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19720 PR inline-asm/56148
19721 * lra-constraints.c (process_alt_operands): Match early clobber
19722 operand with itself. Check conflicts with earlyclobber only if
19723 the operand is not reloaded. Prefer to reload conflicting operand
19724 if earlyclobber and matching operands are the same.
19726 2013-02-12 Richard Biener <rguenther@suse.de>
19729 * lto-streamer-out.c (write_symbol): Do not output symbols
19730 for hard register variables.
19732 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
19735 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19736 (umulsidi3_insn, mulsidi3_insn): New insns.
19738 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
19740 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19741 (struct tune_params): Add vec_costs field.
19742 * config/arm/arm.c (arm_builtin_vectorization_cost)
19743 (arm_add_stmt_cost): New functions.
19744 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19745 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19746 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19747 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19748 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19749 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19750 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19752 2013-02-12 Richard Biener <rguenther@suse.de>
19755 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19756 decls again if possible.
19758 2013-02-12 Richard Biener <rguenther@suse.de>
19760 PR middle-end/56288
19761 * tree-ssa.c (verify_ssa_name): Fix check, move
19762 SSA_NAME_IN_FREE_LIST check up.
19764 2013-02-12 Jakub Jelinek <jakub@redhat.com>
19765 Steven Bosscher <steven@gcc.gnu.org>
19767 PR rtl-optimization/56151
19768 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19769 equal to op0 or op1, and last_insn pattern is CODE operation
19770 with MEM dest and one of the operands matches that MEM.
19772 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19774 * doc/extend.texi: Document Function Multiversioning and "default"
19775 parameter string to target attribute.
19776 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19777 target attribute parameter is "default".
19778 (ix86_compare_version_priority): Remove checks for target attribute.
19779 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19780 Remove check for target attribute equal to NULL. Add assert.
19781 (ix86_generate_version_dispatcher_body): Change error to sorry.
19783 2013-02-11 Iain Sandoe <iain@codesourcery.com>
19784 Jack Howarth <howarth@bromo.med.uc.edu>
19785 Patrick Marlier <patrick.marlier@gmail.com>
19788 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19789 define ENDFILE_SPEC as TM_DESTRUCTOR.
19790 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19792 2013-02-11 Alexander Potapenko <glider@google.com>
19793 Jack Howarth <howarth@bromo.med.uc.edu>
19794 Jakub Jelinek <jakub@redhat.com>
19797 * config/darwin.c (cdtor_record): Rename ctor_record.
19798 (sort_cdtor_records): Rename sort_ctor_records.
19799 (finalize_dtors): New routine to sort destructors by
19800 priority before use in assemble_integer.
19801 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19803 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
19805 PR rtl-optimization/56275
19806 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19807 offset is non-negative and less than cmode size before
19808 calling simplify_subreg.
19810 2013-02-11 Richard Biener <rguenther@suse.de>
19812 PR tree-optimization/56264
19813 * cfgloop.h (fix_loop_structure): Adjust prototype.
19814 * loop-init.c (fix_loop_structure): Return the number of
19815 newly discovered loops.
19816 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19817 are discovered, do a full loop-closed SSA rewrite.
19819 2013-02-11 Richard Biener <rguenther@suse.de>
19821 PR tree-optimization/56273
19822 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19824 (check_array_ref): Fix missing newline in dumps.
19825 (search_for_addr_array): Likewise.
19827 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
19829 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19831 2013-02-09 Jakub Jelinek <jakub@redhat.com>
19834 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19836 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19838 PR rtl-optimization/56246
19839 * lra-constraints.c (simplify_operand_subreg): Try to reuse
19841 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19842 constraints are satisfied.
19844 2013-02-08 Jeff Law <law@redhat.com>
19847 * emit-rtl.c (reg_is_parm_p): New function.
19848 * regs.h (reg_is_parm_p): New prototype.
19849 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19850 callee-clobbered registers.
19852 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19855 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19856 If there is no implicit builtin declaration, just return NULL.
19858 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19860 * config/i386/sse.md (FMAMODEM): New mode iterator.
19861 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19862 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19864 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19866 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19867 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19868 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19870 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19872 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19873 (microblaze*-*-elf): Likewise.
19874 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19876 * config/microblaze/microblaze-c.c: Add builtin defines for
19877 _LITTLE_ENDIAN and _BIG_ENDIAN.
19878 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19879 add to TARGET_DEFAULT flags.
19880 Expand ASM_SPEC and LINK_SPEC.
19881 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19882 * config/microblaze/microblaze.md: Update extendsidi2 and
19883 movdi_internal instructions to use low-order / high-order reg
19885 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19886 options and inversemask / mask of LITTLE_ENDIAN.
19887 * config/microblaze/t-microblaze: Expand multilib options to
19888 include mlittle-endian (le) and update exceptions patterns.
19890 2013-02-08 Jakub Jelinek <jakub@redhat.com>
19892 PR rtl-optimization/56195
19893 * lra-constraints.c (get_reload_reg): Don't reuse regs
19894 if they have smaller mode than requested, if they have
19895 wider mode than requested, try to return a SUBREG.
19897 PR tree-optimization/56250
19898 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19899 if type is unsigned and code isn't MULT_EXPR.
19901 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19903 PR tree-optimization/56064
19904 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19905 bits according to mode.
19906 * fixed-value.h (fixed_from_double_int)
19907 (const_fixed_from_double_int): Adjust comments.
19909 2013-02-08 Richard Biener <rguenther@suse.de>
19912 * lto-streamer.h (struct data_in): Remove current_file, current_line
19913 and current_col members.
19914 * lto-streamer-out.c (lto_output_location): Stream changed bits
19915 en-block for efficiency.
19916 * lto-streamer-in.c (clear_line_info): Remove.
19917 (lto_input_location): Cache current file, line and column
19918 globally via local statics. Read changed bits en-block.
19919 (input_function): Do not call clear_line_info.
19920 (lto_read_body): Likewise.
19921 (lto_input_toplevel_asms): Likewise.
19923 2013-02-08 Michael Matz <matz@suse.de>
19925 PR tree-optimization/52448
19926 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19927 (nt_call_phase): New static.
19928 (add_or_mark_expr): Only mark accesses with newer phase than any
19930 (nonfreeing_call_p): New.
19931 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19932 (nt_fini_block): Keep blocks marked as visited.
19933 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19935 2013-02-08 Richard Biener <rguenther@suse.de>
19937 * ira.c (ira): Free broken dominator information.
19939 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
19941 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19943 2013-02-08 Marek Polacek <polacek@redhat.com>
19945 * cfgloop.c (verify_loop_structure): Add more checking of headers.
19947 2013-02-08 Richard Biener <rguenther@suse.de>
19949 PR middle-end/56181
19950 * cfgloop.h (flow_loops_find): Adjust.
19951 (bb_loop_header_p): Declare.
19952 * cfgloop.c (bb_loop_header_p): New function split out from ...
19953 (flow_loops_find): ... here. Adjust function signature,
19954 support incremental loop structure update.
19955 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19956 * cfgloopmanip.c (fix_loop_structure): Move ...
19957 * loop-init.c (fix_loop_structure): ... here.
19958 (apply_loop_flags): Split out from ...
19959 (loop_optimizer_init): ... here.
19960 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19961 in incremental mode, only remove dead loops here.
19963 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
19966 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19967 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19968 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19969 (*round<mode>3.libgcc): New insns for fixed-modes.
19970 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19971 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19972 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19973 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19974 implementations. Define to __builtin_avr_absFX,
19975 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19976 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19977 __builtin_avr_countlsFX, respectively.
19978 * config/avr/avr-c.c (target.h): Include it.
19979 (enum avr_builtin_id): New enum.
19980 (avr_resolve_overloaded_builtin): New static function.
19981 (avr_register_target_pragmas): Use it to set
19982 targetm.resolve_overloaded_builtin.
19983 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19984 tree nodes used by DEF_BUILTIN.
19985 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19986 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19987 <AVR_BUILTIN_xxBITS>: Same.
19989 2013-02-08 Richard Biener <rguenther@suse.de>
19991 * cfgloop.c (verify_loop_structure): Properly handle
19992 a loop exiting to another loop header.
19993 * ira-int.h (ira_loops): Remove.
19994 * ira.c (ira_loops): Remove.
19995 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19996 (do_reload): Use loop_optimizer_finalize.
19997 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19998 number_of_loops to access the loop tree.
19999 (more_one_region_p): Likewise.
20000 (finish_loop_tree_nodes): Likewise.
20001 (rebuild_regno_allocno_maps): Likewise.
20002 (mark_loops_for_removal): Likewise.
20003 (mark_all_loops_for_removal): Likewise.
20004 (remove_unnecessary_regions): Likewise.
20005 (ira_build): Likewise.
20006 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
20008 2013-02-08 Richard Biener <rguenther@suse.de>
20010 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
20011 * ipa-pure-const.c (analyze_function): Avoid calling
20012 mark_irreducible_loops twice.
20013 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
20015 2013-02-07 David S. Miller <davem@davemloft.net>
20017 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
20019 * var-tracking.c (vt_add_function_parameter): Test the presence of
20020 HAVE_window_save properly and do not remap argument registers when
20021 we have a leaf function.
20023 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
20026 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
20028 * config/i386/i386.c (ix86_print_operand): Ditto.
20030 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20032 * lra-constraints.c (process_alt_operands): Fix recently added comment.
20034 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
20036 PR rtl-optimization/56225
20037 * lra-constraints.c (process_alt_operands): Check that reload hard
20038 reg can hold value for strict_low_part.
20040 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20043 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
20044 dwarf2out_end_function.
20045 (in_first_function_p, maybe_at_text_label_p,
20046 first_loclabel_num_not_at_text_label): New variables.
20047 (dwarf2out_var_location): In the first function find out
20048 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
20049 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
20052 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
20054 PR rtl-optimization/56178
20055 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
20056 SUBREG of a register. Tidy up related block of code.
20057 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
20058 note if the source is a register or a SUBREG of a register.
20060 2013-02-07 Jakub Jelinek <jakub@redhat.com>
20063 * config/rs6000/rs6000.md (ptrm): New mode attr.
20064 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
20065 call_value_indirect_aix<pttrsize>,
20066 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
20069 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20071 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
20072 if -bnortl. Convert to strcmp and strncmp.
20074 2013-02-07 Alan Modra <amodra@gmail.com>
20077 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
20078 addresses won't wrap when offsetting.
20079 (rs6000_secondary_reload): Provide secondary reloads needed for
20080 wrapping LO_SUM addresses.
20082 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
20084 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
20085 MACH, just __MACH__.
20087 2013-02-06 Richard Biener <rguenther@suse.de>
20089 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
20090 instead of calling fix_loop_structure.
20092 2013-02-06 Jakub Jelinek <jakub@redhat.com>
20094 PR middle-end/56217
20095 * omp-low.c (use_pointer_for_field): Return false if
20096 lower_send_shared_vars doesn't generate any copy-out code.
20098 2013-02-06 Tom de Vries <tom@codesourcery.com>
20100 PR rtl-optimization/56131
20101 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
20102 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
20103 of the label is NULL. Add comment.
20105 2013-02-05 Jakub Jelinek <jakub@redhat.com>
20107 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
20110 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
20111 (STATIC_LIBTSAN_LIBS): Likewise.
20112 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
20113 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
20114 is defined, don't add anything else beyond that.
20115 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
20116 (LINK_COMMAND_SPEC): Use them.
20118 PR tree-optimization/56205
20119 * tree-stdarg.c (check_all_va_list_escapes): Return true if
20120 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
20121 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
20123 2013-02-05 Richard Biener <rguenther@suse.de>
20125 PR tree-optimization/53342
20126 PR tree-optimization/53185
20127 * tree-vectorizer.h (vect_check_strided_load): Remove.
20128 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
20129 not disallow peeling for vectorized strided loads.
20130 (vect_check_strided_load): Make static and simplify.
20131 (vect_analyze_data_refs): Adjust.
20132 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
20133 correctly when vectorizing strided loads.
20135 2013-02-05 Richard Biener <rguenther@suse.de>
20137 * doc/install.texi: Refer to ISL, not PPL.
20139 2013-02-05 Jan Hubicka <jh@suse.cz>
20141 PR tree-optimization/55789
20142 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
20144 2013-02-05 Jan Hubicka <jh@suse.cz>
20146 PR tree-optimization/55789
20147 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
20148 the dead call anyway.
20150 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
20153 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
20155 2013-02-04 Alexander Potapenko <glider@google.com>
20156 Jack Howarth <howarth@bromo.med.uc.edu>
20157 Jakub Jelinek <jakub@redhat.com>
20160 * config/darwin.c (sort_ctor_records): Stabilized qsort
20161 on constructor priority by using original position.
20162 (finalize_ctors): New routine to sort constructors by
20163 priority before use in assemble_integer.
20164 (machopic_asm_out_constructor): Use finalize_ctors if needed.
20166 2013-02-04 Jakub Jelinek <jakub@redhat.com>
20169 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
20170 about visibility on artificial decls.
20171 * config/sol2.c (solaris_assemble_visibility): Likewise.
20173 2013-02-04 Kai Tietz <ktietz@redhat.com>
20176 * config/i386/i386.c (function_value_ms_64): Add additional valtype
20177 argument and improve checking of return-argument types for 16-byte
20179 (ix86_function_value_1): Add additional valtype argument on call
20180 of function_value_64.
20181 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
20182 handling infunction_value_64 function.
20184 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
20186 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
20188 2013-02-04 Richard Biener <rguenther@suse.de>
20190 PR tree-optimization/56188
20191 * tree-ssa-structalias.c (label_visit): Consider case with
20192 initially non-empty points-to set.
20193 (perform_var_substitution): Dump node mapping and clean up.
20195 2013-02-04 Richard Guenther <rguenther@suse.de>
20198 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
20199 node prevail as last resort.
20200 (lto_symtab_merge_decls): Remove guard on LTRANS here.
20201 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
20203 2013-02-04 Richard Biener <rguenther@suse.de>
20205 PR tree-optimization/56113
20206 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
20208 (equiv_class_lookup_or_add): ... this.
20209 (label_visit): Adjust and fix error in previous patch.
20210 (perform_var_substitution): Adjust.
20212 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
20214 * config/sh/divtab.c: Fix formatting and comments throughout the file.
20215 * config/sh/sh4-300.md: Likewise.
20216 * config/sh/sh4a.md: Likewise.
20217 * config/sh/constraints.md: Likewise.
20218 * config/sh/sh.md: Likewise.
20219 * config/sh/netbsd-elf.h: Likewise.
20220 * config/sh/predicates.md: Likewise.
20221 * config/sh/sh-protos.h: Likewise.
20222 * config/sh/ushmedia.h: Likewise.
20223 * config/sh/linux.h: Likewise.
20224 * config/sh/sh.c: Likewise.
20225 * config/sh/superh.h: Likewise.
20226 * config/sh/elf.h: Likewise.
20227 * config/sh/sh4.md: Likewise.
20228 * config/sh/sh.h: Likewise.
20230 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20232 * config/pa/constraints.md: Adjust unused letters. Change "T"
20233 constraint to match_test floating_point_store_memory_operand().
20234 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
20235 (base14_operand): New.
20236 (floating_point_store_memory_operand): New.
20237 (integer_store_memory_operand): Revise to use base14_operand and
20238 reg_plus_base_memory_operand.
20239 (move_dest_operand): Allow symbolic_memory_operands.
20240 (symbolic_memory_operand): Check for LO_SOM.
20241 (symbolic_operand): Change default case to break.
20242 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
20243 CONST_DOUBLE values to be reloaded by putting them into memory when
20244 the destination is a floating point register.
20245 (movdf): Remove code to handle CONST_DOUBLE.
20247 (reload_indf_r1): New.
20248 (reload_insf_r1): New.
20249 Consistently use "Q" and "T" constraints with integer and floating
20250 point move instructions, respectively.
20251 (movdi): Remove FAIL.
20252 Change predicate for source operand unamed DImode move from
20253 general_operand to move_src_operand.
20254 (umulsidi3): Change predicate for destination operand to
20256 Likewise for similar unamed patterns.
20257 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
20258 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
20259 (hppa_legitimize_address): Simplify mask calculation.
20260 (pa_emit_move_sequence): Revised handling of secondary reloads from
20261 REG+D addresses for floating point loads and stores. Directly handle
20262 loading CONST0_RTX (mode) to a floating point register.
20263 (pa_secondary_reload): Handle reloading DF and SFmode constant values
20264 to floating point registers. Don't restrict secondary reloads to
20265 floating point registers to integer modes. Revise some comments and
20267 (TARGET_LEGITIMATE_ADDRESS_P): Define.
20268 (pa_legitimate_address_p): New.
20269 (pa_legitimize_reload_address): New.
20270 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
20271 (STRICT_REG_OK_FOR_BASE_P): New.
20272 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
20273 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
20275 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
20276 Andrew Dixie <andrewd@gentrack.com>
20278 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
20281 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
20283 * expmed.c (extract_bit_field_1): Pass the full width of the
20284 structure to get_best_reg_extraction_insn.
20286 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20289 * configure.ac (use_cxa_atexit): Add AIX.
20290 * configure: Regenerate.
20292 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
20294 2013-02-01 Jakub Jelinek <jakub@redhat.com>
20297 * final.c (need_profile_function): New variable.
20298 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
20299 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
20300 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
20301 notes, targetm.asm_out.function_prologue doesn't emit anything,
20302 HAVE_prologue and profiler should be emitted before prologue,
20303 set need_profile_function instead of emitting it.
20304 (final_scan_insn): If need_profile_function, emit
20305 profile_function on the first NOTE_INSN_BASIC_BLOCK or
20306 NOTE_INSN_FUNCTION_BEG note.
20308 2013-02-01 Richard Henderson <rth@redhat.com>
20310 * config/rs6000/rs6000.md (smulditi3): New.
20313 * config/alpha/alpha.md (umulditi3): New.
20315 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
20317 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
20318 (ASM_OUTPUT_ALIGNED_LOCAL): New.
20320 2013-02-01 Richard Biener <rguenther@suse.de>
20322 PR tree-optimization/56113
20323 * tree-ssa-structalias.c (label_visit): Reduce work for
20324 single-predecessor nodes.
20326 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
20328 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
20329 range isn't testing for zero.
20331 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
20333 PR middle-end/56113
20334 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
20336 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
20337 Nick Clifton <nickc@redhat.com>
20339 * config/v850/constraints.md (Q): Define as a memory constraint.
20340 * config/v850/predicates.md (label_ref_operand): New predicate.
20341 (e3v5_shift_operand): New predicate.
20342 (ior_operator): New predicate.
20343 * config/v850/t-v850: Add e3v5 multilib.
20344 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
20345 (v850_gen_movdi): Prototype.
20346 * config/v850/v850.c: Add support for e3v5 architecture.
20347 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
20349 (construct_save_jarl): Add e3v5 long JARL support.
20350 (v850_adjust_insn_length): New function. Adjust length of call
20351 insns when using e3v5 instructions.
20352 (v850_gen_movdi): New function: Generate instructions to move a
20354 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
20355 (CPP_SPEC): Define __v850e3v5__ as appropriate.
20356 (TARGET_USE_FPU): Enable for e3v5.
20357 (CONST_OK_FOR_W): New macro.
20358 (ADJUST_INSN_LENGTH): Define.
20359 * config/v850/v850.md (UNSPEC_LOOP): Define.
20360 (attr cpu): Add v850e3v5.
20361 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
20362 (movdi): New pattern.
20363 (movdi_internal): New pattern.
20364 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
20365 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
20366 (cstoresf4): Likewise.
20367 (cstoredf4): Likewise.
20368 (insv): New pattern.
20369 (rotlso3_a): New pattern.
20370 (rotlsi3_b): New pattern
20371 (rotlsi3_v850e3v5): New pattern.
20372 (doloop_begin): New pattern.
20373 (fix_loop_counter): New pattern.
20374 (doloop_end): New pattern.
20375 (branch_normal): Add e3v5 long branch support.
20376 (branch_invert): Likewise.
20377 (branch_z_normal): Likewise.
20378 (branch_z_invert): Likewise.
20379 (branch_nz_normal): Likewise.
20380 (branch_nz_invert): Likewise.
20381 (call_internal_short): Add e3v5 register-indirect JARL support.
20382 (call_internal_long): Likewise.
20383 (call_value_internal_short): Likewise.
20384 (call_value_internal_long): Likewise.
20385 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
20386 (mloop): New option.
20387 * config.gcc: Add support for configuring v840e3v5 target.
20388 * doc/invoke.texi: Document new v850 specific command line options.
20390 2013-01-31 Paul Koning <ni1d@arrl.net>
20394 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
20395 children if parent is a class.
20396 (prune_unused_types_prune): Don't add DW_AT_declaration.
20398 2013-01-31 Richard Biener <rguenther@suse.de>
20400 PR tree-optimization/56157
20401 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
20402 match up operand with SLP child.
20404 2013-01-31 Jason Merrill <jason@redhat.com>
20407 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
20408 parameters the first time.
20409 (gen_scheduled_generic_parms_dies): Check completeness here.
20411 2013-01-31 Richard Biener <rguenther@suse.de>
20413 PR middle-end/53073
20414 * common.opt (faggressive-loop-optimizations): New flag,
20415 enabled by default.
20416 * doc/invoke.texi (faggressive-loop-optimizations): Document.
20417 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
20418 infer_loop_bounds_from_undefined by it.
20420 2013-01-31 Richard Biener <rguenther@suse.de>
20422 PR tree-optimization/56150
20423 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
20424 visit virtual operands.
20425 (find_uses_to_rename_bb): Likewise.
20427 2013-01-31 Richard Biener <rguenther@suse.de>
20429 PR tree-optimization/56150
20430 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
20431 mixed store non-store stmts.
20433 2013-01-30 Jakub Jelinek <jakub@redhat.com>
20436 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
20437 LIBASAN_EARLY_SPEC is defined.
20438 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
20439 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
20441 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
20444 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
20445 invalid args instead of ICEing on it.
20446 (ix86_valid_target_attribute_tree): Return error_mark_node if
20447 ix86_valid_target_attribute_inner_p failed.
20448 (ix86_valid_target_attribute_p): Return false only if
20449 ix86_valid_target_attribute_tree returned error_mark_node. Allow
20450 target("default") attribute.
20451 (sorted_attr_string): Change argument from const char * to tree,
20452 merge in all target attribute arguments rather than just one.
20453 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
20454 instead of free. Avoid using strcat.
20455 (ix86_mangle_function_version_assembler_name): Mangle
20456 target("default") as if no target attribute is present. Adjust
20457 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20458 instead of xmalloc and XDELETEVEC instead of free.
20459 (ix86_function_versions): Don't return true if one of the decls
20460 doesn't have target attribute. If they don't and one of the decls
20461 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20462 sorted_attr_string caller. Use XDELETEVEC instead of free.
20463 (ix86_supports_function_versions): Remove.
20464 (make_name): Fix up formatting.
20465 (make_dispatcher_decl): Remove resolver_name and its initialization.
20466 Avoid leaking memory.
20467 (is_function_default_version): Return true if there is
20468 target("default") attribute rather than no target attribute at all.
20469 (make_resolver_func): Avoid leaking memory.
20470 (ix86_generate_version_dispatcher_body): Likewise.
20471 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20472 * target.def (supports_function_versions): Remove.
20473 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20474 * doc/tm.texi: Regenerated.
20476 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20478 PR rtl-optimization/56144
20479 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20480 for values with side effects.
20482 2013-01-30 Richard Biener <rguenther@suse.de>
20484 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20485 (sparseset_pop): Likewise.
20486 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20487 to be able to use quick_push in the worker loop.
20489 2013-01-30 Marek Polacek <polacek@redhat.com>
20491 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20493 2013-01-30 Richard Biener <rguenther@suse.de>
20496 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
20498 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
20500 PR tree-optimization/56064
20501 * fixed-value.c (fixed_from_double_int): New function.
20502 * fixed-value.h (fixed_from_double_int): New prototype.
20503 (const_fixed_from_double_int): New static inline function.
20504 * fold-const.c (native_interpret_fixed): New static function.
20505 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20506 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20507 (native_encode_fixed): New static function.
20508 (native_encode_expr) <FIXED_CST>: Use it.
20509 (native_interpret_int): Move double_int worker code to...
20510 * double-int.c (double_int::from_buffer): ...this new static method.
20511 * double-int.h (double_int::from_buffer): Prototype it.
20513 2013-01-30 Richard Biener <rguenther@suse.de>
20515 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20516 New pointer-map and obstack.
20517 (init_alias_vars): Allocate pointer-map and obstack.
20518 (delete_points_to_sets): Free them.
20519 (find_what_var_points_to): Cache result.
20520 (find_what_p_points_to): Adjust for changed interface of
20521 find_what_var_points_to.
20522 (compute_points_to_sets): Likewise.
20523 (ipa_pta_execute): Likewise.
20525 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20527 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20528 * configure: Regenerate.
20529 * config.in: Regenerate.
20530 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20531 #nobits/#progbits if supported.
20533 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20536 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20537 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20538 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20540 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20542 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20543 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
20545 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20547 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20548 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
20550 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20552 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20554 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20555 * config/arm/cortex-a7.md: New bypasses using
20556 arm_mac_accumulator_is_result.
20558 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20560 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
20561 (cortex_a7_neon_mla): Likewise.
20562 (cortex_a7_fpfmad): New reservation.
20563 (cortex_a7_fpmacs): Use ffmas and update required units.
20564 (cortex_a7_fpmuld): Update required units and latency.
20565 (cortex_a7_fpmacd): Likewise.
20566 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20567 (cortex_a7_neon). Likewise.
20568 (bypass) Update participating units.
20570 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20572 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20573 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20575 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20576 (vfp_fmul): Use ffmad.
20577 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20578 (cortex_r4_fmacd): Use ffmad.
20579 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20580 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20581 (cortex_a9_fmacd): Use ffmad.
20582 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20583 (cortex_a8_vfp_macd): Use ffmad.
20584 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20585 (cortex_a5_fpmacd): Use ffmad.
20586 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20587 (cortex_a15_vfp_macd): Use ffmad.
20588 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
20590 2013-01-29 Jason Merrill <jason@redhat.com>
20593 * varasm.c (default_assemble_visibility): Don't warn about
20594 visibility on artificial decls.
20596 2013-01-29 Richard Biener <rguenther@suse.de>
20598 PR tree-optimization/56113
20599 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20601 (label_visit): Free duplicate bitmaps and record the leader instead.
20602 (perform_var_substitution): Adjust.
20604 2013-01-29 Richard Biener <rguenther@suse.de>
20606 PR tree-optimization/55270
20607 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20608 the CFG, schedule loops for fixup.
20610 2013-01-29 Nick Clifton <nickc@redhat.com>
20612 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20615 2013-01-28 Leif Ekblad <leif@rdos.net>
20617 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20618 * config/i386/i386.h (TARGET_RDOS): New macro.
20619 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20620 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20621 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20622 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20623 DEFAULT_LARGE_SECTION_THRESHOLD.
20624 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20625 * config/i386/rdos.h: New file.
20626 * config/i386/rdos64.h: New file.
20628 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20631 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20634 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20636 PR rtl-optimization/56117
20637 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20638 call cselib_lookup_from_insn on the MEM before calling
20639 add_insn_mem_dependence.
20641 2013-01-28 Richard Biener <rguenther@suse.de>
20643 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20644 to a stmt that didn't have one.
20645 (copy_phis_for_bb): Likewise for PHI arguments.
20646 (copy_debug_stmt): Likewise for debug stmts.
20648 2013-01-28 Richard Biener <rguenther@suse.de>
20650 PR tree-optimization/56034
20651 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
20652 (partition_builtin_p): Adjust.
20653 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20654 it is the last partition.
20655 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20656 up the vertex for the definition.
20657 (classify_partition): Classify whether a partition is a
20658 PKIND_REDUCTION, thus has uses outside of the loop.
20659 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20660 Merge all PKIND_REDUCTION partitions into the last partition.
20661 (tree_loop_distribution): Seed partitions from reductions as well.
20663 2013-01-28 Jakub Jelinek <jakub@redhat.com>
20665 PR tree-optimization/56125
20666 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20667 pow(x,c) into sqrt(x) * powi(x, n/2) or
20668 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20669 optimizing for size.
20670 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20671 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20674 PR tree-optimization/56094
20675 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20676 to UNKNOWN_LOCATION while gimplifying expr.
20678 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
20681 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20682 operand 0 in movabs insn template for -masm=intel asm alternative.
20683 (*movabs<mode>_2): Ditto for operand 1.
20685 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
20688 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20691 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20693 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20696 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20698 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20699 Undef to avoid warning.
20701 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20703 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20704 * configure: Regenerate.
20706 2013-01-25 Jakub Jelinek <jakub@redhat.com>
20708 PR tree-optimization/56098
20709 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20710 for stmts with volatile ops.
20711 (cond_store_replacement): Don't optimize if assign has volatile ops.
20712 (cond_if_else_store_replacement_1): Don't optimize if either
20713 then_assign or else_assign have volatile ops.
20714 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20717 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20719 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20721 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20723 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20724 missing ':' in asm example.
20726 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20728 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20729 entries into lane and laneq entries.
20730 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20731 Remove AdvSIMD scalar modes.
20732 (aarch64_sq<r>dmulh_laneq<mode>): New.
20733 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20735 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20736 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20737 * config/aarch64/iterators.md (VCOND): New.
20740 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
20743 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20744 Add NULL LIBNAME argument to existing definitions.
20745 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20746 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20747 * config/avr/avr.c (DEF_BUILTIN): Same.
20748 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20749 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20750 implementation is available (DECL_ASSEMBLER_NAME is set).
20751 (avr_fold_absfx): New static function.
20752 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20753 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20754 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20755 AVR_BUILTIN_ABSLLK.
20756 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20757 (abshk, absk, abslk, absllk): Provide as static inline functions.
20759 2013-01-25 Marek Polacek <polacek@redhat.com>
20761 PR tree-optimization/56035
20762 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20764 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
20766 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20767 (*movtf_internal_rex64): Add (!o,C) alternative
20768 (*movxf_internal_rex64): Ditto.
20769 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20771 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
20773 * doc/invoke.texi: fix typo.
20774 * doc/objc.texi: fix typo.
20776 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20778 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20779 for the first two alternatives.
20781 2013-01-24 Diego Novillo <dnovillo@google.com>
20783 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
20784 (ggc-zone.o): Remove.
20785 * configure.ac: Remove option --with-gc.
20786 * configure: Re-generate.
20787 * doc/install.texi: Remove documentation for --with-gc.
20788 * gengtype.c (write_enum_defn): Remove. Update all users.
20789 (write_Types_process_field): Remove generation of gt_e_* argument.
20790 (output_type_enum): Remove. Update all users.
20791 (write_enum_defn): Remove. Update all users.
20792 (enum alloc_zone): Remove. Update all users.
20793 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
20794 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20795 Update all callers.
20796 (struct ptr_data): Remove field TYPE. Update all users.
20797 (gt_pch_note_object): Remove argument TYPE. Update all users.
20798 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20800 * ggc-none.c (ggc_alloc_typed_stat): Remove.
20801 (struct alloc_zone): Remove.
20802 (ggc_internal_alloc_zone_stat): Remove.
20803 (ggc_internal_cleared_alloc_zone_stat): Remove.
20804 * ggc-page.c (ggc_alloc_typed_stat): Remove.
20805 (ggc_pch_count_object): Remove last argument. Update all users.
20806 (ggc_pch_alloc_object): Remove last argument. Update all users.
20807 (struct alloc_zone): Remove.
20808 * ggc-zone.c: Remove.
20809 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
20810 (struct alloc_zone): Remove.
20811 (ggc_alloc_typed_stat): Remove.
20812 (ggc_alloc_typed): Remove.
20813 (ggc_splay_alloc): Remove first argument.
20814 (rtl_zone): Remove. Update all users.
20815 (tree_zone): Remove. Update all users.
20816 (tree_id_zone): Remove. Update all users.
20817 (ggc_internal_zone_alloc_stat): Remove. Update all users.
20818 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
20819 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
20820 * tree-ssanames.c: Remove references to zone allocator in comments.
20822 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
20824 * config/avr/avr.c (avr_out_fract): Make register numbers that
20825 might be outside of source operand signed.
20827 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
20829 * config/i386/constraints.md (Yf): New constraint.
20830 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20831 of f constraint to conditionaly disable x87 register preferences.
20832 (*movdf_internal): Ditto.
20833 (*movsf_internal): Ditto.
20835 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20837 PR inline-asm/55934
20838 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20839 that have operands with impossible constraints.
20840 Add a FIXME for a speed-up opportunity.
20841 * lra-constraints.c (process_alt_operands): Verify that a class
20842 selected from constraints on asms is valid for the operand mode.
20843 (curr_insn_transform): Remove incorrect comment.
20845 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
20847 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20848 TOC operand is a valid symbol ref in the constant pool.
20850 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
20852 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
20854 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
20857 * config/avr/stdfix.h: New file.
20858 * t-avr (stdfix-gcc.h): New rule to build it.
20859 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20861 2013-01-23 Kostya Serebryany <kcc@google.com>
20863 * config/darwin.h: remove dependency on
20864 CoreFoundation (asan on Mac OS).
20866 2013-01-23 Jakub Jelinek <jakub@redhat.com>
20869 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20870 instead of cmpdi_operand for first comparison operand.
20871 Don't assert that comparison operands aren't both constants.
20873 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20875 * doc/install.texi (Downloading the Source): Update references to
20876 downloading separate components.
20878 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20880 * doc/extend.texi (__int128): Improve grammar.
20882 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
20885 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20886 alternative to (o,r).
20887 (*movdi_internal_rex64): Remove (!o,n) alternative.
20888 (DImode immediate->memory splitter): Remove.
20889 (DImode immediate->memory peephole2): Remove.
20890 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20891 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20892 alternative to (!o,*r).
20893 (*movtf_internal_sse): New pattern.
20894 (*movxf_internal_rex64): New pattern.
20895 (*movxf_internal): Disable for TARGET_64BIT.
20896 (*movdf_internal_rex64): Remove (!o,F) alternative.
20898 2013-01-22 Jakub Jelinek <jakub@redhat.com>
20900 PR middle-end/56074
20901 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20902 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20903 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20904 stmt locations where LOCATION_LOCUS of the stmt location is
20905 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20908 * config/i386/i386.md (UNSPEC_STOS): New.
20909 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20910 *strsetqi_1): Add UNSPEC_STOS.
20912 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20915 * doc/invoke.texi: Remove left over -Wsynth example.
20917 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20919 PR tree-optimization/56051
20920 * fold-const.c (fold_binary_loc): Don't fold
20921 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20922 a narrowing conversion, or widening conversion from signed
20925 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20927 PR rtl-optimization/56023
20928 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20929 dependent on debug instruction.
20931 2013-01-21 Martin Jambor <mjambor@suse.cz>
20933 PR middle-end/56022
20934 * function.c (allocate_struct_function): Call
20935 invoke_set_current_function_hook earlier.
20937 2013-01-21 Jakub Jelinek <jakub@redhat.com>
20939 * reload1.c (init_reload): Only initialize reload_obstack
20940 during the first call.
20942 2013-01-21 Marek Polacek <polacek@redhat.com>
20944 * cfgloop.c (verify_loop_structure): Fix up grammar.
20946 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20948 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20949 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20951 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20954 * config/arm/marvell-pj4.md: Update copyright year.
20955 Fix up use of alu to alu_reg and simple_alu_imm.
20957 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
20959 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20961 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20964 * lra-constraints.c (curr_insn_transform): Don't reuse original
20965 insn for secondary memory move when memory mode should be different.
20967 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20969 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20970 atomic_storedi_1): New patterns.
20972 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20974 btver2 pipeline descriptions.
20975 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20977 * config/i386/i386.md (btver2_decode): New type attributes.
20978 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
20980 * config/i386/btver2.md: New file describing btver2 pipelines.
20982 2013-01-19 Andrew Pinski <apinski@cavium.com>
20984 PR tree-optimization/52631
20985 * tree-ssa-sccvn (visit_use): Before looking up the original
20986 statement, try looking up the simplified expression.
20988 2013-01-19 Anthony Green <green@moxielogic.com>
20990 * config/moxie/moxie.c (moxie_expand_prologue): Set
20991 current_function_static_stack_size.
20993 2013-01-18 Jakub Jelinek <jakub@redhat.com>
20995 PR tree-optimization/56029
20996 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20997 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20999 2013-01-18 Sharad Singhai <singhai@google.com>
21001 PR tree-optimization/55995
21002 * dumpfile.c (dump_loc): Print location only if available.
21003 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
21005 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
21008 * lra-constraints.c (curr_insn_transform): Reuse original insn for
21009 secondary memory move.
21010 (inherit_reload_reg): Use rclass instead of cl for
21011 check_secondary_memory_needed_p.
21013 2013-01-18 Jakub Jelinek <jakub@redhat.com>
21015 PR middle-end/56015
21016 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
21017 the case where writing real complex part of target modifies op1.
21019 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
21021 * config/aarch64/aarch64-simd.md
21022 (aarch64_vcond_internal<mode>): Handle unordered cases.
21023 * config/aarch64/iterators.md (v_cmp_result): New.
21025 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
21026 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21028 * config/arm/marvell-pj4.md: New file.
21029 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
21030 * config/arm/arm.md (generic_sched): Add marvell_pj4.
21031 (generic_vfp): Likewise.
21032 * config/arm/arm-cores.def: Add marvell-pj4.
21033 * config/arm/arm-tune.md: Regenerate.
21034 * config/arm/arm-tables.opt: Regenerate.
21035 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
21036 * doc/invoke.texi: Document marvell-pj4.
21038 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
21040 * config/aarch64/arm_neon.h: Map scalar types to standard types.
21042 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21047 * var-tracking.c (negative_power_of_two_p): New.
21048 (global_get_addr_cache, local_get_addr_cache): New.
21049 (get_addr_from_global_cache, get_addr_from_local_cache): New.
21050 (vt_canonicalize_addr): Rewrite using the above. Adjust the
21052 (vt_stack_offset_p): Remove.
21053 (vt_canon_true_dep): Always canonicalize loc's address.
21054 (clobber_overlapping_mems): Make sure we have a MEM.
21055 (local_get_addr_clear_given_value): New.
21056 (val_reset): Clear local cached entries.
21057 (compute_bb_dataflow): Create and release the local cache.
21058 Disable duplicate MEMs clobbering.
21059 (emit_notes_in_bb): Clobber MEMs likewise.
21060 (vt_emit_notes): Create and release the local cache.
21061 (vt_initialize, vt_finalize): Create and release the global
21062 cache, respectively.
21063 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
21065 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21067 PR libmudflap/53359
21068 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
21069 not found in the symtab.
21071 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
21074 PR rtl-optimization/55547
21075 PR rtl-optimization/53827
21078 * alias.c (offset_overlap_p): New, factored out of...
21079 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
21080 the conservative special case for symbolic constants. Don't
21081 adjust zero sizes on alignment.
21083 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
21085 PR rtl-optimization/52573
21086 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
21087 REG_UNUSED for the same register.
21089 2013-01-17 Richard Biener <rguenther@suse.de>
21090 Marek Polacek <polacek@redhat.com>
21092 PR rtl-optimization/55833
21093 * loop-unswitch.c (unswitch_loops): Move loop verification...
21094 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
21095 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
21096 Set it to true when we're removing a loop from hierarchy tree in
21097 an irreducible region.
21098 (fix_bb_placements): Adjust caller.
21099 (fix_loop_placements): Likewise.
21101 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
21103 * config/avr/builtins.def (DEF_BUILTIN): Factor out
21104 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
21105 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
21106 Remove ID. Adjust comments.
21107 * config/avr/avr-c.c (avr_builtin_name): Remove.
21108 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
21109 * config/avr/avr.c (avr_tolower): New static function.
21110 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
21111 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
21112 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
21115 2013-01-17 Jan Hubicka <jh@suse.cz>
21117 PR tree-optimization/55273
21118 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
21120 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
21123 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
21124 store through atomic_store<mode>_1.
21125 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
21127 2013-01-17 Martin Jambor <mjambor@suse.cz>
21129 PR tree-optimizations/55264
21130 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
21131 for virtual methods.
21132 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
21133 virtual methods before inlining is over.
21134 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
21136 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
21139 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
21141 PR rtl-optimization/56005
21142 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
21143 pending reads for prefetch.
21145 2013-01-16 Ian Bolton <ian.bolton@arm.com>
21147 * config/aarch64/aarch64.md
21148 (*cstoresi_neg_uxtw): New pattern.
21149 (*cmovsi_insn_uxtw): New pattern.
21150 (*<optab>si3_uxtw): New pattern.
21151 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
21152 (*<optab>si3_insn_uxtw): New pattern.
21153 (*bswapsi2_uxtw): New pattern.
21155 2013-01-16 Richard Biener <rguenther@suse.de>
21157 * tree-inline.c (tree_function_versioning): Remove set but
21158 never used variable.
21160 2013-01-16 Richard Biener <rguenther@suse.de>
21162 PR tree-optimization/55964
21163 * tree-flow.h (rename_variables_in_loop): Remove.
21164 (rename_variables_in_bb): Likewise.
21165 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
21166 (copy_loop_before): Adjust and delete update-ssa status.
21167 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
21168 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
21169 (rename_variables_in_loop): Remove.
21170 (slpeel_update_phis_for_duplicate_loop): Likewise.
21171 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
21172 use available cfg machinery instead of duplicating it.
21173 Update PHI nodes and perform poor-mans SSA update here.
21174 (slpeel_tree_peel_loop_to_edge): Adjust.
21176 2013-01-16 Richard Biener <rguenther@suse.de>
21178 PR tree-optimization/54767
21179 PR tree-optimization/53465
21180 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
21181 (vrp_visit_phi_node): For PHI arguments coming via backedges
21182 drop all symbolical range information.
21183 (execute_vrp): Compute backedges.
21185 2013-01-16 Richard Biener <rguenther@suse.de>
21187 * doc/install.texi: Update CLooG and ISL requirements to
21190 2013-01-16 Christian Bruel <christian.bruel@st.com>
21193 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
21194 (broken_move): Handle UNSPECV_SP_SWITCH_B.
21195 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
21197 2013-01-16 DJ Delorie <dj@redhat.com>
21199 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
21200 (UNSPECV_SP_SWITCH_E): New.
21201 (sp_switch_1): Change to an unspec.
21202 (sp_switch_2): Change to an unspec. Don't use post-inc when we
21205 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
21207 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
21208 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
21209 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
21210 (expand_mem_thread_fence): Ditto.
21211 (expand_mem_signal_fence): Ditto.
21212 (expand_atomic_load): Ditto.
21213 (expand_atomic_store): Ditto.
21215 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
21217 PR rtl-optimization/55547
21218 PR rtl-optimization/53827
21221 * alias.c (memrefs_conflict_p): Set sizes to negative after
21224 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21227 * function.c (thread_prologue_and_epilogue_insns): Always
21228 add crtl->drap_reg to set_up_by_prologue.set, even if
21229 stack_realign_drap is false.
21231 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
21233 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
21234 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
21235 *call): Fix indention.
21237 2013-01-15 Tom de Vries <tom@codesourcery.com>
21240 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
21243 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
21245 PR rtl-optimization/55153
21246 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
21248 2013-01-15 Martin Jambor <mjambor@suse.cz>
21250 PR tree-optimization/55920
21251 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
21252 accesses as grp_to_be_debug_replaced.
21254 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21256 PR tree-optimization/55920
21257 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
21258 there is non-useless type conversion needed from debug rhs to lhs,
21259 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
21261 2013-01-15 Joseph Myers <joseph@codesourcery.com>
21262 Mikael Pettersson <mikpe@it.uu.se>
21265 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
21267 (ASM_OUTPUT_CASE_LABEL): Remove.
21268 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
21269 * final.c (shorten_branches): Update alignment of labels before
21270 jump tables if CASE_VECTOR_SHORTEN_MODE.
21272 2013-01-15 Richard Biener <rguenther@suse.de>
21275 * system.h: Do not include gmp.h for building host tools.
21277 2013-01-15 Richard Biener <rguenther@suse.de>
21279 PR middle-end/55882
21280 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
21281 account for bitpos when computing alignment.
21283 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
21285 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
21286 (ix86_target_macros_internal): Likewise.
21288 * config/i386/i386.c (m_CORE2I7): Removed.
21289 (m_CORE_HASWELL): New macro.
21290 (m_CORE_ALL): Likewise.
21291 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
21292 (initial_ix86_arch_features): Likewise.
21293 (processor_target_table): Initializations for Core avx2.
21294 (cpu_names): New names "core-avx2".
21295 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
21296 PROCESSOR_CORE_HASWELL.
21297 (ix86_issue_rate): New case.
21298 (ia32_multipass_dfa_lookahead): Likewise.
21299 (ix86_sched_init_global): Likewise.
21301 * config/i386/i386.h (TARGET_HASWELL): New macro.
21302 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
21303 (processor_type): New PROCESSOR_HASWELL.
21305 2013-01-15 Jakub Jelinek <jakub@redhat.com>
21307 PR tree-optimization/55955
21308 * tree-vect-loop.c (vectorizable_reduction): Give up early on
21309 *SHIFT_EXPR and *ROTATE_EXPR codes.
21311 PR tree-optimization/48766
21312 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
21313 -ftrapv disable -fwrapv.
21315 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21318 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
21319 etc. to 1 and not to __flash.
21320 Use LL suffix for __INT24_MAX__ with -mint8.
21321 Use ULL suffix for __UINT24_MAX__ with -mint8.
21323 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21325 * config/avr/avr-arch.h
21326 (struct base_arch_s): Use typedef avr_arch_t instead.
21327 (struct arch_info_s): Use typedef avr_arch_info_t instead.
21328 (struct mcu_type_s): Use typedef avr_mcu_t instead.
21329 * config/avr/avr.c: Same.
21330 * config/avr/avr-devices.c: Same.
21331 * config/avr/driver-avr.c: Same.
21332 * config/avr/gen-avr-mmcu-texi.c: Same.
21333 * config/avr/avr-mcus.def: Adjust comment.
21335 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
21337 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
21338 * config/aarch64/iterators.md (VALLDI): New.
21340 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
21341 Andi Kleen <ak@linux.intel.com>
21344 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
21345 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
21348 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21350 * config/avr/avr-stdint.h: Remove trailing blanks.
21351 * config/avr/avr-log.h: Same.
21352 * config/avr/avr-arch.h: Same.
21353 * config/avr/avr-devices.c: Same.
21354 * config/avr/avr-dimode.md: Same.
21355 * config/avr/predicates.md: Same.
21356 * config/avr/avr-c.c: Same. And fix typo.
21358 * config/avr/avr-protos.h: Same. And:
21359 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
21360 (init_cumulative_args): Rename to avr_init_cumulative_args.
21361 (expand_prologue): Rename to avr_expand_prologue.
21362 (expand_epilogue): Rename to avr_expand_epilogue.
21363 (adjust_insn_length): Rename to avr_adjust_insn_length.
21364 (notice_update_cc): Rename to avr_notice_update_cc.
21365 (final_prescan_insn): Rename to avr_final_prescan_insn.
21366 * config/avr/avr.c: Same.
21367 * config/avr/avr.h: Same.
21368 * config/avr/avr.md: Remove trailing blanks.
21369 (prologue): Use avr_expand_prologue.
21370 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
21372 2013-01-14 Richard Biener <rguenther@suse.de>
21374 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
21375 verify_location, collect_subblocks): New functions.
21376 (verify_gimple_in_cfg): Verify that locations only reference
21377 BLOCKs in the functions BLOCK tree.
21379 2013-01-14 Richard Biener <rguenther@suse.de>
21381 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
21383 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
21385 (insert_out_of_ssa_copy_on_edge): Likewise.
21386 (rewrite_close_phi_out_of_ssa): Likewise.
21387 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
21389 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
21390 propagated constants.
21391 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
21394 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
21396 * config/avr/avr-modes.def: Add GPL copyright notice.
21398 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
21400 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
21401 MEMMODEL_MASK to determine memory model.
21402 (atomic_store<mode>): Ditto from operands[2].
21403 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
21405 2013-01-13 Jakub Jelinek <jakub@redhat.com>
21408 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
21409 (fold_gimple_assign): Don't call unshare_expr here.
21410 (fold_ctor_reference): Call unshare_expr.
21412 2013-01-13 Terry Guo <terry.guo@arm.com>
21414 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
21415 * doc/fragments.texi: Document MULTILIB_REUSE.
21416 * gcc.c (multilib_reuse): New internal spec.
21417 (set_multilib_dir): Also search multilib from multilib_reuse.
21418 * genmultilib (tmpmultilib3): Refactor code.
21419 (tmpmultilib4): Ditto.
21420 (multilib_reuse): New multilib argument.
21422 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
21424 * Makefile.in: Update copyright.
21426 2013-01-12 Tom de Vries <tom@codesourcery.com>
21428 PR middle-end/55890
21429 * calls.c (expand_call): Check if arg_nr is valid.
21431 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
21433 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
21434 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
21435 documentation. Add missing '__' in front of
21436 __builtin_ia32_packssdw256.
21438 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21441 * config/s390/s390.c (s390_preferred_reload_class): Do not return
21442 NO_REGS for larl operands.
21443 (s390_reload_larl_operand): Use s390_load_address instead of
21446 2013-01-11 Richard Biener <rguenther@suse.de>
21448 * tree-cfg.c (verify_node_sharing_1): Split out from ...
21449 (verify_node_sharing): ... here.
21450 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
21452 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
21454 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
21455 Substitute TREECHECKING.
21456 * configure: Regenerate.
21457 * Makefile.in (TREECHECKING): New.
21459 2013-01-11 Richard Guenther <rguenther@suse.de>
21461 PR tree-optimization/44061
21462 * tree-vrp.c (extract_range_basic): Compute zero as
21463 value-range for __builtin_constant_p of function parameters.
21465 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21467 Update copyright years.
21469 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21471 PR rtl-optimization/55672
21472 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21473 const to be eliminable.
21475 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
21477 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21478 * configure: Regenerate.
21480 2013-01-10 Richard Biener <rguenther@suse.de>
21482 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
21484 2013-01-10 Richard Biener <rguenther@suse.de>
21487 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21488 locations for virtual PHI arguments.
21489 (rewrite_update_phi_arguments): Likewise.
21491 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21493 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21496 2013-01-10 Jakub Jelinek <jakub@redhat.com>
21498 PR tree-optimization/55921
21499 * tree-complex.c (expand_complex_asm): New function.
21500 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21502 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21505 * config/s390/s390.c (s390_symref_operand_p)
21506 (s390_loadrelative_operand_p): Merge the two functions.
21507 (s390_check_qrst_address, print_operand_address): Add parameters
21508 to s390_loadrelative_operand_p invokation.
21509 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21510 (s390_reload_larl_operand, s390_secondary_reload): Use
21511 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21512 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21514 2013-01-09 Mike Stump <mikestump@comcast.net>
21516 * dse.c (record_store): Remove unnecessary assert.
21518 2013-01-09 Jan Hubicka <jh@suse.cz>
21520 PR tree-optimization/55569
21521 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21522 * cfgloop.h (scale_loop_profile): Likewise.
21524 2013-01-09 Jan Hubicka <jh@suse.cz>
21527 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21529 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21531 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21533 PR middle-end/55114
21534 * expr.h (maybe_emit_group_store): Declare.
21535 * expr.c (maybe_emit_group_store): New function.
21536 * builtins.c (expand_builtin_int_roundingfn): Call it.
21537 (expand_builtin_int_roundingfn_2): Likewise.
21539 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21541 PR rtl-optimization/55829
21542 * lra-constraints.c (match_reload): Add code for absent output.
21543 (curr_insn_transform): Add code for reloads of matched inputs
21546 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
21548 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21549 attribute of movddup insn to DF.
21550 (*vec_interleave_lowv2df): Ditto.
21551 (vec_dupv2df): Ditto.
21553 2013-01-09 Jan Hubicka <jh@suse.cz>
21555 PR tree-optimiation/55875
21556 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21557 EVERY_ITERATION parameter.
21558 (number_of_iterations_exit): Check if exit is executed every iteration.
21559 (idx_infer_loop_bounds): Similarly here.
21560 (n_of_executions_at_most): Simplify
21561 to only test for cases where statement is dominated by the
21562 particular bound; handle correctly the "postdominance" test.
21563 (scev_probably_wraps_p): Use max loop iterations info
21564 as a global bound first.
21566 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
21567 Nick Clifton <nickc@redhat.com>
21569 * config/v850/v850.md (cbranchsf4): New pattern.
21570 (cstoresf4): New pattern.
21571 (cbranchdf4): New pattern.
21572 (cstoredf4): New pattern.
21573 (movsicc): Disallow floating point comparisons.
21574 (cmpsf_le_insn): Fix order of operators.
21575 (cmpsf_lt_insn): Likewise.
21576 (cmpsf_eq_insn): Likewise.
21577 (cmpdf_le_insn): Likewise.
21578 (cmpdf_lt_insn): Likewise.
21579 (cmpdf_eq_insn): Likewise.
21580 (cmpsf_ge_insn): Use LE comparison.
21581 (cmpdf_ge_insn): Likewise.
21582 (cmpsf_gt_insn): Use LT comparison.
21583 (cmpdf_gt_insn): Likewise.
21584 (cmpsf_ne_insn): Delete pattern.
21585 (cmpdf_ne_insn): Delete pattern.
21586 * config/v850/v850.c (v850_gen_float_compare): Use
21587 gen_cmpdf_eq_insn for NE comparison.
21588 (v850_float_z_comparison_operator)
21589 (v850_float_nz_comparison_operator): Move from here ...
21590 * config/v850/predicates.md: ... to here. Move GT and GE
21591 comparisons into v850_float_z_comparison_operator.
21592 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21594 (v850_float_nz_comparison_operator): Likewise.
21596 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21598 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21599 with calls to gen_insvsi/gen_insvdi.
21601 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21603 * config/i386/i386.c (initial_ix86_tune_features): Set up
21604 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21606 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21607 Jakub Jelinek <jakub@redhat.com>
21609 PR tree-optimization/48189
21610 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21611 If nitercst is 0, don't predict the exit edge.
21613 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21615 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21616 in asm_fprintf with reg_names.
21617 (aarch64_print_operand_address): Likewise.
21618 (aarch64_return_addr): Likewise.
21619 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21621 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21623 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21624 (INT_U6_BITS): Likewise.
21625 * config/pa/predicates.md (uint6_operand): New predicate.
21626 (shift5_operand, shift6_operand): Likewise.
21627 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21629 (lshrdi3): Use shift6_operand.
21630 (shrpsi4, shrpdi4): New insn patterns.
21631 (extzv): Delete expander.
21632 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21633 predicates in unamed zero extract patterns. Tighten common constraint.
21634 (extv): Delete expander.
21635 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21636 predicates in unamed sign extract patterns. Tighten common constraint.
21637 (insv): Delete expander.
21638 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21639 predicates in unamed insert patterns. Tighten common constraint.
21640 Change uint32_operand predicate to uint6_operand predicate in unamed
21641 DImode pattern to insert constant values of type 1...1xxxx.
21643 2013-01-04 Jan Hubicka <jh@suse.cz>
21645 PR tree-optimization/55823
21646 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21649 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21650 Uros Bizjak <ubizjak@gmail.com>
21652 PR rtl-optimization/55845
21653 * df-problems.c (can_move_insns_across): Stop scanning at
21654 volatile_insn_p source instruction or give up if
21655 across_from .. across_to range contains any volatile_insn_p
21658 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21660 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21661 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21663 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
21664 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
21666 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21669 * asan.c (asan_clear_shadow): New function.
21670 (asan_emit_stack_protection): Use it.
21672 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21674 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21675 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21676 with tab instead of space.
21678 2013-01-08 Nick Clifton <nickc@redhat.com>
21680 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21681 register bank 0 at the start of an interrupt handler.
21682 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21685 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21687 * config/aarch64/aarch64-simd.md
21688 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21689 (aarch64_simd_bsl): Likewise.
21690 (aarch64_vcond_internal<mode>): Likewise.
21691 (vcond<mode><mode>): Likewise.
21692 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21693 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21695 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21697 * config/aarch64/aarch64-builtins.c
21698 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21700 2013-01-08 Martin Jambor <mjambor@suse.cz>
21703 * tree-sra.c (analyze_access_subtree): Return true also after
21704 potentially creating a debug-only replacement.
21706 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21708 PR middle-end/55890
21709 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21711 PR tree-optimization/54120
21712 * tree-vrp.c (range_fits_type_p): Don't allow
21713 src_precision < precision from signed vr to unsigned_p
21714 if vr->min or vr->max is negative.
21715 (simplify_float_conversion_using_ranges): Test can_float_p
21716 against CODE_FOR_nothing.
21718 2013-01-08 Jakub Jelinek <jakub@redhat.com>
21719 Richard Biener <rguenther@suse.de>
21721 PR middle-end/55851
21722 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21723 types instead of just INTEGER_TYPE types.
21725 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
21727 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21730 2013-01-07 Steve Ellcey <sellcey@mips.com>
21733 * config/mips/mips.opt: Change mad to mmad to match documentation.
21735 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21738 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21739 .progmemx.data now.
21741 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21744 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21745 (avr_addrspace_t): Add .section_name field.
21746 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21748 (avr_addrspace): Same. Initialize .section_name. Remove last
21749 NULL entry. Put __memx into .progmemx.data.
21750 (progmem_section_prefix): Remove.
21751 (avr_asm_init_sections): No need to initialize progmem_section.
21752 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21753 section name prefix.
21754 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21755 retrieve the progmem section.
21756 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21757 boundary to run over avr_addrspace[].
21758 (avr_register_target_pragmas): Ditto.
21760 2013-01-06 Jakub Jelinek <jakub@redhat.com>
21762 * varasm.c (output_constant_def_contents): For asan_protect_global
21763 protected strings, adjust DECL_ALIGN if needed, before testing for
21765 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21766 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21768 (output_object_block): For asan protected decls, emit asan padding
21769 after their contents.
21770 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21771 (asan_finish_file): Test it here instead.
21773 2013-01-07 Nick Clifton <nickc@redhat.com>
21774 Matthias Klose <doko@debian.org>
21775 Doug Kwan <dougkwan@google.com>
21776 H.J. Lu <hongjiu.lu@intel.com>
21779 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21781 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21783 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21785 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
21787 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21789 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21792 * doc/install.texi (Cross-Compiler-Specific Options): Document
21795 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21797 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21798 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21799 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21800 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21801 vqmovun_high_s64): Fix source operand number and update copyright.
21803 2013-01-07 Richard Biener <rguenther@suse.de>
21805 PR middle-end/55890
21806 * gimple.h (gimple_call_builtin_p): New overload.
21807 * gimple.c (validate_call): New function.
21808 (gimple_call_builtin_p): Likewise.
21809 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21810 Use gimple_call_builtin_p.
21811 (find_func_clobbers): Likewise.
21812 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21813 (strlen_optimize_stmt): Likewise.
21815 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21817 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21818 (vld1q_dup_*): Likewise.
21819 (vld1_*): Likewise.
21820 (vld1q_*): Likewise.
21821 (vld1_lane_*): Likewise.
21822 (vld1q_lane_*): Likewise.
21824 2013-01-07 Richard Biener <rguenther@suse.de>
21826 * lto-streamer.h (LTO_minor_version): Bump to 2.
21828 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21830 * config/aarch64/aarch64-protos.h
21831 (aarch64_const_double_zero_rtx_p): Rename to...
21832 (aarch64_float_const_zero_rtx_p): ...this.
21833 (aarch64_float_const_representable_p): New.
21834 (aarch64_output_simd_mov_immediate): Likewise.
21835 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21836 move immediate case.
21837 * config/aarch64/aarch64.c
21838 (aarch64_const_double_zero_rtx_p): Rename to...
21839 (aarch64_float_const_zero_rtx_p): ...this.
21840 (aarch64_print_operand): Allow printing of new constants.
21841 (aarch64_valid_floating_const): New.
21842 (aarch64_legitimate_constant_p): Check for valid floating-point
21844 (aarch64_simd_valid_immediate): Likewise.
21845 (aarch64_vect_float_const_representable_p): New.
21846 (aarch64_float_const_representable_p): Likewise.
21847 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21848 (aarch64_output_simd_mov_immediate): New.
21849 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21850 (*movdf_aarch64): Likewise.
21851 * config/aarch64/constraints.md (Ufc): New.
21852 (Y): call aarch64_float_const_zero_rtx.
21853 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21855 2013-01-07 Richard Biener <rguenther@suse.de>
21857 PR tree-optimization/55888
21858 PR tree-optimization/55862
21859 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21860 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21861 not if it is contained therein.
21863 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21865 * config/avr/t-avr: Typo.
21867 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
21870 * config/avr/t-avr: Don't automatically rebuild
21871 $(srcdir)/config/avr/t-multilib
21872 $(srcdir)/config/avr/avr-tables.opt
21873 $(srcdir)/doc/avr-mmcu.texi
21874 (avr-mcus): New phony target to build them on request.
21875 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21876 * avr/avr-mcus.def: Adjust comments.
21878 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
21880 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21882 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21884 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21886 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21888 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21890 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
21892 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21893 to generate profiling.
21894 * config/rs6000/aix64.h (LIB_SPEC): Same.
21896 2013-01-04 Andrew Pinski <apinski@cavium.com>
21898 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21900 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21902 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
21904 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21906 (ix86_expand_move): Ditto.
21907 (ix86_zero_extend_to_Pmode): Ditto.
21908 (ix86_expand_call): Ditto.
21909 (ix86_expand_special_args_builtin): Ditto.
21910 (ix86_expand_builtin): Ditto.
21912 2013-01-04 Richard Biener <rguenther@suse.de>
21914 PR tree-optimization/55862
21915 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21916 translating them through PHI nodes.
21918 2013-01-04 Martin Jambor <mjambor@suse.cz>
21920 PR tree-optimization/55755
21921 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21922 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21924 2013-01-04 Marek Polacek <polacek@redhat.com>
21926 PR middle-end/55859
21927 * opts.c (default_options_optimization): Clarify error message.
21929 2013-01-04 Richard Biener <rguenther@suse.de>
21931 PR middle-end/55863
21932 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21935 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21938 * config/pa/pa.md (movsi): Revert previous change.
21939 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21942 2013-01-03 Richard Henderson <rth@redhat.com>
21944 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21945 after eliminating TLS symbols.
21947 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
21950 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21951 * graphite-poly.c (debug_gmp_value): Likewise.
21953 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
21956 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21957 selected code model, define __code_mode_small__, __code_model_medium__,
21958 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21959 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21960 xchg temporary register with %k. Declare temporary register as
21962 [__x86_64__]: For medium and large code models, preserve %rbx register.
21964 2013-01-03 Richard Biener <rguenther@suse.de>
21966 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
21967 (dump_subscript): Adjust.
21968 (finalize_ddr_dependent): Do not dump redundant info.
21969 (analyze_siv_subscript): Adjust.
21970 (subscript_dependence_tester): Likewise.
21971 (compute_affine_dependence): Likewise.
21973 2013-01-03 Richard Biener <rguenther@suse.de>
21976 2013-01-03 Richard Biener <rguenther@suse.de>
21978 PR tree-optimization/55857
21979 * tree-vect-stmts.c (vectorizable_load): Do not setup
21980 re-alignment for invariant loads.
21982 2013-01-02 Richard Biener <rguenther@suse.de>
21984 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
21985 invariant load do not generate a vector load from the scalar location.
21987 2013-01-03 Richard Biener <rguenther@suse.de>
21989 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21990 for not vectorizing.
21991 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21992 not build INDIRECT_REFs, call get_name once only.
21993 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21994 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21996 2013-01-03 Richard Biener <rguenther@suse.de>
21998 PR tree-optimization/55857
21999 * tree-vect-stmts.c (vectorizable_load): Do not setup
22000 re-alignment for invariant loads.
22002 2013-01-03 Richard Biener <rguenther@suse.de>
22005 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
22006 prefer a built-in decl.
22008 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22010 * gcc.c (process_command): Update copyright notice dates.
22011 * gcov.c (print_version): Likewise.
22012 * gcov-dump.c (print_version): Likewise.
22014 PR rtl-optimization/55838
22015 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
22016 iv0.step, iv1.step and step.
22018 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22019 Marc Glisse <marc.glisse@inria.fr>
22021 PR tree-optimization/55832
22022 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
22023 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
22024 integer_{one,zero}_node.
22026 2013-01-03 Jakub Jelinek <jakub@redhat.com>
22029 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
22030 * var-tracking.c (reverse_op): Don't add reverse ops to
22031 VALUEs that have already
22032 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
22034 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
22036 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
22038 2013-01-02 Teresa Johnson <tejohnson@google.com>
22040 * dumpfile.c (dump_loc): Print filename with location.
22041 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
22042 new location_t parameter to emit complete unroll message with
22043 new dump framework.
22044 (canonicalize_loop_induction_variables): Compute loops location
22045 and pass to try_unroll_loop_completely.
22046 * loop-unroll.c (report_unroll_peel): New function.
22047 (peel_loops_completely): Use new dump format with location
22048 for main dumpfile message, and invoke report_unroll_peel on success.
22049 (decide_unrolling_and_peeling): Ditto.
22050 (decide_peel_once_rolling): Remove old dumpfile message subsumed
22051 by report_unroll_peel.
22052 (decide_peel_completely): Ditto.
22053 (decide_unroll_constant_iterations): Ditto.
22054 (decide_unroll_runtime_iterations): Ditto.
22055 (decide_peel_simple): Ditto.
22056 (decide_unroll_stupid): Ditto.
22057 * cfgloop.c (get_loop_location): New function.
22058 * cfgloop.h (get_loop_location): Declare.
22060 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22062 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
22065 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22067 PR middle-end/55198
22068 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
22069 BLKmode objects when EXPAND_MEMORY is specified.
22071 2013-01-02 Sriraman Tallam <tmsriram@google.com>
22073 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
22075 (fold_builtin_cpu): Do not share cpu model decls across statements.
22077 2013-01-02 Jason Merrill <jason@redhat.com>
22080 * tree.c (build_array_type_1): Revert earlier change.
22082 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
22084 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
22086 * config/aarch64/aarch64-tune.md: Re-generate.
22088 2013-01-02 Richard Biener <rguenther@suse.de>
22090 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
22091 invariant load do not generate a vector load from the scalar location.
22093 2013-01-02 Richard Biener <rguenther@suse.de>
22096 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
22097 * configure: Regenerate.
22099 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22101 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
22102 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
22103 (expand_builtin_int_roundingfn_2): Keep the original target around
22104 for the fallback case.
22106 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
22108 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
22109 to be clear for sign changes.
22111 2013-01-01 Jan Hubicka <jh@suse.cz>
22113 * ipa-inline-analysis.c: Fix formatting.
22115 2013-01-01 Jakub Jelinek <jakub@redhat.com>
22117 PR tree-optimization/55831
22118 * tree-vect-loop.c (get_initial_def_for_induction): Use
22119 gsi_after_labels instead of gsi_start_bb.
22121 Copyright (C) 2013 Free Software Foundation, Inc.
22123 Copying and distribution of this file, with or without modification,
22124 are permitted in any medium without royalty provided the copyright
22125 notice and this notice are preserved.