1 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
4 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
5 * config/arm/arm-cores.def (License): Same.
7 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
9 * tree.h (decl_comdat_group): Declare.
10 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
11 * tree.c (decl_comdat_group): Here.
13 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
15 PR rtl-optimization/61222
16 * combine.c (simplify_shift_const_1): When moving a PLUS outside
17 the shift, truncate the PLUS operand to the result mode.
19 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
22 * config/i386/i386.c (ix86_rtx_costs)
23 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
26 2014-05-26 Martin Jambor <mjambor@suse.cz>
28 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
31 2014-05-26 Richard Biener <rguenther@suse.de>
33 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
34 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
35 Provide specializations.
36 (wi::int_traits <HOST_WIDE_INT>,
37 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
39 2014-05-26 Alan Modra <amodra@gmail.com>
42 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
43 params and return a bool. Remove dead code. Update comment.
44 Assert we have a const_int source. Remove bogus code from
45 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
46 handling of constants > 2G and reg_equal note, from..
47 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
48 return value. Update comment. If we can, use a new pseudo
49 for intermediate calculations.
50 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
52 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
53 call to rs6000_emit_set_const in splitter.
54 (movdi_internal64+2, +3): Likewise.
56 2014-05-26 Richard Biener <rguenther@suse.de>
58 * system.h: Define __STDC_FORMAT_MACROS before
60 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
61 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
62 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
63 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
64 HOST_WIDEST_INT_C): Remove.
65 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
66 if C99 inttypes.h is not available.
67 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
68 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
69 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
70 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
71 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
72 (struct output_info): Likewise.
73 (print_statistics): Adjust.
74 (dump_bitmap_statistics): Likewise.
75 * bt-load.c (migrate_btr_defs): Print with PRId64.
76 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
77 (MAX_SAFE_MULTIPLIER): Adjust.
78 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
79 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
80 dump_cgraph_node): Likewise.
81 * final.c (dump_basic_block_info): Likewise.
82 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
83 * gcov.c (format_gcov): Likewise.
84 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
86 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
87 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
88 (inline_small_functions, dump_overall_stats, dump_inline_stats):
89 Use PRId64 for dumping.
90 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
91 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
92 (add_allocno_hard_regs): Adjust.
93 * loop-doloop.c (doloop_modify): Print using PRId64.
94 * loop-iv.c (inverse): Compute in uint64_t.
95 (determine_max_iter, iv_number_of_iterations): Likewise.
96 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
98 * lto-streamer-out.c (write_symbol): Use uint64_t.
99 * mcf.c (CAP_INFINITY): Use int64_t maximum.
100 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
101 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
102 * modulo-sched.c (const_iteration_count): Use int64_t.
103 (sms_schedule): Dump using PRId64.
104 * predict.c (dump_prediction): Likewise.
105 * pretty-print.h (pp_widest_integer): Remove.
106 * profile.c (get_working_sets, is_edge_inconsistent,
107 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
108 * tree-pretty-print.c (pp_double_int): Remove case handling
109 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
110 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
112 (pass_optimize_bswap::execute): Remove restriction on hosts.
113 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
114 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
115 * tree.c (widest_int_cst_value): Remove.
116 * tree.h (widest_int_cst_value): Likewise.
117 * value-prof.c (dump_histogram_value): Print using PRId64.
118 * gengtype.c (main): Also inject int64_t.
119 * ggc-page.c (struct max_alignment): Use int64_t.
120 * alloc-pool.c (struct allocation_object_def): Likewise.
121 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
123 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
124 * doc/tm.texi: Regenerated.
125 * gengtype-lex.l (IWORD): Handle [u]int64_t.
126 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
127 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
128 mmix_output_register_setting): Use [u]int64_t in prototypes.
129 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
130 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
131 mmix_output_octa, mmix_output_shifted_value): Adjust.
132 (mmix_intval): Adjust. Remove unreachable case.
133 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
135 2014-05-26 Richard Biener <rguenther@suse.de>
137 * configure.ac: Drop __int64 type check. Insist that we
138 found uint64_t and int64_t.
139 * hwint.h (HOST_BITS_PER___INT64): Remove.
140 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
141 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
142 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
143 (HOST_WIDEST_FAST_INT): Remove __int64 case.
144 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
145 for dst_q_src_df_rms_cdt.
146 * configure: Regenerate.
147 * config.in: Likewise.
149 2014-05-26 Michael Tautschnig <mt@debian.org>
152 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
153 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
155 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
157 PR rtl-optimization/61278
158 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
160 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
162 PR rtl-optimization/61220
163 Part of PR rtl-optimization/61225
164 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
165 insn; skip split_edge for a block with only one successor.
167 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
169 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
172 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
174 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
175 (update_vtable_references): New function.
176 (function_and_variable_visibility): Rewrite also vtable initializers.
177 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
179 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
181 * ggc.h (ggc_grow): New function.
182 * ggc-none.c (ggc_grow): New function.
183 * ggc-page.c (ggc_grow): Likewise.
185 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
187 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
188 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
189 comdat_can_be_unshared_p, cgraph_externally_visible_p,
190 varpool_externally_visible_p, can_replace_by_local_alias,
191 update_visibility_by_resolution_info, function_and_variable_visibility,
192 pass_data_ipa_function_and_variable_visibility,
193 make_pass_ipa_function_and_variable_visibility,
194 whole_program_function_and_variable_visibility,
195 pass_data_ipa_whole_program_visibility,
196 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
197 * cgraph.h (cgraph_local_node_p): Declare.
198 * ipa-visibility.c: New file.
199 * Makefile.in (OBJS): Add ipa-visiblity.o
201 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
203 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
204 that var decl is available.
206 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
208 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
210 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
211 (find_decls_types_r): Do not walk COMDAT_GROUP.
212 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
213 * varasm.c (make_decl_one_only): Use set_comdat_group;
214 create node if needed.
215 * ipa-inline-transform.c (save_inline_function_body): Update
216 way we decl->symtab mapping.
217 * symtab.c (symtab_hash, hash_node, eq_node
218 symtab_insert_node_to_hashtable): Remove.
219 (symtab_register_node): Update.
220 (symtab_unregister_node): Update.
221 (symtab_get_node): Reimplement as inline function.
222 (symtab_add_to_same_comdat_group): Update.
223 (symtab_dissolve_same_comdat_group_list): Update.
224 (dump_symtab_base): Update.
225 (verify_symtab_base): Update.
226 (symtab_make_decl_local): Update.
227 (fixup_same_cpp_alias_visibility): Update.
228 (symtab_nonoverwritable_alias): Update.
229 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
230 * ipa.c (update_visibility_by_resolution_info): UPdate.
231 * bb-reorder.c: Include cgraph.h
232 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
234 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
235 * cgraph.c (cgraph_get_create_node): Update.
236 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
238 (symtab_get_node): Make inline.
239 (symtab_insert_node_to_hashtable): Remove.
240 (symtab_can_be_discarded): Update.
241 (decl_comdat_group): New function.
242 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
244 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
246 (read_comdat_group): New function.
247 (input_node, input_varpool_node): Use it.
248 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
250 * mips.c (mips_start_unique_function): Likewise.
251 (ix86_code_end): Likewise.
252 (rs6000_code_end): Likweise.
253 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
255 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
257 * gengtype-state.c (fatal_reading_state): Bring offline.
258 * optabs.c (widening_optab_handler): Bring offline.
259 * optabs.h (widening_optab_handler): Likewise.
260 * final.c (get_attr_length_1): Likewise.
262 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
264 * sched-int.h (sd_iterator_cond): Manually tail recurse.
266 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
268 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
269 (ppc440-compare): Include shift with dot.
270 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
271 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
273 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
275 (e6500_sfx2): Include it.
276 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
277 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
278 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
279 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
280 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
281 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
282 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
283 *lshiftrt_internal1le, *lshiftrt_internal1be,
284 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
285 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
286 *rotldi3_internal10le, *rotldi3_internal10be,
287 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
288 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
289 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
290 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
291 define_insns): Use type "shift" in the appropriate alternatives.
293 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
295 * config/rs6000/rs6000.md (type): Add "logical". Delete
297 (dot): Adjust comment.
298 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
299 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
300 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
301 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
302 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
303 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
304 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
305 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
307 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
308 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
309 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
310 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
311 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
312 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
313 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
314 * config/rs6000/8540.md (ppc8540_su): Adjust.
315 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
316 cell-cmp-microcoded): Adjust.
317 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
318 * config/rs6000/e500mc.md (e500mc_su): Adjust.
319 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
320 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
321 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
322 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
323 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
324 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
325 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
327 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
328 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
329 Adjust. Adjust comment.
330 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
331 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
333 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
335 * config/rs6000/rs6000.md (type): Add "add".
336 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
337 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
338 define_insns): Use it.
339 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
341 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
342 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
343 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
344 * config/rs6000/601.md (ppc601-integer): Adjust.
345 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
346 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
347 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
348 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
349 * config/rs6000/8540.md (ppc8540_su): Adjust.
350 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
351 cell-cmp-microcoded): Adjust.
352 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
353 * config/rs6000/e500mc.md (e500mc_su): Adjust.
354 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
355 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
356 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
357 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
358 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
359 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
360 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
362 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
363 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
364 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
365 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
367 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
369 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
370 "delayed_compare", "var_delayed_compare".
371 (var_shift): New attribute.
372 (cell_micro): Adjust.
373 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
374 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
375 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
376 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
377 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
378 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
379 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
380 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
381 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
382 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
383 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
384 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
385 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
386 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
387 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
388 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
389 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
390 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
391 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
392 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
393 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
394 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
395 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
396 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
397 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
399 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
400 * config/rs6000/440.md (ppc440-integer): Adjust.
401 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
402 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
403 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
404 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
405 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
406 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
407 * config/rs6000/8540.md (ppc8540_su): Adjust.
408 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
409 cell-cmp-microcoded): Adjust.
410 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
411 * config/rs6000/e500mc.md (e500mc_su): Adjust.
412 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
413 e500mc64_delayed): Adjust.
414 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
415 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
416 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
417 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
418 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
419 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
420 power6-delayed-compare, power6-var-delayed-compare): Adjust.
421 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
422 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
424 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
425 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
427 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
429 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
430 (bits): New mode_attr.
431 (idiv_ldiv): Delete mode_attr.
432 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
433 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
434 rs6000_adjust_priority, is_nonpipeline_insn,
435 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
437 * config/rs6000/40x.md (ppc403-idiv): Adjust.
438 * config/rs6000/440.md (ppc440-idiv): Adjust.
439 * config/rs6000/476.md (ppc476-idiv): Adjust.
440 * config/rs6000/601.md (ppc601-idiv): Adjust.
441 * config/rs6000/603.md (ppc603-idiv): Adjust.
442 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
443 ppc620-ldiv): Adjust.
444 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
445 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
446 * config/rs6000/8540.md (ppc8540_divide): Adjust.
447 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
448 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
449 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
450 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
451 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
452 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
453 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
454 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
455 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
456 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
457 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
458 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
459 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
460 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
461 * config/rs6000/titan.md (titan_fxu_div): Adjust.
463 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
465 * config/rs6000/rs6000.md (type): Delete "insert_word",
466 "insert_dword". Add "insert".
467 (size): Update comment.
468 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
469 insn_must_be_first_in_group): Adjust.
470 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
471 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
472 *insvsi_internal6, insvdi_internal): Adjust.
474 * config/rs6000/40x.md (ppc403-integer): Adjust.
475 * config/rs6000/440.md (ppc440-integer): Adjust.
476 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
477 * config/rs6000/601.md (ppc601-integer): Adjust.
478 * config/rs6000/603.md (ppc603-integer): Adjust.
479 * config/rs6000/6xx.md (ppc604-integer): Adjust.
480 * config/rs6000/7450.md (ppc7450-integer): Adjust.
481 * config/rs6000/7xx.md (ppc750-integer): Adjust.
482 * config/rs6000/8540.md (ppc8540_su): Adjust.
483 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
484 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
485 * config/rs6000/e500mc.md (e500mc_su): Adjust.
486 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
487 * config/rs6000/e5500.md (e5500_sfx): Adjust.
488 * config/rs6000/e6500.md (e6500_sfx): Adjust.
489 * config/rs6000/mpc.md (mpccore-integer): Adjust.
490 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
491 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
492 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
493 * config/rs6000/power7.md (power7-integer): Adjust.
494 * config/rs6000/power8.md (power8-1cyc): Adjust.
495 * config/rs6000/rs64.md (rs64a-integer): Adjust.
496 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
498 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
500 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
501 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
502 (size): New attribute.
503 (dot): New attribute.
504 (cell_micro): Adjust.
505 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
506 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
507 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
508 umuldi3_highpart): Adjust.
509 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
510 rs6000_adjust_priority, is_nonpipeline_insn,
511 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
513 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
514 ppc405-imul3): Adjust.
515 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
516 * config/rs6000/476.md (ppc476-imul): Adjust.
517 * config/rs6000/601.md (ppc601-imul): Adjust.
518 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
519 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
520 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
521 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
522 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
524 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
525 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
526 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
528 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
529 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
530 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
531 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
532 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
533 * config/rs6000/mpc.md (mpccore-imul): Adjust.
534 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
535 power4-lmul, power4-imul, power4-imul3): Adjust.
536 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
537 power5-lmul, power5-imul, power5-imul3): Adjust.
538 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
539 power6-lmul, power6-imul, power6-imul3): Adjust.
540 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
541 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
543 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
545 * config/rs6000/titan.md (titan_imul): Adjust.
547 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
549 * config/rs6000/rs6000.md (type): Add new value "halfmul".
550 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
551 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
552 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
553 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
554 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
555 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
556 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
557 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
558 * config/rs6000/titan.md: Delete nonsensical comment.
559 (titan_imul): Add type imul3.
560 (titan_mulhw): Remove type imul3; add type halfmul.
562 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
564 * config/rs6000/rs6000.md (type): Reorder, reformat.
566 2014-05-23 Martin Jambor <mjambor@suse.cz>
568 PR tree-optimization/53787
569 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
570 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
571 analysis_done, update all uses.
572 * ipa-prop.c: Include domwalk.h
573 (param_analysis_info): Removed.
574 (param_aa_status): New type.
575 (ipa_bb_info): Likewise.
576 (func_body_info): Likewise.
577 (ipa_get_bb_info): New function.
578 (aa_overwalked): Likewise.
579 (find_dominating_aa_status): Likewise.
580 (parm_bb_aa_status_for_bb): Likewise.
581 (parm_preserved_before_stmt_p): Changed to use new param AA info.
582 (load_from_unmodified_param): Accept func_body_info as a parameter
583 instead of parms_ainfo.
584 (parm_ref_data_preserved_p): Changed to use new param AA info.
585 (parm_ref_data_pass_through_p): Likewise.
586 (ipa_load_from_parm_agg_1): Likewise. Update callers.
587 (compute_complex_assign_jump_func): Changed to use new param AA info.
588 (compute_complex_ancestor_jump_func): Likewise.
589 (ipa_compute_jump_functions_for_edge): Likewise.
590 (ipa_compute_jump_functions): Removed.
591 (ipa_compute_jump_functions_for_bb): New function.
592 (ipa_analyze_indirect_call_uses): Likewise, moved variable
594 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
595 and info, moved variable declarations down.
596 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
598 (ipa_analyze_stmt_uses): Likewise.
599 (ipa_analyze_params_uses): Removed.
600 (ipa_analyze_params_uses_in_bb): New function.
601 (ipa_analyze_controlled_uses): Likewise.
602 (free_ipa_bb_info): Likewise.
603 (analysis_dom_walker): New class.
604 (ipa_analyze_node): Handle node-specific forbidden analysis,
605 initialize and free func_body_info, use dominator walker.
606 (ipcp_modif_dom_walker): New class.
607 (ipcp_transform_function): Create and free func_body_info, use
608 ipcp_modif_dom_walker, moved a lot of functionality there.
610 2014-05-23 Marek Polacek <polacek@redhat.com>
611 Jakub Jelinek <jakub@redhat.com>
613 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
614 * gcc.c (sanitize_spec_function): Likewise.
615 * convert.c (convert_to_integer): Include "ubsan.h". Add
616 floating-point to integer instrumentation.
617 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
618 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
620 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
621 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
622 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
623 * ubsan.c: Include "realmpfr.h" and "dfp.h".
624 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
625 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
626 float/double/long double.
627 (ubsan_instrument_float_cast): New function.
628 * ubsan.h (ubsan_instrument_float_cast): Declare.
630 2014-05-23 Jiong Wang <jiong.wang@arm.com>
632 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
634 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
635 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
636 Adjust for tailcalling through registers.
637 * config/aarch64/aarch64.h (enum reg_class): New caller save
639 (REG_CLASS_NAMES): Likewise.
640 (REG_CLASS_CONTENTS): Likewise.
641 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
642 Allow tailcalling without decls.
644 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
646 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
647 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
649 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
650 gsi, and variables v_* to v*.
652 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
654 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
656 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
658 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
659 * omp-low.c: Update accordingly.
661 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
662 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
663 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
664 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
665 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
666 GF_OMP_TARGET_KIND_UPDATE.
668 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
669 Explicitly enumerate the expected region types.
671 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
674 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
675 documentation; the old documentation didn't clearly state the
676 constraints on the contents of the pointed-to storage.
678 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
680 Fix bootstrap error on ia64
681 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
682 Return default value.
684 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
686 PR tree-optimization/54733
687 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
689 (find_bswap_or_nop_load): New.
690 (find_bswap_1): Renamed to ...
691 (find_bswap_or_nop_1): This. Also add support for memory source.
692 (find_bswap): Renamed to ...
693 (find_bswap_or_nop): This. Also add support for memory source and
694 detection of bitwise operations equivalent to load in host endianness.
695 (execute_optimize_bswap): Likewise. Also move its leading comment back
696 in place and split statement transformation into ...
697 (bswap_replace): This.
699 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
701 PR rtl-optimization/61215
702 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
703 simplify_gen_subreg until final substitution.
705 2014-05-23 Alan Modra <amodra@gmail.com>
708 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
709 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
710 Use "Y" constraint rather than "m".
712 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
714 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
716 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
717 New function declaration.
718 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
719 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
720 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
721 (aarch64_init_builtins) : Initialize builtins
722 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
723 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
724 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
725 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
726 and __builtins_aarch64_set_fpsr.
727 (aarch64_atomic_assign_expand_fenv): New function.
728 * config/aarch64/aarch64.md (set_fpcr): New pattern.
729 (get_fpcr) : Likewise.
730 (set_fpsr) : Likewise.
731 (get_fpsr) : Likewise.
732 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
733 and UNSPECV_SET_FPSR.
734 * doc/extend.texi (AARCH64 Built-in Functions) : Document
735 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
736 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
738 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
740 PR rtl-optimization/60969
741 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
742 constraints. Set up mem cost for NO_REGS case.
744 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
746 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
748 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
750 * config/darwin.c: Include "lto-section-names.h".
751 (LTO_SEGMENT_NAME): Don't define.
752 * config/i386/winnt.c: Include "lto-section-names.h".
753 * lto-streamer.c: Include "lto-section-names.h".
754 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
755 * lto-wrapper.c: Include "lto-section-names.h".
756 (LTO_SECTION_NAME_PREFIX): Don't define.
757 * lto-section-names.h: New file.
758 * cgraphunit.c: Include "lto-section-names.h".
760 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
762 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
764 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
767 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
769 2014-05-22 Nick Clifton <nickc@redhat.com>
771 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
773 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
775 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
776 -> (T)A transformation to integer types.
778 2014-05-22 Teresa Johnson <tejohnson@google.com>
780 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
781 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
782 (gcov_rewrite): Use gcov_nonruntime_assert.
784 (gcov_write_words): Ditto.
785 (gcov_write_length): Ditto.
786 (gcov_read_words): Use gcov_nonruntime_assert, and remove
787 gcc_assert from IN_LIBGCOV code.
788 (gcov_read_summary): Use gcov_error to flag profile corruption.
789 (gcov_sync): Use gcov_nonruntime_assert.
790 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
791 (gcov_histo_index): Use gcov_nonruntime_assert.
792 (static void gcov_histogram_merge): Ditto.
793 (compute_working_sets): Ditto.
794 * gcov-io.h (gcov_nonruntime_assert): Define.
795 (gcov_error): Define for !IN_LIBGCOV
797 2014-05-22 Richard Biener <rguenther@suse.de>
799 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
800 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
801 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
802 and deallocation site.
803 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
804 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
805 passing through the incoming points-to set.
806 (handle_lhs_call): Use flags argument instead of recomputing it.
807 (find_func_aliases_for_call): Call handle_lhs_call with proper
810 2014-05-22 Jakub Jelinek <jakub@redhat.com>
812 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
813 all padding bits in REAL_VALUE_TYPE are cleared.
815 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
817 Cleanup and improve multipass_dfa_lookahead_guard
818 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
819 (core2i7_first_cycle_multipass_begin,)
820 (core2i7_first_cycle_multipass_issue,)
821 (core2i7_first_cycle_multipass_backtrack): Update signature.
823 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
824 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
825 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
827 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
828 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
830 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
832 * doc/tm.texi: Regenerate.
834 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
835 * haifa-sched.c (ready_try): Make signed to allow negative values.
836 (rebug_ready_list_1): Update.
837 (choose_ready): Simplify.
838 (sched_extend_ready_list): Update.
840 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
842 Remove IA64 speculation tweaking flags
843 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
844 speculation tuning flags.
845 (msched-prefer-non-data-spec-insns,)
846 (msched-prefer-non-control-spec-insns): Obsolete options.
847 * haifa-sched.c (choose_ready): Remove handling of
848 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
849 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
850 and PREFER_NON_DATA_SPEC.
851 * sel-sched.c (process_spec_exprs): Remove handling of
852 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
854 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
856 Improve scheduling debug output
857 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
858 (advance_one_cycle): Update.
859 (schedule_insn, queue_to_ready): Add debug printouts.
860 (debug_ready_list_1): New static function.
861 (debug_ready_list): Update.
862 (max_issue): Add debug printouts.
863 (dump_insn_stream): New static function.
864 (schedule_block): Use it. Also better indent printouts.
866 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
868 Fix sched_insn debug counter
869 * haifa-sched.c (schedule_insn): Update.
870 (struct haifa_saved_data): Add nonscheduled_insns_begin.
871 (save_backtrack_point, restore_backtrack_point): Update.
872 (first_nonscheduled_insn): New static function.
873 (queue_to_ready, choose_ready): Use it.
874 (schedule_block): Init nonscheduled_insns_begin.
875 (sched_emit_insn): Update.
878 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
880 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
882 (aarch64_secondary_reload) : LikeWise.
883 (aarch64_class_max_nregs) : Remove CORE_REGS.
884 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
885 (REG_CLASS_NAMES) : Likewise.
886 (REG_CLASS_CONTENTS) : LikeWise.
887 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
889 2014-05-21 Guozhi Wei <carrot@google.com>
892 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
894 (vqdmulhq_n_s16): Likewise.
896 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
898 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
900 2014-05-21 Marek Polacek <polacek@redhat.com>
903 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
905 2014-05-21 Martin Jambor <mjambor@suse.cz>
907 * doc/invoke.texi (Optimize Options): Document parameters
908 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
909 ipa-cp-array-index-hint-bonus.
911 2014-05-21 Mark Wielaard <mjw@redhat.com>
914 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
915 version >= 3 or not strict DWARF.
916 * langhooks.h (struct lang_hooks_for_types): Add
917 enum_underlying_base_type.
918 * langhooks.c (lhd_enum_underlying_base_type): New function.
919 * gcc/langhooks.h (struct lang_hooks_for_types): Add
920 enum_underlying_base_type.
921 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
922 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
923 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
925 2014-05-21 Richard Biener <rguenther@suse.de>
927 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
929 2014-05-21 John Marino <gnugcc@marino.st>
931 * config.gcc (*-*-dragonfly*): New target.
932 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
933 * configure: Regenerate.
934 * config/dragonfly-stdint.h: New.
935 * config/dragonfly.h: New.
936 * config/dragonfly.opt: New.
937 * config/i386/dragonfly.h: New.
938 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
940 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
942 * tree.def (VOID_CST): New.
943 * tree-core.h (TI_VOID): New.
944 * tree.h (void_node): New.
945 * tree.c (tree_node_structure_for_code, tree_code_size)
946 (iterative_hash_expr): Handle VOID_CST.
947 (build_common_tree_nodes): Initialize void_node.
949 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
951 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
953 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
955 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
958 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
959 flag_reorder_blocks_and_partition.
960 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
962 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
965 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
967 (*addc_r_t): Add new insn_and_split.
969 2014-05-21 Jakub Jelinek <jakub@redhat.com>
972 * omp-low.c (handle_simd_reference): New function.
973 (lower_rec_input_clauses): Use it. Defer adding reference
974 initialization even for reduction without placeholder if in simd,
975 handle it properly later on.
977 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
979 PR tree-optimization/60899
980 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
981 assume all static symbols will have definition wile parsing and
982 check the do have definition later in compilation; check that
983 variable referring symbol will be output before concluding that
984 reference is safe; be conservative for referring local statics;
985 be more precise about when comdat is output in other partition.
987 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
990 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
992 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
993 (ipa_inline): Loop inline_to_all_callers until no more aliases
996 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
998 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
999 set writeonly flag only for vars actually written to.
1001 2014-05-20 Dehao Chen <dehao@google.com>
1003 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1004 and callee count to get clone count.
1005 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1008 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1010 PR rtl-optimization/61243
1011 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1013 2014-05-20 Xinliang David Li <davidxl@google.com>
1015 * cgraphunit.c (walk_polymorphic_call_targets): Add
1016 dbgcnt and fopt-info support.
1017 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1018 * ipa-devirt.c (ipa_devirt): Ditto.
1019 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1020 * ipa.c (walk_polymorphic_call_targets): Ditto.
1021 * gimple-fold.c (fold_gimple_assign): Ditto.
1022 (gimple_fold_call): Ditto.
1023 * dbgcnt.def: New counter.
1025 2014-05-20 DJ Delorie <dj@redhat.com>
1027 * config/msp430/msp430.md (split): Don't allow subregs when
1028 splitting SImode adds.
1029 (andneghi): Fix subtraction logic.
1030 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1032 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1034 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1036 * except.c (switch_to_exception_section, resolve_unique_section,
1037 get_named_text_section, default_function_rodata_section,
1038 align_variable, get_block_for_decl, default_section_type_flags):
1039 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1040 * symtab.c (symtab_add_to_same_comdat_group,
1041 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1042 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1044 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1045 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1046 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1047 (c6x_function_in_section_p): Likewise.
1048 * config/darwin.c (machopic_select_section): Likewise.
1049 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1050 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1051 * config/mep/mep.c (mep_select_section): LIkewise.
1052 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1054 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1056 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1057 EH region of calls to pure functions that can throw an exception.
1058 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1059 (copy_reference_ops_from_call): Also copy the EH region of the call if
1060 it can throw an exception.
1062 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1064 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1065 nested VEC_SELECTs that are inverses of each other.
1067 2014-05-20 Richard Biener <rguenther@suse.de>
1069 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1070 (extract_and_process_scc_for_name): not here.
1071 (cond_dom_walker::before_dom_children): Only process
1072 stmts that end the BB in interesting ways.
1073 (run_scc_vn): Mark param uses as visited.
1075 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1077 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1080 2014-05-20 Nick Clifton <nickc@redhat.com>
1082 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1083 (msp430_gimplify_va_arg_expr): New function.
1084 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1086 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1087 operand 0 in order to prevent confusion about the number of
1090 2014-05-20 Richard Biener <rguenther@suse.de>
1092 PR tree-optimization/61221
1093 * tree-ssa-pre.c (el_to_update): Remove.
1094 (eliminate_dom_walker::before_dom_children): Handle released
1095 VDEFs by value-numbering them to the associated VUSE. Update
1096 stmt immediately for substituted call address.
1097 (eliminate): Remove delayed stmt updating code.
1098 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1099 possibly late re-numbered vuses.
1100 (vn_reference_lookup_2): Adjust.
1101 (vn_reference_lookup_pieces): Likewise.
1102 (vn_reference_lookup): Likewise.
1104 2014-05-20 Richard Biener <rguenther@suse.de>
1106 * config.gcc: Remove need_64bit_hwint.
1107 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1108 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1110 * config.in: Regenerate.
1111 * configure: Likewise.
1113 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1115 * doc/extend.texi: Create Label Attributes section,
1116 move all label attributes into it and reference it.
1118 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1120 * arm.c (thumb1_reorg): When scanning backwards skip anything
1121 that's not a proper insn.
1123 2014-05-19 Richard Biener <rguenther@suse.de>
1125 PR tree-optimization/61221
1126 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1127 Do nothing for unreachable blocks.
1128 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1129 Improve unreachability detection.
1131 2014-05-19 Richard Biener <rguenther@suse.de>
1133 PR tree-optimization/61209
1134 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1136 2014-05-19 Nick Clifton <nickc@redhat.com>
1138 * except.c (init_eh): Fix computation of builtin setjmp buffer
1139 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1141 2014-05-19 Richard Biener <rguenther@suse.de>
1143 PR tree-optimization/61184
1144 * tree-vrp.c (is_negative_overflow_infinity): Use
1145 TREE_OVERFLOW_P and do that check first.
1146 (is_positive_overflow_infinity): Likewise.
1147 (is_overflow_infinity): Likewise.
1148 (vrp_operand_equal_p): Properly treat operands with
1149 differing overflow as not equal.
1151 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1153 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1154 shift simplification where it was intended.
1156 2014-05-19 Christian Bruel <christian.bruel@st.com>
1159 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1161 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1164 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1167 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1169 * reg-notes.def (CROSSING_JUMP): Likewise.
1170 * rtl.h (rtx_def): Update comment for jump flag.
1171 (CROSSING_JUMP_P): Define.
1172 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1173 of a REG_CROSSING_JUMP note.
1174 * cfghooks.c (tidy_fallthru_edges): Likewise.
1175 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1176 * emit-rtl.c (try_split): Likewise.
1177 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1178 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1179 * jump.c (redirect_jump_2): Likewise.
1180 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1181 (relax_delay_slots): Likewise.
1182 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1183 (bbit_di): Likewise.
1184 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1185 * config/sh/sh.md (jump_compact): Likewise.
1186 * bb-reorder.c (rotate_loop): Likewise.
1187 (pass_duplicate_computed_gotos::execute): Likewise.
1188 (add_reg_crossing_jump_notes): Rename to...
1189 (update_crossing_jump_flags): ...this.
1190 (pass_partition_blocks::execute): Update accordingly.
1192 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1194 * tree.h: Remove extraneous template <>.
1196 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1198 * ipa.c (symtab_remove_unreachable_nodes): Remove
1199 symbol from comdat group if its body was eliminated.
1200 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1201 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1202 (symtab_unregister_node): ... this one.
1203 (verify_symtab_base): More strict checking of comdats.
1204 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1206 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1208 * tree-pass.h (make_pass_ipa_comdats): New pass.
1209 * timevar.def (TV_IPA_COMDATS): New timevar.
1210 * passes.def (pass_ipa_comdats): Add.
1211 * Makefile.in (OBJS): Add ipa-comdats.o
1212 * ipa-comdats.c: New file.
1214 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1216 * ipa.c (update_visibility_by_resolution_info): New function.
1217 (function_and_variable_visibility): Use it.
1219 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1221 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1223 (FOR_EACH_DEFINED_SYMBOL): New macro.
1224 (varpool_first_static_initializer, varpool_next_static_initializer,
1225 varpool_first_defined_variable, varpool_next_defined_variable):
1227 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1229 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1231 * ggc-page.c (ggc_handle_finalizers): Add comment.
1233 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1235 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1236 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1237 (ggc_internal_cleared_alloc): Likewise.
1238 * ggc-page.c (finalizer): New class.
1239 (vec_finalizer): Likewise.
1240 (globals::finalizers): New member.
1241 (globals::vec_finalizers): Likewise.
1242 (ggc_internal_alloc): Record the finalizer if any for the block being
1244 (ggc_handle_finalizers): New function.
1245 (ggc_collect): Call ggc_handle_finalizers.
1246 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1248 (ggc_internal_cleared_alloc): Likewise.
1249 (finalize): New function.
1250 (need_finalization_p): Likewise.
1251 (ggc_alloc): Install the type's destructor as the finalizer if it
1253 (ggc_cleared_alloc): Likewise.
1254 (ggc_vec_alloc): Likewise.
1255 (ggc_cleared_vec_alloc): Likewise.
1257 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1259 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1261 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1263 * alias.c (record_alias_subset): Adjust.
1264 * bitmap.c (bitmap_element_allocate): Likewise.
1265 (bitmap_gc_alloc_stat): Likewise.
1266 * cfg.c (init_flow): Likewise.
1267 (alloc_block): Likewise.
1268 (unchecked_make_edge): Likewise.
1269 * cfgloop.c (alloc_loop): Likewise.
1270 (flow_loops_find): Likewise.
1271 (rescan_loop_exit): Likewise.
1272 * cfgrtl.c (init_rtl_bb_info): Likewise.
1273 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1274 (cgraph_allocate_node): Likewise.
1275 (cgraph_create_edge_1): Likewise.
1276 (cgraph_allocate_init_indirect_info): Likewise.
1277 * cgraphclones.c (cgraph_clone_edge): Likewise.
1278 * cgraphunit.c (add_asm_node): Likewise.
1279 (init_lowered_empty_function): Likewise.
1280 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1281 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1282 (alpha_use_linkage): Likewise.
1283 * config/arc/arc.c (arc_init_machine_status): Likewise.
1284 * config/arm/arm.c (arm_init_machine_status): Likewise.
1285 * config/avr/avr.c (avr_init_machine_status): Likewise.
1286 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1287 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1288 * config/cris/cris.c (cris_init_machine_status): Likewise.
1289 * config/darwin.c (machopic_indirection_name): Likewise.
1290 (darwin_build_constant_cfstring): Likewise.
1291 (darwin_enter_string_into_cfstring_table): Likewise.
1292 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1293 * config/frv/frv.c (frv_init_machine_status): Likewise.
1294 * config/i386/i386.c (get_dllimport_decl): Likewise.
1295 (ix86_init_machine_status): Likewise.
1296 (assign_386_stack_local): Likewise.
1297 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1298 (i386_pe_maybe_record_exported_symbol): Likewise.
1299 (i386_pe_record_stub): Likewise.
1300 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1301 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1302 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1303 (m32c_note_pragma_address): Likewise.
1304 * config/mep/mep.c (mep_init_machine_status): Likewise.
1305 (mep_note_pragma_flag): Likewise.
1306 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1307 (mips16_local_alias): Likewise.
1308 (mips_init_machine_status): Likewise.
1309 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1310 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1311 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1312 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1313 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1314 * config/pa/pa.c (pa_init_machine_status): Likewise.
1315 (pa_get_deferred_plabel): Likewise.
1316 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1317 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1318 (rs6000_init_machine_status): Likewise.
1319 (output_toc): Likewise.
1320 * config/s390/s390.c (s390_init_machine_status): Likewise.
1321 * config/score/score.c (score_output_external): Likewise.
1322 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1323 * config/spu/spu.c (spu_init_machine_status): Likewise.
1324 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1325 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1326 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1327 * coverage.c (coverage_end_function): Likewise.
1328 * dbxout.c (dbxout_init): Likewise.
1329 * doc/gty.texi: Don't mention variable_size attribute.
1330 * dwarf2cfi.c (new_cfi): Adjust.
1331 (new_cfi_row): Likewise.
1332 (copy_cfi_row): Likewise.
1333 (create_cie_data): Likewise.
1334 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1335 (new_loc_descr): Likewise.
1336 (find_AT_string_in_table): Likewise.
1337 (add_addr_table_entry): Likewise.
1338 (new_die): Likewise.
1339 (add_var_loc_to_decl): Likewise.
1340 (clone_die): Likewise.
1341 (clone_as_declaration): Likewise.
1342 (break_out_comdat_types): Likewise.
1343 (new_loc_list): Likewise.
1344 (add_loc_descr_to_each): Likewise.
1345 (add_location_or_const_value_attribute): Likewise.
1346 (add_linkage_name): Likewise.
1347 (lookup_filename): Likewise.
1348 (dwarf2out_var_location): Likewise.
1349 (new_line_info_table): Likewise.
1350 (dwarf2out_init): Likewise.
1351 (mem_loc_descriptor): Likewise.
1352 (loc_descriptor): Likewise.
1353 (add_const_value_attribute): Likewise.
1354 (tree_add_const_value_attribute): Likewise.
1355 (comp_dir_string): Likewise.
1356 (dwarf2out_vms_debug_main_pointer): Likewise.
1357 (string_cst_pool_decl): Likewise.
1358 * emit-rtl.c (set_mem_attrs): Likewise.
1359 (get_reg_attrs): Likewise.
1360 (start_sequence): Likewise.
1361 (init_emit): Likewise.
1362 (init_emit_regs): Likewise.
1363 * except.c (init_eh_for_function): Likewise.
1364 (gen_eh_region): Likewise.
1365 (gen_eh_region_catch): Likewise.
1366 (gen_eh_landing_pad): Likewise.
1367 (add_call_site): Likewise.
1368 * function.c (add_frame_space): Likewise.
1369 (insert_temp_slot_address): Likewise.
1370 (assign_stack_temp_for_type): Likewise.
1371 (get_hard_reg_initial_val): Likewise.
1372 (allocate_struct_function): Likewise.
1373 (prepare_function_start): Likewise.
1374 (types_used_by_var_decl_insert): Likewise.
1375 * gengtype.c (variable_size_p): Remove function.
1376 (enum alloc_quantity): Remove enum.
1377 (write_typed_alloc_def): Remove function.
1378 (write_typed_struct_alloc_def): Likewise.
1379 (write_typed_typedef_alloc_def): Likewise.
1380 (write_typed_alloc_defns): Likewise.
1382 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1383 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1384 * ggc.h (ggc_alloc): new function.
1385 (ggc_cleared_alloc): Likewise.
1386 (ggc_vec_alloc): Template on type of vector element, and remove
1387 element size argument.
1388 (ggc_cleared_vec_alloc): Likewise.
1389 * gimple.c (gimple_build_omp_for): Adjust.
1390 (gimple_copy): Likewise.
1391 * ipa-cp.c (get_replacement_map): Likewise.
1392 (find_aggregate_values_for_callers_subset): Likewise.
1393 (known_aggs_to_agg_replacement_list): Likewise.
1394 * ipa-devirt.c (get_odr_type): Likewise.
1395 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1396 (read_agg_replacement_chain): Likewise.
1397 * loop-iv.c (get_simple_loop_desc): Likewise.
1398 * lto-cgraph.c (input_node_opt_summary): Likewise.
1399 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1400 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1401 (input_eh_region): Likewise.
1402 (input_eh_lp): Likewise.
1403 (input_cfg): Likewise.
1404 * optabs.c (set_optab_libfunc): Likewise.
1405 (init_tree_optimization_optabs): Likewise.
1406 (set_conv_libfunc): Likewise.
1407 * passes.c (do_per_function_toporder): Likewise.
1408 * rtl.h: Don't use variable_size gty attribute.
1409 * sese.c (if_region_set_false_region): Adjust.
1410 * stringpool.c (gt_pch_save_stringpool): Likewise.
1411 * target-globals.c (save_target_globals): Likewise.
1412 * toplev.c (general_init): Likewise.
1413 * trans-mem.c (record_tm_replacement): Likewise.
1414 (split_bb_make_tm_edge): Likewise.
1415 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1416 * tree-data-ref.h (lambda_vector_new): Likewise.
1417 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1418 * tree-iterator.c (tsi_link_before): Likewise.
1419 (tsi_link_after): Likewise.
1420 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1421 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1422 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1423 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1424 * tree-ssa.c (init_tree_ssa): Adjust.
1425 * tree-ssanames.c (set_range_info): Likewise.
1426 (get_ptr_info): Likewise.
1427 (duplicate_ssa_name_ptr_info): Likewise.
1428 (duplicate_ssa_name_range_info): Likewise.
1429 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1430 (unpack_ts_fixed_cst_value_fields): Likewise.
1431 * tree.c (build_fixed): Likewise.
1432 (build_real): Likewise.
1433 (build_string): Likewise.
1434 (decl_priority_info): Likewise.
1435 (decl_debug_expr_insert): Likewise.
1436 (decl_value_expr_insert): Likewise.
1437 (decl_debug_args_insert): Likewise.
1438 (type_hash_add): Likewise.
1439 (build_omp_clause): Likewise.
1440 * ubsan.c (decl_for_type_insert): Likewise.
1441 * varasm.c (get_unnamed_section): Likewise.
1442 (get_noswitch_section): Likewise.
1443 (get_section): Likewise.
1444 (get_block_for_section): Likewise.
1445 (create_block_symbol): Likewise.
1446 (build_constant_desc): Likewise.
1447 (create_constant_pool): Likewise.
1448 (force_const_mem): Likewise.
1449 (record_tm_clone_pair): Likewise.
1450 * varpool.c (varpool_create_empty_node): Likewise.
1452 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1454 * dwarf2out.c (tree_add_const_value_attribute): Call
1455 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1456 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1457 instead of ggc_internal_<x>alloc_stat.
1458 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1459 (ggc_realloc): Likewise.
1460 * ggc-none.c (ggc_internal_alloc): Likewise.
1461 (ggc_internal_cleared_alloc): Likewise.
1462 * ggc-page.c: Likewise.
1463 * ggc.h (ggc_internal_alloc_stat): Likewise.
1464 (ggc_internal_alloc): Remove macro.
1465 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1466 (ggc_internal_cleared_alloc): Remove macro.
1467 (GGC_RESIZEVEC): Adjust.
1468 (ggc_resizevar): Remove macro.
1469 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1470 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1471 (ggc_internal_vec_cleared_alloc): Remove macro.
1472 (ggc_alloc_atomic_stat): Drop _stat suffix.
1473 (ggc_alloc_atomic): Remove macro.
1474 (ggc_alloc_cleared_atomic): Remove macro.
1475 (ggc_alloc_string_stat): Drop _stat suffix.
1476 (ggc_alloc_string): Remove macro.
1477 (ggc_alloc_rtx_def_stat): Adjust.
1478 (ggc_alloc_tree_node_stat): Likewise.
1479 (ggc_alloc_cleared_tree_node_stat): Likewise.
1480 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1481 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1482 * gimple.c (gimple_build_omp_for): Likewise.
1483 (gimple_copy): Likewise.
1484 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1485 * toplev.c (realloc_for_line_map): Adjust.
1486 * tree-data-ref.h (lambda_vector_new): Likewise.
1487 * tree-phinodes.c (allocate_phi_node): Likewise.
1488 * tree.c (grow_tree_vec_stat): Likewise.
1489 * vec.h (va_gc::reserve): Adjust.
1491 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1493 * config/microblaze/microblaze.c (break_handler): New Declaration.
1494 (microblaze_break_function_p,microblaze_is_break_handler): New.
1495 (compute_frame_size): Use microblaze_break_function_p.
1496 Add the test of break_handler.
1497 (microblaze_function_prologue) : Add the test of variable
1498 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1499 (microblaze_function_epilogue) : Add the test of break_handler.
1500 (microblaze_globalize_label) : Add the test of break_handler.
1501 Check the name by BREAK_HANDLER_NAME.
1503 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1505 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1506 microblaze_is_break_handler test.
1507 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1508 Use SYMBOL_REF_DECL.
1510 * config/microblaze/microblaze-protos.h
1511 (microblaze_break_function_p,microblaze_is_break_handler):
1514 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1515 new MicroBlaze break_handler functions.
1517 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1519 * doc/extend.texi (Size of an asm): Move node text according
1520 to its @menu entry position.
1522 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1524 PR tree-optimization/61140
1525 PR tree-optimization/61150
1526 PR tree-optimization/61197
1527 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1529 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1531 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1533 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1535 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1536 __SIZEOF_INT128__ is defined.
1538 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1540 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1541 (rs6000_delegitimize_address): Use it.
1543 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1545 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1546 inplace argument. Store the new address in the original MEM when true.
1547 * emit-rtl.c (change_address_1): Likewise.
1548 (adjust_address_1, adjust_automodify_address_1, offset_address):
1550 * rtl.h (plus_constant): Add an inplace argument.
1551 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1552 when true. Avoid generating (plus X (const_int 0)).
1553 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1554 in-place. Pass true to plus_constant.
1555 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1557 2014-05-16 Dehao Chen <dehao@google.com>
1559 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1561 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1564 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1566 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1568 2014-05-16 Dehao Chen <dehao@google.com>
1570 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1571 optimize_function_for_size_p.
1572 * regs.h (REG_FREQ_FROM_BB): Likewise.
1574 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1577 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1578 negt_reg_operand cases.
1579 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1581 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1583 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1585 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1588 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1591 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1593 * tree-cfg.c (dump_function_to_file): Dump the return type of
1594 functions, in a line to itself before the function body, mimicking
1595 the layout of a C function.
1597 2014-05-16 Dehao Chen <dehao@google.com>
1599 * cfghooks.c (make_forwarder_block): Use direct computation to
1600 get fall-through edge's count and frequency.
1602 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1604 * config/arc/arc.c (arc_init): Fix typo in error message.
1605 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1606 (split_stack_prologue_scratch_regno): Likewise.
1607 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1608 word from error message.
1610 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1612 * ira-costs.c: Fix typo in comment.
1614 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1616 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1618 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1620 * varpool.c (dump_varpool_node): Dump write-only flag.
1621 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1623 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1624 write-only variables.
1625 * ipa.c (process_references): New function.
1626 (set_readonly_bit): New function.
1627 (set_writeonly_bit): New function.
1628 (clear_addressable_bit): New function.
1629 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1630 fix handling of aliases.
1631 * cgraph.h (struct varpool_node): Add writeonly flag.
1633 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1635 PR rtl-optimization/60969
1636 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1637 Calculate costs for this case.
1639 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1641 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1642 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1644 2014-05-16 Richard Biener <rguenther@suse.de>
1646 PR tree-optimization/61194
1647 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1648 bool patterns ending in a COND_EXPR.
1650 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1652 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1654 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1656 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1657 where we were unable to cost an RTX.
1659 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1661 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1664 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1665 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1667 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1669 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1670 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1672 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1673 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1675 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1676 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1678 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1681 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1682 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1684 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1687 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1688 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1690 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1691 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1693 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1694 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1696 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1699 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1700 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1702 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1703 ZERO_EXTEND and SIGN_EXTEND better.
1705 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1706 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1708 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1711 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1712 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1714 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1715 costs when costing loads and stores to memory.
1717 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1718 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1720 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1723 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1725 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1727 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1728 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1730 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1732 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1734 (aarch64_rtx_mult_cost): New.
1735 (aarch64_rtx_costs): Use it, refactor as appropriate.
1737 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1738 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1740 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1741 emit instructions, return number of instructions which would
1743 (aarch64_add_constant): Update call to aarch64_build_constant.
1744 (aarch64_output_mi_thunk): Likewise.
1745 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1748 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1750 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1751 (TARGET_RTX_COSTS): Call it.
1753 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1755 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1756 (cortexa57_vector_cost): Likewise.
1757 (cortexa57_tunings): Use them.
1759 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1761 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1762 (cpu_addrcost_table): Use it.
1763 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1764 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1767 2014-05-16 Richard Biener <rguenther@suse.de>
1769 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1770 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1771 (visit_phi): Ignore edges marked as not executable.
1772 (class cond_dom_walker): New.
1773 (cond_dom_walker::before_dom_children): Value-number
1774 control statements and mark successor edges as not
1775 executable if possible.
1776 (run_scc_vn): First walk all control statements in
1777 dominator order, marking edges as not executable.
1778 * tree-inline.c (copy_edges_for_bb): Be not confused
1779 about random edge flags.
1781 2014-05-16 Richard Biener <rguenther@suse.de>
1783 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1785 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1788 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1789 (__TM_simple_begin): Use it.
1790 (__TM_begin): Likewise.
1792 2014-05-15 Martin Jambor <mjambor@suse.cz>
1795 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1796 type_preserved flag when the indirect edge is polymorphic.
1798 2014-05-15 Martin Jambor <mjambor@suse.cz>
1800 PR tree-optimization/61090
1801 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1802 build_ref_for_model.
1804 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1806 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1807 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1809 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1811 PR tree-optimization/61158
1812 * fold-const.c (fold_binary_loc): If X is zero-extended and
1813 shiftc >= prec, make sure zerobits is all ones instead of
1814 invoking undefined behavior.
1816 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1818 * regcprop.h: New file.
1819 * regcprop.c (skip_debug_insn_p): New decl.
1820 (replace_oldest_value_reg): Check skip_debug_insn_p.
1821 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1822 * shrink-wrap.c: Include regcprop.h.
1823 (prepare_shrink_wrap): Call
1824 copyprop_hardreg_forward_bb_without_debug_insn.
1826 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1828 * shrink-wrap.h: Update comment.
1829 * shrink-wrap.c: Update comment.
1830 (next_block_for_reg): Rename to live_edge_for_reg.
1831 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1832 (move_insn_for_shrink_wrap): Split live_edge.
1833 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1835 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1837 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1839 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1840 * config/sparc/sparc.md (fptype_ut699): New attribute.
1841 (in_branch_delay): Return false if -mfix-ut699 is specified and
1842 fptype_ut699 is set to single.
1843 (truncdfsf2): Add fptype_ut699 attribute.
1844 (fix_truncdfsi2): Likewise.
1845 (floatsisf2): Change fptype attribute.
1846 (fix_truncsfsi2): Likewise.
1847 (negtf2_notv9): Delete.
1848 (negtf2_v9): Likewise.
1849 (negtf2_hq): New instruction.
1850 (negtf2): New instruction and splitter.
1851 (negdf2_notv9): Rewrite.
1852 (abstf2_notv9): Delete.
1853 (abstf2_hq_v9): Likewise.
1854 (abstf2_v9): Likewise.
1855 (abstf2_hq): New instruction.
1856 (abstf2): New instruction and splitter.
1857 (absdf2_notv9): Rewrite.
1859 2014-05-14 Cary Coutant <ccoutant@google.com>
1862 * opts.c (common_handle_option): Don't special-case "-g".
1863 (set_debug_level): Default to at least level 2 with "-g".
1865 2014-05-14 DJ Delorie <dj@redhat.com>
1867 * config/msp430/msp430.c (msp430_builtin): Add
1868 MSP430_BUILTIN_DELAY_CYCLES.
1869 (msp430_init_builtins): Register void __delay_cycles(long long).
1870 (msp430_builtin_decl): Add it.
1871 (cg_magic_constant): New.
1872 (msp430_expand_delay_cycles): New.
1873 (msp430_expand_builtin): Call it.
1874 (msp430_print_operand_raw): Change integer printing from "int" to
1876 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1877 (delay_cycles_start): New.
1878 (delay_cycles_end): New.
1879 (delay_cycles_32): New.
1880 (delay_cycles_32x): New.
1881 (delay_cycles_16): New.
1882 (delay_cycles_16x): New.
1883 (delay_cycles_2): New.
1884 (delay_cycles_1): New.
1885 * doc/extend.texi: Document __delay_cycles().
1887 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1889 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1890 length attribute computation.
1892 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1895 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1897 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1900 * config/sparc/sparc.md: Fix types of low and high in DI constant
1901 splitter. Use gen_int_mode in some other splitters.
1903 2014-05-14 Martin Jambor <mjambor@suse.cz>
1906 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1908 2014-05-14 James Norris <jnorris@codesourcery.com>
1910 * omp-low.c (expand_parallel_call): Remove shadow variable.
1911 (expand_omp_taskreg): Likewise.
1913 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1915 * common/config/i386/i386-common.c
1916 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1917 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1918 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1919 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1920 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1921 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1922 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1923 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1924 xsavecintrin.h, xsavesintrin.h.
1925 (x86_64-*-*): Ditto.
1926 * config/i386/clflushoptintrin.h: New.
1927 * config/i386/xsavecintrin.h: Ditto.
1928 * config/i386/xsavesintrin.h: Ditto.
1929 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1930 (bit_XSAVES): Ditto.
1931 (bit_XSAVES): Ditto.
1932 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1933 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1935 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1936 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1937 OPTION_MASK_ISA_XSAVES.
1938 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1940 (PTA_CLFLUSHOPT) Define.
1941 (PTA_XSAVEC): Ditto.
1942 (PTA_XSAVES): Ditto.
1943 (ix86_option_override_internal): Handle new options.
1944 (ix86_valid_target_attribute_inner_p): Ditto.
1945 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1946 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1947 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1948 (bdesc_special_args): Add __builtin_ia32_xsaves,
1949 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1950 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1951 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1952 (ix86_expand_builtin): Handle new builtins.
1953 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1954 (TARGET_CLFLUSHOPT_P): Ditto.
1955 (TARGET_XSAVEC): Ditto.
1956 (TARGET_XSAVEC_P): Ditto.
1957 (TARGET_XSAVES): Ditto.
1958 (TARGET_XSAVES_P): Ditto.
1959 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1960 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1961 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1963 (ANY_XRSTOR64): Ditto.
1965 (xrstor): Change into <xrstor>.
1966 (xrstor_rex64): Change into <xrstor>_rex64.
1967 (xrstor64): Change into <xrstor>64
1969 * config/i386/i386.opt (mclflushopt): New.
1972 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1974 * doc/invoke.texi: Document new options.
1976 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1978 PR rtl-optimization/60866
1979 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1980 Default it to -1. Pass it down to init_simplejump_data.
1981 (init_simplejump_data): New parameter old_seqno. Pass it down
1982 to get_seqno_for_a_jump.
1983 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1984 initializing new jump seqno as a last resort. Add comment.
1985 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1986 jump and pass it down to sel_init_new_insn.
1987 (sel_redirect_edge_and_branch_force): Likewise.
1989 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1991 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1992 shifted values to avoid build warning.
1994 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1996 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1997 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1998 (cfg_layout_merge_blocks): Likewise.
1999 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2001 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2003 PR rtl-optimization/60901
2004 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2005 bb predecessor belongs to the same scheduling region. Adjust comment.
2007 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2009 * doc/sourcebuild.texi: (dfp_hw): Document.
2010 (p8vector_hw): Likewise.
2011 (powerpc_eabi_ok): Likewise.
2012 (powerpc_elfv2): Likewise.
2013 (powerpc_htm_ok): Likewise.
2014 (ppc_recip_hw): Likewise.
2017 2014-05-13 Cary Coutant <ccoutant@google.com>
2019 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2021 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2023 * gengtype-parse.c (require3): Eliminate in favor of...
2025 (require_template_declaration): Update to support optional single *
2028 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2029 (create_user_defined_type): Handle a single level of explicit
2030 pointerness within template arguments.
2031 (struct write_types_data): Add field "kind".
2032 (filter_type_name): Handle "*" character.
2033 (write_user_func_for_structure_ptr): Require a write_types_data
2034 rather than just a prefix string, so that we can look up the kind
2035 of the wtd and use it as an index into wrote_user_func_for_ptr,
2036 ensuring that such functions are written at most once. Support
2037 subclasses by invoking the marking function of the ultimate base class.
2038 (write_user_func_for_structure_body): Require a write_types_data
2039 rather than just a prefix string, so that we can pass this to
2040 write_user_func_for_structure_ptr.
2041 (write_func_for_structure): Likewise.
2042 (ggc_wtd): Add initializer of new "kind" field.
2043 (pch_wtd): Likewise.
2045 * gengtype.h (enum write_types_kinds): New.
2046 (struct type): Add field wrote_user_func_for_ptr to the "s"
2049 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2051 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2052 instead of const_binop.
2053 (fold_binary_loc): Likewise.
2055 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2057 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2058 calculation to match get_ref_base_and_extent.
2060 2014-05-13 Catherine Moore <clm@codesourcery.com>
2061 Sandra Loosemore <sandra@codesourcery.com>
2063 * configure.ac: Fix assembly for explicit JALR relocation check.
2064 * configure: Regenerate.
2066 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2068 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2069 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2070 Remove associated type declarations and initialisations.
2071 (arm_expand_neon_builtin): Likewise.
2072 (neon_emit_pair_result_insn): Delete.
2073 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2074 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2075 (neon_vzip<mode>): Likewise.
2076 (neon_vuzp<mode>): Likewise.
2078 2014-05-13 Richard Biener <rguenther@suse.de>
2081 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2082 it needs revisiting whether the call still may be tail-called.
2084 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2086 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2087 * rtl.h (block_symbol): Reduce number of fields to 2.
2088 (rtx_def): Add u2.symbol_ref_flags.
2089 (SYMBOL_REF_FLAGS): Use it.
2090 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2091 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2092 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2093 Lower index of SYMBOL_REF_DATA.
2094 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2095 Print SYMBOL_REF_FLAGS at the same time.
2096 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2098 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2100 * rtl.def (VAR_LOCATION): Remove "i" field.
2101 * rtl.h (rtx_def): Add u2.var_location_status.
2102 (PAT_VAR_LOCATION_STATUS): Use it.
2103 (gen_rtx_VAR_LOCATION): Declare.
2104 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2105 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2106 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2108 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2110 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2111 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2113 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2115 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2116 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2117 * rtl.h (rtx_def): Add insn_uid to u2 field.
2118 (RTX_FLAG_CHECK8): Delete in favor of...
2119 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2120 (INSN_DELETED_P): Update accordingly.
2121 (INSN_UID): Use u2.insn_uid.
2122 (INSN_CHAIN_CODE_P): Define.
2123 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2124 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2125 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2126 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2127 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2128 indices accordingly.
2129 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2130 Update indices for insn-chain rtxes.
2131 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2132 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2133 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2134 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2135 * combine.c (try_combine): Likewise.
2136 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2138 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2140 * rtl.def (REG): Remove middle field.
2141 * rtl.h (rtx_def): Add orignal_regno to u2.
2142 (ORIGINAL_REGNO): Use it instead of field 1.
2143 (REG_ATTRS): Lower field index accordingly.
2144 * gengtype.c (adjust_field_rtx_def): Remove handling of
2145 ORIGINAL_REGNO. Move REG_ATTRS index down.
2146 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2147 code that prints the REGNO.
2149 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2151 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2154 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2156 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2158 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2160 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2161 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2163 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2165 * config/aarch64/aarch64-protos.h
2166 (aarch64_hard_regno_caller_save_mode): New prototype.
2167 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2169 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2171 2014-05-13 Christian Bruel <christian.bruel@st.com>
2173 * target.def (mode_switching): New hook vector.
2174 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2175 (mode_exit, modepriority_to_mode): Likewise.
2176 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2177 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2178 * target.h: Include tm.h and hard-reg-set.h.
2179 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2180 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2181 * doc/tm.texi Regenerate.
2182 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2183 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2184 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2185 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2186 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2187 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2188 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2189 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2190 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2191 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2192 (ix86_emit_mode_set): Hookify.
2193 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2195 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2196 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2197 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2198 (epiphany_mode_priority_to_mode): Remove declaration.
2199 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2200 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2201 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2203 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2204 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2205 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2207 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2210 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2211 is const0_rtx, return immediately. Don't test count == 0 when
2214 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2216 * Makefile.in: add shrink-wrap.o.
2217 * config/i386/i386.c: include "shrink-wrap.h"
2218 * function.c: Likewise.
2219 (requires_stack_frame_p, next_block_for_reg,
2220 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2221 dup_block_and_redirect): Move to shrink-wrap.c
2222 (thread_prologue_and_epilogue_insns): Extract three code segments
2223 as functions in shrink-wrap.c
2224 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2226 * shrink-wrap.c: New file.
2227 * shrink-wrap.h: New file.
2229 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2231 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2232 reference to Solaris.
2234 2014-05-12 Mike Stump <mikestump@comcast.net>
2237 * genattrtab.c (filename): Add.
2238 (convert_set_attr_alternative): Improve error message.
2239 (check_defs): Restore read_md_filename for error messages.
2240 (gen_insn): Save filename.
2242 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2244 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2245 -fno-local-ivars and -fivar-visibility.
2246 * c-family/c.opt: Make -Wshadow also implicitly enable
2249 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2251 * doc/tm.texi: Remove reference to deleted macro.
2252 * doc/tm.texi.in: Likewise.
2254 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2257 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2260 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2263 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2264 * config/arm/aout.h (License): Same.
2265 * config/arm/bpabi.h (License): Same.
2266 * config/arm/elf.h (License): Same.
2267 * config/arm/linux-elf.h (License): Same.
2268 * config/arm/linux-gas.h (License): Same.
2269 * config/arm/netbsd-elf.h (License): Same.
2270 * config/arm/uclinux-eabi.h (License): Same.
2271 * config/arm/uclinux-elf.h (License): Same.
2272 * config/arm/vxworks.h (License): Same.
2274 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2276 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2277 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2278 number of operands to 3.
2279 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2280 * tree-nested.c (convert_nonlocal_omp_clauses,
2281 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2282 * gimplify.c (gimplify_scan_omp_clauses): Handle
2283 OMP_CLAUSE_LINEAR_STMT.
2284 * omp-low.c (lower_rec_input_clauses): Fix typo.
2285 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2286 cast between Fortran boolean_type_node and C _Bool if
2289 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2291 PR tree-optimization/61136
2292 * wide-int.h (multiple_of_p): Define a version that doesn't return
2294 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2295 integer_zerop/const_binop pair.
2296 (multiple_of_p): Likewise, converting both operands to widest_int
2299 2014-05-09 Teresa Johnson <tejohnson@google.com>
2301 * cgraphunit.c (analyze_functions): Use correct dump file.
2303 2014-05-09 Florian Weimer <fweimer@redhat.com>
2305 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2307 (stack_protect_return_slot_p): New function.
2308 (expand_used_vars): Call stack_protect_decl_p and
2309 stack_protect_return_slot_p for -fstack-protector-strong.
2311 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2314 * fold-const.c (fold_binary_loc): Changed width of mask.
2316 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2318 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2319 unsigned int initializers for regno_in, regno_out.
2321 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2324 * config/avr/avr.md (cc): Add new attribute set_vzn.
2325 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2326 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2327 with INC, DEC or NEG.
2328 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2329 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2330 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2332 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2335 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2337 * wide-int.cc (UTItype): Define.
2338 (UDWtype): Define for appropriate W_TYPE_SIZE.
2340 2014-05-09 Richard Biener <rguenther@suse.de>
2342 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2343 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2344 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2345 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2346 ssa_propagate): Adjust.
2348 2014-05-08 Jeff Law <law@redhat.com>
2350 PR tree-optimization/61009
2351 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2352 tri-state rather than a boolean. When a block is too big to
2353 thread through, inform caller via negative return value.
2354 (thread_across_edge): If a block was too big for normal threading,
2355 then it's too big for a joiner too, so remove temporary equivalences
2356 and return immediately.
2358 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2359 Matthias Klose <doko@ubuntu.com>
2362 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2364 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2367 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2369 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2372 * config/alpha/alpha.c: Include gimple-iterator.h.
2373 (alpha_gimple_fold_builtin): New function. Move
2374 ALPHA_BUILTIN_UMULH folding from ...
2375 (alpha_fold_builtin): ... here.
2376 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2378 2014-05-08 Wei Mi <wmi@google.com>
2381 * config/i386/i386.c (ix86_compute_frame_layout): Update
2382 preferred_stack_boundary for call, expanded from tls descriptor.
2383 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2384 to depend on SP register.
2385 (*tls_local_dynamic_base_32_gnu): Ditto.
2386 (*tls_local_dynamic_32_once): Ditto.
2387 (tls_global_dynamic_64_<mode>): Set
2388 ix86_tls_descriptor_calls_expanded_in_cfun.
2389 (tls_local_dynamic_base_64_<mode>): Ditto.
2390 (tls_global_dynamic_32): Set
2391 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2392 to depend on SP register.
2393 (tls_local_dynamic_base_32): Ditto.
2395 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2397 * config/arm/arm_neon.h: Update comment.
2398 * config/arm/neon-docgen.ml: Delete.
2399 * config/arm/neon-gen.ml: Delete.
2400 * doc/arm-neon-intrinsics.texi: Update comment.
2402 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2404 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2406 (vand, vorr, veor, vorn, vbic): Remove.
2407 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2409 (neon_vsub_unspec): Likewise.
2410 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2412 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2414 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2415 (vadd_s16): Likewise.
2416 (vadd_s32): Likewise.
2417 (vadd_f32): Likewise.
2418 (vadd_u8): Likewise.
2419 (vadd_u16): Likewise.
2420 (vadd_u32): Likewise.
2421 (vadd_s64): Likewise.
2422 (vadd_u64): Likewise.
2423 (vaddq_s8): Likewise.
2424 (vaddq_s16): Likewise.
2425 (vaddq_s32): Likewise.
2426 (vaddq_s64): Likewise.
2427 (vaddq_f32): Likewise.
2428 (vaddq_u8): Likewise.
2429 (vaddq_u16): Likewise.
2430 (vaddq_u32): Likewise.
2431 (vaddq_u64): Likewise.
2432 (vmul_s8): Likewise.
2433 (vmul_s16): Likewise.
2434 (vmul_s32): Likewise.
2435 (vmul_f32): Likewise.
2436 (vmul_u8): Likewise.
2437 (vmul_u16): Likewise.
2438 (vmul_u32): Likewise.
2439 (vmul_p8): Likewise.
2440 (vmulq_s8): Likewise.
2441 (vmulq_s16): Likewise.
2442 (vmulq_s32): Likewise.
2443 (vmulq_f32): Likewise.
2444 (vmulq_u8): Likewise.
2445 (vmulq_u16): Likewise.
2446 (vmulq_u32): Likewise.
2447 (vsub_s8): Likewise.
2448 (vsub_s16): Likewise.
2449 (vsub_s32): Likewise.
2450 (vsub_f32): Likewise.
2451 (vsub_u8): Likewise.
2452 (vsub_u16): Likewise.
2453 (vsub_u32): Likewise.
2454 (vsub_s64): Likewise.
2455 (vsub_u64): Likewise.
2456 (vsubq_s8): Likewise.
2457 (vsubq_s16): Likewise.
2458 (vsubq_s32): Likewise.
2459 (vsubq_s64): Likewise.
2460 (vsubq_f32): Likewise.
2461 (vsubq_u8): Likewise.
2462 (vsubq_u16): Likewise.
2463 (vsubq_u32): Likewise.
2464 (vsubq_u64): Likewise.
2465 (vand_s8): Likewise.
2466 (vand_s16): Likewise.
2467 (vand_s32): Likewise.
2468 (vand_u8): Likewise.
2469 (vand_u16): Likewise.
2470 (vand_u32): Likewise.
2471 (vand_s64): Likewise.
2472 (vand_u64): Likewise.
2473 (vandq_s8): Likewise.
2474 (vandq_s16): Likewise.
2475 (vandq_s32): Likewise.
2476 (vandq_s64): Likewise.
2477 (vandq_u8): Likewise.
2478 (vandq_u16): Likewise.
2479 (vandq_u32): Likewise.
2480 (vandq_u64): Likewise.
2481 (vorr_s8): Likewise.
2482 (vorr_s16): Likewise.
2483 (vorr_s32): Likewise.
2484 (vorr_u8): Likewise.
2485 (vorr_u16): Likewise.
2486 (vorr_u32): Likewise.
2487 (vorr_s64): Likewise.
2488 (vorr_u64): Likewise.
2489 (vorrq_s8): Likewise.
2490 (vorrq_s16): Likewise.
2491 (vorrq_s32): Likewise.
2492 (vorrq_s64): Likewise.
2493 (vorrq_u8): Likewise.
2494 (vorrq_u16): Likewise.
2495 (vorrq_u32): Likewise.
2496 (vorrq_u64): Likewise.
2497 (veor_s8): Likewise.
2498 (veor_s16): Likewise.
2499 (veor_s32): Likewise.
2500 (veor_u8): Likewise.
2501 (veor_u16): Likewise.
2502 (veor_u32): Likewise.
2503 (veor_s64): Likewise.
2504 (veor_u64): Likewise.
2505 (veorq_s8): Likewise.
2506 (veorq_s16): Likewise.
2507 (veorq_s32): Likewise.
2508 (veorq_s64): Likewise.
2509 (veorq_u8): Likewise.
2510 (veorq_u16): Likewise.
2511 (veorq_u32): Likewise.
2512 (veorq_u64): Likewise.
2513 (vbic_s8): Likewise.
2514 (vbic_s16): Likewise.
2515 (vbic_s32): Likewise.
2516 (vbic_u8): Likewise.
2517 (vbic_u16): Likewise.
2518 (vbic_u32): Likewise.
2519 (vbic_s64): Likewise.
2520 (vbic_u64): Likewise.
2521 (vbicq_s8): Likewise.
2522 (vbicq_s16): Likewise.
2523 (vbicq_s32): Likewise.
2524 (vbicq_s64): Likewise.
2525 (vbicq_u8): Likewise.
2526 (vbicq_u16): Likewise.
2527 (vbicq_u32): Likewise.
2528 (vbicq_u64): Likewise.
2529 (vorn_s8): Likewise.
2530 (vorn_s16): Likewise.
2531 (vorn_s32): Likewise.
2532 (vorn_u8): Likewise.
2533 (vorn_u16): Likewise.
2534 (vorn_u32): Likewise.
2535 (vorn_s64): Likewise.
2536 (vorn_u64): Likewise.
2537 (vornq_s8): Likewise.
2538 (vornq_s16): Likewise.
2539 (vornq_s32): Likewise.
2540 (vornq_s64): Likewise.
2541 (vornq_u8): Likewise.
2542 (vornq_u16): Likewise.
2543 (vornq_u32): Likewise.
2544 (vornq_u64): Likewise.
2546 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2548 * wide-int.cc (UTItype): Define.
2549 (UDWtype): Define for appropriate W_TYPE_SIZE.
2551 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2553 PR tree-optimization/59100
2554 * tree-ssa-phiopt.c: Include tree-inline.h.
2555 (neutral_element_p, absorbing_element_p): New functions.
2556 (value_replacement): Handle conditional binary operations with a
2557 neutral or absorbing element.
2559 2014-05-08 Richard Biener <rguenther@suse.de>
2561 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2562 folding the expression.
2563 (valueize_expr): Remove.
2564 (visit_reference_op_load): Do not valueize the result of
2566 (simplify_binary_expression): Likewise.
2567 (simplify_unary_expression): Likewise.
2569 2014-05-08 Richard Biener <rguenther@suse.de>
2571 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2572 looking at TYPE_ARG_TYPES.
2574 2014-05-08 Richard Biener <rguenther@suse.de>
2576 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2577 pointer propagation special-case.
2579 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2581 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2582 core part of address expressions.
2584 2014-05-08 Alan Modra <amodra@gmail.com>
2587 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2588 loads and stores when -mno-strict-align at any alignment.
2589 (expand_block_clear): Similarly. Also correct calculation of
2592 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2595 * tree-complex.c (expand_complex_move): Keep line info when expanding
2597 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2598 of complex expression. Use new argument to display correct location
2599 for values coming from phi statement.
2600 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2601 (warn_uninitialized_phi): Pass location of phi argument to
2603 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2604 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2606 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2608 * config/rs6000/predicates.md (indexed_address_mem): New.
2609 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2610 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2611 fpstore_ux, fpstore_u.
2612 (sign_extend, indexed, update): New.
2613 (cell_micro): Adjust.
2614 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2615 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2616 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2617 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2618 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2619 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2620 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2621 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2622 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2623 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2624 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2625 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2626 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2627 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2628 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2630 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2631 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2632 *vsx_extract_<mode>_store): Adjust.
2633 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2634 is_cracked_insn, insn_must_be_first_in_group,
2635 insn_must_be_last_in_group): Adjust.
2637 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2639 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2640 ppc440-fpstore): Adjust.
2641 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2642 ppc476-fpstore): Adjust.
2643 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2644 ppc601-fpstore): Adjust.
2645 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2647 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2649 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2650 ppc7450-fpstore): Adjust.
2651 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2652 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2653 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2655 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2656 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2657 cell-fpstore, cell-fpstore-update): Adjust.
2658 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2659 ppce300c3_store, ppce300c3_fpstore): Adjust.
2660 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2661 e500mc_fpstore): Adjust.
2662 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2663 e500mc64_store, e500mc64_fpstore): Adjust.
2664 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2665 e5500_fpstore): Adjust.
2666 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2667 e6500_fpstore): Adjust.
2668 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2670 * config/rs6000/power4.md (power4-load, power4-load-ext,
2671 power4-load-ext-update, power4-load-ext-update-indexed,
2672 power4-load-update-indexed, power4-load-update, power4-fpload,
2673 power4-fpload-update, power4-store, power4-store-update,
2674 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2676 * config/rs6000/power5.md (power5-load, power5-load-ext,
2677 power5-load-ext-update, power5-load-ext-update-indexed,
2678 power5-load-update-indexed, power5-load-update, power5-fpload,
2679 power5-fpload-update, power5-store, power5-store-update,
2680 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2682 * config/rs6000/power6.md (power6-load, power6-load-ext,
2683 power6-load-update, power6-load-update-indexed,
2684 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2685 power6-fpload-update, power6-store, power6-store-update,
2686 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2688 * config/rs6000/power7.md (power7-load, power7-load-ext,
2689 power7-load-update, power7-load-update-indexed,
2690 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2691 power7-fpload-update, power7-store, power7-store-update,
2692 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2694 * config/rs6000/power8.md (power8-load, power8-load-update,
2695 power8-load-ext, power8-load-ext-update, power8-fpload,
2696 power8-fpload-update, power8-store, power8-store-update-indexed,
2697 power8-fpstore, power8-fpstore-update): Adjust.
2698 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2700 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2701 titan_lsu_store, titan_lsu_fpstore): Adjust.
2702 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2704 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2707 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2708 unrolled byte insns. Emit address increments after move insns.
2710 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2712 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2713 const_gimple, rather than a gimple.
2714 (gimple_call_builtin_p): Likewise, for the three variants.
2716 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2717 (gimple_call_builtin_p): Likewise, for the three variants.
2719 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2721 PR tree-optimization/61095
2722 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2724 2014-05-07 Richard Biener <rguenther@suse.de>
2726 PR tree-optimization/61034
2727 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2728 (maybe_skip_until): Use translate to take into account
2729 lattices when trying to do disambiguations.
2730 (get_continuation_for_phi_1): Likewise.
2731 (get_continuation_for_phi): Adjust for added translate arguments.
2732 (walk_non_aliased_vuses): Likewise.
2733 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2734 (walk_non_aliased_vuses): Likewise.
2735 (call_may_clobber_ref_p_1): Declare.
2736 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2737 calls. Stop early if we are only supposed to disambiguate.
2738 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2740 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2742 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2743 Emit an error when the function has arguments.
2745 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2747 * cfgloop.h (unswitch_loops): Remove.
2748 * doc/passes.texi: Remove references to loop-unswitch.c
2749 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2751 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2753 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2754 check for loads group of length 3.
2755 (vect_permute_load_chain): New permutations for loads group of
2757 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2758 of vec_perm_shuffle for the new permutations.
2760 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2762 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2763 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2764 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2765 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2766 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2767 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2768 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2769 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2771 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2773 * loop-unswitch.c: Delete.
2775 2014-05-07 Richard Biener <rguenther@suse.de>
2777 * config.gcc: Always set need_64bit_hwint to yes.
2779 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2781 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2782 of using optimize_size.
2784 2014-05-06 Mike Stump <mikestump@comcast.net>
2786 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2788 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2790 * config/i386/sse.md (*mov<mode>_internal)
2791 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2792 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2793 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2794 (*<code><mode>3, *andnot<mode>3<mask_name>)
2795 (<mask_codefor><code><mode>3<mask_name>): Only consider
2796 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2798 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2801 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2803 * lra-constraints.c (valid_address_p): Move earlier in file.
2804 Add a constraint argument to the address_info version.
2805 (satisfies_memory_constraint_p): New function.
2806 (satisfies_address_constraint_p): Likewise.
2807 (process_alt_operands, curr_insn_transform): Use them.
2808 (process_address): Pass the constraint to valid_address_p when
2809 checking address operands.
2811 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2813 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2814 to their respective blocks. Fix inadvertent use of "node".
2816 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2818 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2820 (init_emit_once): ...here.
2821 * rtl.h (init_derived_machine_modes): Declare.
2822 * toplev.c (do_compile): Call it even if no_backend.
2824 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2825 Mike Stump <mikestump@comcast.net>
2826 Richard Sandiford <rdsandiford@googlemail.com>
2827 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2829 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2830 (rtx_equal_for_memref_p): Update comment.
2831 (adjust_offset_for_component_ref): Use wide-int interfaces.
2832 * builtins.c (get_object_alignment_2): Likewise.
2833 (c_readstr): Likewise.
2834 (target_char_cast): Add comment.
2835 (determine_block_size): Use wide-int interfaces.
2836 (expand_builtin_signbit): Likewise.
2837 (fold_builtin_int_roundingfn): Likewise.
2838 (fold_builtin_bitop): Likewise.
2839 (fold_builtin_bswap): Likewise.
2840 (fold_builtin_logarithm): Use signop.
2841 (fold_builtin_pow): Likewise.
2842 (fold_builtin_memory_op): Use wide-int interfaces.
2843 (fold_builtin_object_size): Likewise.
2844 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2845 nb_iterations_estimate.
2846 (record_niter_bound): Use wide-int interfaces.
2847 (get_estimated_loop_iterations_int): Likewise.
2848 (get_estimated_loop_iterations): Likewise.
2849 (get_max_loop_iterations): Likewise.
2850 * cfgloop.h: Include wide-int.h.
2851 (struct nb_iter_bound): Change bound to widest_int.
2852 (struct loop): Change nb_iterations_upper_bound and
2853 nb_iterations_estimate to widest_int.
2854 (record_niter_bound): Switch to use widest_int.
2855 (get_estimated_loop_iterations): Likewise.
2856 (get_max_loop_iterations): Likewise.
2857 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2858 update for wide-int.
2859 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2860 * combine.c (try_combine): Likewise.
2861 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2862 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2864 (aarch64_float_const_representable_p): Likewise.
2865 * config/arc/arc.c: Include wide-int.h.
2866 (arc_can_use_doloop_p): Use wide-int interfaces.
2867 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2868 (vfp3_const_double_index): Likewise.
2869 * config/avr/avr.c (avr_out_round): Likewise.
2870 (avr_fold_builtin): Likewise.
2871 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2872 (bfin_can_use_doloop_p): Likewise.
2873 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2874 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2875 * config/i386/i386.c: Include wide-int.h.
2876 (ix86_data_alignment): Use wide-int interfaces.
2877 (ix86_local_alignment): Likewise.
2878 (ix86_emit_swsqrtsf): Update real_from_integer.
2879 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2880 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2881 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2882 (zero_constant): Likewise.
2883 (input_operand): Likewise.
2884 (splat_input_operand): Likewise.
2885 (non_logical_cint_operand): Change const_double to const_wide_int.
2886 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2887 (easy_altivec_constant): Remove comment.
2888 (paired_expand_vector_init): Use CONSTANT_P.
2889 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2890 (rs6000_emit_move): Update checks.
2891 (rs6000_aggregate_candidate): Use wide-int interfaces.
2892 (rs6000_expand_ternop_builtin): Likewise.
2893 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2894 (rs6000_assemble_integer): Likewise.
2895 (rs6000_hash_constant): Likewise.
2896 (output_toc): Likewise.
2897 (rs6000_rtx_costs): Likewise.
2898 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2899 * config/rs6000/rs6000-c.c: Include wide-int.h.
2900 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2901 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2902 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2903 Handle CONST_WIDE_INT.
2904 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2905 Use tree_fits_uhwi_p.
2906 * config/sparc/sparc.c: Include wide-int.h.
2907 (sparc_fold_builtin): Use wide-int interfaces.
2908 * config/vax/vax.c: Include wide-int.h.
2909 (vax_float_literal): Use real_from_integer.
2910 * coretypes.h (struct hwivec_def): New.
2912 (const_hwivec): New.
2913 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2914 (equiv_constant): Handle CONST_WIDE_INT.
2915 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2916 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2917 * dbxout.c (stabstr_U): Use wide-int interfaces.
2918 (dbxout_type): Update to use cst_fits_shwi_p.
2919 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2920 (TARGET_SUPPORTS_WIDE_INT): Add default.
2921 * dfp.c: Include wide-int.h.
2922 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2923 decimal_real_to_integer.
2924 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2925 decimal_real_to_integer.
2926 * doc/generic.texi (Constant expressions): Update for wide_int.
2927 * doc/rtl.texi (const_double): Likewise.
2928 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2929 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2930 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2931 (REAL_VALUE_FROM_INT): Remove.
2932 (TARGET_SUPPORTS_WIDE_INT): New.
2933 * doc/tm.texi: Regenerate.
2934 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2935 * double-int.h: Include wide-int.h.
2936 (struct wi::int_traits): New.
2937 * dwarf2out.c (get_full_len): New.
2938 (dw_val_equal_p): Add case dw_val_class_wide_int.
2939 (size_of_loc_descr): Likewise.
2940 (output_loc_operands): Likewise.
2941 (insert_double): Remove.
2942 (insert_wide_int): New.
2944 (print_die): Add case dw_val_class_wide_int.
2945 (attr_checksum): Likewise.
2946 (attr_checksum_ordered): Likewise.
2947 (same_dw_val_p): Likewise.
2948 (size_of_die): Likewise.
2949 (value_format): Likewise.
2950 (output_die): Likewise.
2951 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2953 (clz_loc_descriptor): Use wide-int interfaces.
2954 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2955 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2956 (round_up_to_align): Use wide-int interfaces.
2957 (field_byte_offset): Likewise.
2958 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2959 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2960 CONST_DOUBLE handling. Use wide-int interfaces.
2961 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2962 (gen_enumeration_type_die): Use add_AT_wide.
2963 (hash_loc_operands): Add case dw_val_class_wide_int.
2964 (compare_loc_operands): Likewise.
2965 * dwarf2out.h: Include wide-int.h.
2966 (wide_int_ptr): New.
2967 (enum dw_val_class): Add dw_val_class_wide_int.
2968 (struct dw_val_struct): Add val_wide.
2969 * emit-rtl.c (const_wide_int_htab): New.
2970 (const_wide_int_htab_hash): New.
2971 (const_wide_int_htab_eq): New.
2972 (lookup_const_wide_int): New.
2973 (const_double_htab_hash): Use wide-int interfaces.
2974 (const_double_htab_eq): Likewise.
2975 (rtx_to_double_int): Conditionally compile for wide-int.
2976 (immed_double_int_const): Rename to immed_wide_int_const and
2977 update for wide-int.
2978 (immed_double_const): Conditionally compile for wide-int.
2979 (init_emit_once): Use wide-int interfaces.
2980 * explow.c (plus_constant): Likewise.
2981 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2982 (lshift_value): Use wide-int interfaces.
2983 (expand_mult): Likewise.
2984 (choose_multiplier): Likewise.
2985 (expand_smod_pow2): Likewise.
2986 (make_tree): Likewise.
2987 * expr.c (convert_modes): Consolidate handling of constants.
2988 Use wide-int interfaces.
2989 (emit_group_load_1): Add note.
2990 (store_expr): Update comment.
2991 (get_inner_reference): Use wide-int interfaces.
2992 (expand_constructor): Update comment.
2993 (expand_expr_real_2): Use wide-int interfaces.
2994 (expand_expr_real_1): Likewise.
2995 (reduce_to_bit_field_precision): Likewise.
2996 (const_vector_from_tree): Likewise.
2997 * final.c: Include wide-int-print.h.
2998 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2999 * fixed-value.c: Include wide-int.h.
3000 (fixed_from_string): Use wide-int interfaces.
3001 (fixed_to_decimal): Likewise.
3002 (fixed_convert_from_real): Likewise.
3003 (real_convert_from_fixed): Likewise.
3004 * fold-const.h (mem_ref_offset): Return an offset_int.
3005 (div_if_zero_remainder): Remove code parameter.
3006 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3007 Use wide-int interfaces.
3008 (may_negate_without_overflow_p): Use wide-int interfaces.
3009 (negate_expr_p): Likewise.
3010 (fold_negate_expr): Likewise.
3011 (int_const_binop_1): Likewise.
3012 (const_binop): Likewise.
3013 (fold_convert_const_int_from_int): Likewise.
3014 (fold_convert_const_int_from_real): Likewise.
3015 (fold_convert_const_int_from_fixed): Likewise.
3016 (fold_convert_const_fixed_from_int): Likewise.
3017 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3018 (sign_bit_p): Use wide-int interfaces.
3019 (make_range_step): Likewise.
3020 (build_range_check): Likewise. Pass an integer of the correct type
3021 instead of using integer_one_node.
3022 (range_predecessor): Pass an integer of the correct type instead
3023 of using integer_one_node.
3024 (range_successor): Likewise.
3025 (merge_ranges): Likewise.
3026 (unextend): Use wide-int interfaces.
3027 (extract_muldiv_1): Likewise.
3028 (fold_div_compare): Likewise.
3029 (fold_single_bit_test): Likewise.
3030 (fold_sign_changed_comparison): Likewise.
3031 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3032 (fold_plusminus_mult_expr): Use wide-int interfaces.
3033 (native_encode_int): Likewise.
3034 (native_interpret_int): Likewise.
3035 (fold_unary_loc): Likewise.
3036 (pointer_may_wrap_p): Likewise.
3037 (size_low_cst): Likewise.
3038 (mask_with_tz): Likewise.
3039 (fold_binary_loc): Likewise.
3040 (fold_ternary_loc): Likewise.
3041 (multiple_of_p): Likewise.
3042 (tree_call_nonnegative_warnv_p): Update calls to
3043 tree_int_cst_min_precision and real_from_integer.
3044 (fold_negate_const): Use wide-int interfaces.
3045 (fold_abs_const): Likewise.
3046 (fold_relational_const): Use tree_int_cst_lt.
3047 (round_up_loc): Use wide-int interfaces.
3048 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3049 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3050 * gengtype.c: Remove include of double-int.h.
3051 (do_typedef): Use wide-int interfaces.
3052 (open_base_files): Add wide-int.h.
3053 (main): Add offset_int and widest_int typedefs.
3054 * gengtype-lex.l: Handle "^".
3055 (CXX_KEYWORD): Add "static".
3056 * gengtype-parse.c (require3): New.
3057 (require_template_declaration): Handle constant template arguments
3058 and nested templates.
3059 * gengtype-state.c: Don't include "double-int.h".
3060 * genpreds.c (write_one_predicate_function): Update comment.
3061 (write_tm_constrs_h): Add check for hval and lval use in
3063 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3064 (add_to_sequence): Likewise.
3065 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3066 and const_double_operand.
3067 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3069 * gimple-fold.c (get_base_constructor): Likewise.
3070 (fold_array_ctor_reference): Likewise.
3071 (fold_nonarray_ctor_reference): Likewise.
3072 (fold_const_aggregate_ref_1): Likewise.
3073 (gimple_val_nonnegative_real_p): Likewise.
3074 (gimple_fold_indirect_ref): Likewise.
3075 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3076 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3077 (struct slsr_cand_d): Change index to be widest_int.
3078 (struct incr_info_d): Change incr to be widest_int.
3079 (alloc_cand_and_find_basis): Use wide-int interfaces.
3080 (slsr_process_phi): Likewise.
3081 (backtrace_base_for_ref): Likewise. Return a widest_int.
3082 (restructure_reference): Take a widest_int instead of a double_int.
3083 (slsr_process_ref): Use wide-int interfaces.
3084 (create_mul_ssa_cand): Likewise.
3085 (create_mul_imm_cand): Likewise.
3086 (create_add_ssa_cand): Likewise.
3087 (create_add_imm_cand): Take a widest_int instead of a double_int.
3088 (slsr_process_add): Use wide-int interfaces.
3089 (slsr_process_cast): Likewise.
3090 (slsr_process_copy): Likewise.
3091 (dump_candidate): Likewise.
3092 (dump_incr_vec): Likewise.
3093 (replace_ref): Likewise.
3094 (cand_increment): Likewise. Return a widest_int.
3095 (cand_abs_increment): Likewise.
3096 (replace_mult_candidate): Take a widest_int instead of a double_int.
3097 (replace_unconditional_candidate): Use wide-int interfaces.
3098 (incr_vec_index): Take a widest_int instead of a double_int.
3099 (create_add_on_incoming_edge): Likewise.
3100 (create_phi_basis): Use wide-int interfaces.
3101 (replace_conditional_candidate): Likewise.
3102 (record_increment): Take a widest_int instead of a double_int.
3103 (record_phi_increments): Use wide-int interfaces.
3104 (phi_incr_cost): Take a widest_int instead of a double_int.
3105 (lowest_cost_path): Likewise.
3106 (total_savings): Likewise.
3107 (analyze_increments): Use wide-int interfaces.
3108 (ncd_with_phi): Take a widest_int instead of a double_int.
3109 (ncd_of_cand_and_phis): Likewise.
3110 (nearest_common_dominator_for_cands): Likewise.
3111 (insert_initializers): Use wide-int interfaces.
3112 (all_phi_incrs_profitable): Likewise.
3113 (replace_one_candidate): Likewise.
3114 (replace_profitable_candidates): Likewise.
3115 * godump.c: Include wide-int-print.h.
3116 (go_output_typedef): Use wide-int interfaces.
3117 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3118 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3119 (build_loop_iteration_domains): Likewise.
3120 * hooks.h: Include wide-int.h rather than double-int.h.
3121 (hook_bool_dint_dint_uint_bool_true): Delete.
3122 (hook_bool_wint_wint_uint_bool_true): Declare.
3123 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3124 (hook_bool_wint_wint_uint_bool_true): New.
3125 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3127 (ubsan_expand_si_overflow_mul_check): Likewise.
3128 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3129 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3130 (get_ancestor_addr_info): Likewise.
3131 (ipa_modify_call_arguments): Likewise.
3132 * loop-doloop.c (doloop_modify): Likewise.
3133 (doloop_optimize): Likewise.
3134 * loop-iv.c (iv_number_of_iterations): Likewise.
3135 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3136 (unroll_loop_constant_iterations): Likewise.
3137 (decide_unroll_runtime_iterations): Likewise.
3138 (unroll_loop_runtime_iterations): Likewise.
3139 (decide_peel_simple): Likewise.
3140 (decide_unroll_stupid): Likewise.
3141 * lto-streamer-in.c (streamer_read_wi): Add.
3142 (input_cfg): Use wide-int interfaces.
3143 (lto_input_tree_1): Likewise.
3144 * lto-streamer-out.c (streamer_write_wi): Add.
3145 (hash_tree): Use wide-int interfaces.
3146 (output_cfg): Likewise.
3147 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3148 (GTFILES): Add wide-int.h and signop.h.
3149 (TAGS): Look for .cc files too.
3150 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3151 * optabs.c (expand_subword_shift): Likewise.
3152 (expand_doubleword_shift): Likewise.
3153 (expand_absneg_bit): Likewise.
3154 (expand_copysign_absneg): Likewise.
3155 (expand_copysign_bit): Likewise.
3156 * postreload.c (reload_cse_simplify_set): Likewise.
3157 * predict.c (predict_iv_comparison): Likewise.
3158 * pretty-print.h: Include wide-int-print.h.
3160 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3161 * print-tree.c: Include wide-int-print.h.
3162 (print_node_brief): Use wide-int interfaces.
3163 (print_node): Likewise.
3164 * read-rtl.c (validate_const_wide_int): New.
3165 (read_rtx_code): Add CONST_WIDE_INT case.
3166 * real.c: Include wide-int.h.
3167 (real_to_integer2): Delete.
3168 (real_to_integer): New function, returning a wide_int.
3169 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3170 (ten_to_ptwo): Update call to real_from_integer.
3171 (real_digit): Likewise.
3172 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3173 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3174 (REAL_VALUE_TO_INT): Delete.
3175 (real_to_integer): Declare a wide-int form.
3176 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3177 * recog.c (const_int_operand): Improve comment.
3178 (const_scalar_int_operand): New.
3179 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3180 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3181 (split_double): Likewise.
3182 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3183 (rtx_size): Likewise.
3184 (rtx_alloc_stat_v): New.
3185 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3186 (cwi_output_hex): New.
3187 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3188 (cwi_check_failed_bounds): New.
3189 * rtl.def (CONST_WIDE_INT): New.
3190 * rtl.h: Include <utility> and wide-int.h.
3191 (struct hwivec_def): New.
3192 (CWI_GET_NUM_ELEM): New.
3193 (CWI_PUT_NUM_ELEM): New.
3194 (struct rtx_def): Add num_elem and hwiv.
3195 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3196 (CASE_CONST_UNIQUE): Likewise.
3197 (CASE_CONST_ANY): Likewise.
3198 (CONST_SCALAR_INT_P): Likewise.
3199 (CONST_WIDE_INT_P): New.
3201 (HWIVEC_CHECK): New.
3202 (cwi_check_failed_bounds): New.
3204 (HWIVEC_CHECK): New.
3205 (CONST_WIDE_INT_VEC) New.
3206 (CONST_WIDE_INT_NUNITS) New.
3207 (CONST_WIDE_INT_ELT) New.
3208 (rtx_mode_t): New type.
3209 (wi::int_traits <rtx_mode_t>): New.
3211 (wi::min_value): New.
3212 (wi::max_value): New.
3214 (const_wide_int_alloc): New.
3215 (immed_wide_int_const): New.
3216 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3217 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3218 * signop.h: New file.
3219 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3220 (simplify_const_unary_operation): Use wide-int interfaces.
3221 (simplify_binary_operation_1): Likewise.
3222 (simplify_const_binary_operation): Likewise.
3223 (simplify_const_relational_operation): Likewise.
3224 (simplify_immed_subreg): Likewise.
3225 * stmt.c (expand_case): Likewise.
3226 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3227 signop rather than a bool.
3228 * stor-layout.c (layout_type): Use wide-int interfaces.
3229 (initialize_sizetypes): Update calls to
3230 set_min_and_max_values_for_integral_type.
3231 (set_min_and_max_values_for_integral_type): Take a signop rather
3232 than a bool. Use wide-int interfaces.
3233 (fixup_signed_type): Update accordingly. Remove
3234 HOST_BITS_PER_DOUBLE_INT limit.
3235 (fixup_unsigned_type): Likewise.
3236 * system.h (STATIC_CONSTANT_P): New.
3237 (STATIC_ASSERT): New.
3238 * target.def (can_use_doloop_p): Take widest_ints rather than
3240 * target.h: Include wide-int.h rather than double-int.h.
3241 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3243 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3244 rather than INT_CST_LT_UNSIGNED.
3245 (can_use_doloop_if_innermost): Take widest_ints rather than
3247 * tree-affine.c: Include wide-int-print.h.
3248 (double_int_ext_for_comb): Delete.
3249 (wide_int_ext_for_comb): New.
3250 (aff_combination_zero): Use wide-int interfaces.
3251 (aff_combination_const): Take a widest_int instead of a double_int.
3252 (aff_combination_elt): Use wide-int interfaces.
3253 (aff_combination_scale): Take a widest_int instead of a double_int.
3254 (aff_combination_add_elt): Likewise.
3255 (aff_combination_add_cst): Likewise.
3256 (aff_combination_add): Use wide-int interfaces.
3257 (aff_combination_convert): Likewise.
3258 (tree_to_aff_combination): Likewise.
3259 (add_elt_to_tree): Take a widest_int instead of a double_int.
3260 (aff_combination_to_tree): Use wide-int interfaces.
3261 (aff_combination_remove_elt): Likewise.
3262 (aff_combination_add_product): Take a widest_int instead of
3264 (aff_combination_mult): Use wide-int interfaces.
3265 (aff_combination_expand): Likewise.
3266 (double_int_constant_multiple_p): Delete.
3267 (wide_int_constant_multiple_p): New.
3268 (aff_combination_constant_multiple_p): Take a widest_int pointer
3269 instead of a double_int pointer.
3270 (print_aff): Use wide-int interfaces.
3271 (get_inner_reference_aff): Take a widest_int pointer
3272 instead of a double_int pointer.
3273 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3274 * tree-affine.h: Include wide-int.h.
3275 (struct aff_comb_elt): Change type of coef to widest_int.
3276 (struct affine_tree_combination): Change type of offset to widest_int.
3277 (double_int_ext_for_comb): Delete.
3278 (wide_int_ext_for_comb): New.
3279 (aff_combination_const): Use widest_int instead of double_int.
3280 (aff_combination_scale): Likewise.
3281 (aff_combination_add_elt): Likewise.
3282 (aff_combination_constant_multiple_p): Likewise.
3283 (get_inner_reference_aff): Likewise.
3284 (aff_comb_cannot_overlap_p): Likewise.
3285 (aff_combination_zero_p): Use wide-int interfaces.
3286 * tree.c: Include tree.h.
3287 (init_ttree): Use make_int_cst.
3288 (tree_code_size): Removed code for INTEGER_CST case.
3289 (tree_size): Add INTEGER_CST case.
3290 (make_node_stat): Update comment.
3291 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3292 (build_int_cst_type): Use wide-int interfaces.
3293 (double_int_to_tree): Likewise.
3294 (double_int_fits_to_tree_p): Delete.
3295 (force_fit_type_double): Delete.
3296 (force_fit_type): New.
3297 (int_cst_hash_hash): Use wide-int interfaces.
3298 (int_cst_hash_eq): Likewise.
3299 (build_int_cst_wide): Delete.
3300 (wide_int_to_tree): New.
3301 (cache_integer_cst): Use wide-int interfaces.
3302 (build_low_bits_mask): Likewise.
3303 (cst_and_fits_in_hwi): Likewise.
3304 (real_value_from_int_cst): Likewise.
3305 (make_int_cst_stat): New.
3306 (integer_zerop): Use wide_int interfaces.
3307 (integer_onep): Likewise.
3308 (integer_all_onesp): Likewise.
3309 (integer_pow2p): Likewise.
3310 (integer_nonzerop): Likewise.
3311 (tree_log2): Likewise.
3312 (tree_floor_log2): Likewise.
3313 (tree_ctz): Likewise.
3314 (int_size_in_bytes): Likewise.
3315 (mem_ref_offset): Return an offset_int rather than a double_int.
3316 (build_type_attribute_qual_variant): Use wide_int interfaces.
3317 (type_hash_eq): Likewise
3318 (tree_int_cst_equal): Likewise.
3319 (tree_int_cst_lt): Delete.
3320 (tree_int_cst_compare): Likewise.
3321 (tree_fits_shwi_p): Use wide_int interfaces.
3322 (tree_fits_uhwi_p): Likewise.
3323 (tree_int_cst_sign_bit): Likewise.
3324 (tree_int_cst_sgn): Likewise.
3325 (tree_int_cst_min_precision): Take a signop rather than a bool.
3326 (simple_cst_equal): Use wide_int interfaces.
3327 (compare_tree_int): Likewise.
3328 (iterative_hash_expr): Likewise.
3329 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3331 (get_type_static_bounds): Use wide_int interfaces.
3332 (tree_int_cst_elt_check_failed): New.
3333 (build_common_tree_nodes): Reordered to set prec before filling in
3335 (int_cst_value): Check cst_and_fits_in_hwi.
3336 (widest_int_cst_value): Use wide_int interfaces.
3337 (upper_bound_in_type): Likewise.
3338 (lower_bound_in_type): Likewise.
3339 (num_ending_zeros): Likewise.
3340 (drop_tree_overflow): Likewise.
3341 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3342 (gen_conditions_for_pow_cst_base): Likewise.
3343 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3344 (group_case_labels_stmt): Use wide-int interfaces.
3345 (verify_gimple_assign_binary): Likewise.
3346 (print_loop): Likewise.
3347 * tree-chrec.c (tree_fold_binomial): Likewise.
3348 * tree-core.h (struct tree_base): Add int_length.
3349 (struct tree_int_cst): Change rep of value.
3350 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3351 (dr_may_alias_p): Likewise.
3352 (max_stmt_executions_tree): Likewise.
3353 * tree.def (INTEGER_CST): Update comment.
3354 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3355 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3356 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3357 (dequeue_and_dump): Use wide-int interfaces.
3358 * tree.h: Include wide-int.h.
3359 (NULL_TREE): Moved to earlier loc in file.
3360 (TREE_INT_CST_ELT_CHECK): New.
3361 (tree_int_cst_elt_check_failed): New.
3363 (TREE_INT_CST): Delete.
3364 (TREE_INT_CST_LOW): Use wide-int interfaces.
3365 (TREE_INT_CST_HIGH): Delete.
3366 (TREE_INT_CST_NUNITS): New.
3367 (TREE_INT_CST_EXT_NUNITS): Likewise.
3368 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3369 (TREE_INT_CST_ELT): Likewise.
3370 (INT_CST_LT): Delete.
3371 (tree_int_cst_elt_check): New (two forms).
3372 (type_code_size): Update comment.
3373 (make_int_cst_stat, make_int_cst): New.
3374 (tree_to_double_int): Delete.
3375 (double_int_fits_to_tree_p): Delete.
3376 (force_fit_type_double): Delete.
3377 (build_int_cstu): Replace with out-of-line function.
3378 (build_int_cst_wide): Delete.
3379 (tree_int_cst_lt): Define inline.
3380 (tree_int_cst_le): New.
3381 (tree_int_cst_compare): Define inline.
3382 (tree_int_cst_min_precision): Take a signop rather than a bool.
3383 (wi::int_traits <const_tree>): New.
3384 (wi::int_traits <tree>): New.
3385 (wi::extended_tree): New.
3386 (wi::int_traits <wi::extended_tree>): New.
3387 (wi::to_widest): New.
3388 (wi::to_offset): New.
3389 (wi::fits_to_tree_p): New.
3390 (wi::min_value): New.
3391 (wi::max_value): New.
3392 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3393 (copy_tree_body_r): Likewise.
3394 * tree-object-size.c (compute_object_offset): Likewise.
3395 (addr_object_size): Likewise.
3396 * tree-predcom.c: Include wide-int-print.h.
3397 (struct dref_d): Change type of offset to widest_int.
3398 (dump_dref): Call wide-int printer.
3399 (aff_combination_dr_offset): Use wide-int interfaces.
3400 (determine_offset): Take a widest_int pointer rather than a
3402 (split_data_refs_to_components): Use wide-int interfaces.
3403 (suitable_component_p): Likewise.
3404 (order_drefs): Likewise.
3405 (add_ref_to_chain): Likewise.
3406 (valid_initializer_p): Likewise.
3407 (determine_roots_comp): Likewise.
3408 * tree-pretty-print.c: Include wide-int-print.h.
3409 (dump_generic_node): Use wide-int interfaces.
3410 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3411 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3412 (move_fixed_address_to_symbol): Likewise.
3413 (move_hint_to_base): Likewise.
3414 (move_pointer_to_base): Likewise.
3415 (move_variant_to_index): Likewise.
3416 (most_expensive_mult_to_index): Likewise.
3417 (addr_to_parts): Likewise.
3418 (copy_ref_info): Likewise.
3419 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3420 (indirect_refs_may_alias_p): Likewise.
3421 (stmt_kills_ref_p_1): Likewise.
3422 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3423 * tree-ssa-ccp.c: Update comment at top of file. Include
3425 (struct prop_value_d): Change type of mask to widest_int.
3426 (extend_mask): New function.
3427 (dump_lattice_value): Use wide-int interfaces.
3428 (get_default_value): Likewise.
3429 (set_constant_value): Likewise.
3430 (set_value_varying): Likewise.
3431 (valid_lattice_transition): Likewise.
3432 (set_lattice_value): Likewise.
3433 (value_to_double_int): Delete.
3434 (value_to_wide_int): New.
3435 (get_value_from_alignment): Use wide-int interfaces.
3436 (get_value_for_expr): Likewise.
3437 (do_dbg_cnt): Likewise.
3438 (ccp_finalize): Likewise.
3439 (ccp_lattice_meet): Likewise.
3440 (bit_value_unop_1): Use widest_ints rather than double_ints.
3441 (bit_value_binop_1): Likewise.
3442 (bit_value_unop): Use wide-int interfaces.
3443 (bit_value_binop): Likewise.
3444 (bit_value_assume_aligned): Likewise.
3445 (evaluate_stmt): Likewise.
3446 (ccp_fold_stmt): Likewise.
3447 (visit_cond_stmt): Likewise.
3448 (ccp_visit_stmt): Likewise.
3449 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3450 (constant_pointer_difference): Likewise.
3451 (associate_pointerplus): Likewise.
3452 (combine_conversions): Likewise.
3453 * tree-ssa-loop.h: Include wide-int.h.
3454 (struct tree_niter_desc): Change type of max to widest_int.
3455 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3456 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3457 (remove_redundant_iv_tests): Likewise.
3458 (canonicalize_loop_induction_variables): Likewise.
3459 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3460 (constant_multiple_of): Take a widest_int pointer instead of
3461 a double_int pointer.
3462 (get_computation_aff): Use wide-int interfaces.
3463 (ptr_difference_cost): Likewise.
3464 (difference_cost): Likewise.
3465 (get_loop_invariant_expr_id): Likewise.
3466 (get_computation_cost_at): Likewise.
3467 (iv_elimination_compare_lt): Likewise.
3468 (may_eliminate_iv): Likewise.
3469 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3470 instead of double_int.
3471 (max_loop_iterations): Likewise.
3472 (max_stmt_executions): Likewise.
3473 (estimated_stmt_executions): Likewise.
3474 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3475 (split_to_var_and_offset): Use wide-int interfaces.
3476 (determine_value_range): Likewise.
3477 (bound_difference_of_offsetted_base): Likewise.
3478 (bounds_add): Take a widest_int instead of a double_int.
3479 (number_of_iterations_ne_max): Use wide-int interfaces.
3480 (number_of_iterations_ne): Likewise.
3481 (number_of_iterations_lt_to_ne): Likewise.
3482 (assert_loop_rolls_lt): Likewise.
3483 (number_of_iterations_lt): Likewise.
3484 (number_of_iterations_le): Likewise.
3485 (number_of_iterations_cond): Likewise.
3486 (number_of_iterations_exit): Likewise.
3487 (finite_loop_p): Likewise.
3488 (derive_constant_upper_bound_assign): Likewise.
3489 (derive_constant_upper_bound): Return a widest_int.
3490 (derive_constant_upper_bound_ops): Likewise.
3491 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3492 (record_estimate): Take a widest_int rather than a double_int.
3493 (record_nonwrapping_iv): Use wide-int interfaces.
3494 (double_int_cmp): Delete.
3495 (wide_int_cmp): New.
3496 (bound_index): Take a widest_int rather than a double_int.
3497 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3498 (maybe_lower_iteration_bound): Likewise.
3499 (estimate_numbers_of_iterations_loop): Likewise.
3500 (estimated_loop_iterations): Take a widest_int pointer than than
3501 a double_int pointer.
3502 (estimated_loop_iterations_int): Use wide-int interfaces.
3503 (max_loop_iterations): Take a widest_int pointer than than
3504 a double_int pointer.
3505 (max_loop_iterations_int): Use wide-int interfaces.
3506 (max_stmt_executions): Take a widest_int pointer than than
3507 a double_int pointer.
3508 (estimated_stmt_executions): Likewise.
3509 (n_of_executions_at_most): Use wide-int interfaces.
3510 (scev_probably_wraps_p): Likewise.
3511 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3513 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3515 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3516 double_ints. Adjust for trailing_wide_ints <3> representation.
3517 (set_nonzero_bits): Likewise.
3518 (get_range_info): Return wide_ints rather than double_ints.
3519 Adjust for trailing_wide_ints <3> representation.
3520 (get_nonzero_bits): Likewise.
3521 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3523 * tree-ssanames.h (struct range_info_def): Replace min, max and
3524 nonzero_bits with a trailing_wide_ints <3>.
3525 (set_range_info): Use wide_int_refs rather than double_ints.
3526 (set_nonzero_bits): Likewise.
3527 (get_range_info): Return wide_ints rather than double_ints.
3528 (get_nonzero_bits): Likewise.
3529 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3530 * tree-ssa-pre.c (phi_translate_1): Likewise.
3531 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3532 (acceptable_pow_call): Likewise.
3533 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3535 (vn_reference_fold_indirect): Likewise.
3536 (vn_reference_maybe_forwprop_address): Likewise.
3537 (valueize_refs_1): Likewise.
3538 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3539 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3540 tree_int_cst_lt and tree_int_cst_le.
3541 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3543 (streamer_alloc_tree): Likewise.
3544 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3545 (streamer_write_tree_header): Likewise.
3546 (streamer_write_integer_cst): Likewise.
3547 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3548 (build_constructors): Likewise.
3549 (array_value_type): Likewise.
3550 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3551 (vect_check_gather): Likewise.
3552 * tree-vect-generic.c (build_replicated_const): Likewise.
3553 (expand_vector_divmod): Likewise.
3554 * tree-vect-loop.c (vect_transform_loop): Likewise.
3555 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3556 (vect_do_peeling_for_alignment): Likewise.
3557 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3558 * tree-vrp.c: Include wide-int.h.
3559 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3560 (extract_range_from_assert): Use wide-int interfaces.
3561 (vrp_int_const_binop): Likewise.
3562 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3563 double_int pointers.
3564 (ranges_from_anti_range): Use wide-int interfaces.
3565 (quad_int_cmp): Delete.
3566 (quad_int_pair_sort): Likewise.
3567 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3568 (extract_range_from_unary_expr_1): Likewise.
3569 (adjust_range_with_scev): Likewise.
3570 (masked_increment): Take and return wide_ints rather than double_ints.
3571 (register_edge_assert_for_2): Use wide-int interfaces.
3572 (check_array_ref): Likewise.
3573 (search_for_addr_array): Likewise.
3574 (maybe_set_nonzero_bits): Likewise.
3575 (union_ranges): Pass an integer of the correct type instead of
3576 using integer_one_node.
3577 (intersect_ranges): Likewise.
3578 (simplify_truth_ops_using_ranges): Likewise.
3579 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3580 (range_fits_type_p): Likewise.
3581 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3583 (simplify_conversion_using_ranges): Use wide-int interfaces.
3584 (simplify_float_conversion_using_ranges): Likewise.
3585 (vrp_finalize): Likewise.
3586 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3587 (gimple_stringops_transform): Likewise.
3588 * varasm.c (decode_addr_const): Likewise.
3589 (const_hash_1): Likewise.
3590 (const_rtx_hash_1): Likewise
3591 (output_constant): Likewise.
3592 (array_size_for_constructor): Likewise.
3593 (output_constructor_regular_field): Likewise.
3594 (output_constructor_bitfield): Likewise.
3595 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3596 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3598 * gencheck.c: Define BITS_PER_UNIT.
3601 * wide-int-print.cc: New.
3602 * wide-int-print.h: New.
3604 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3606 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3608 2014-05-06 Richard Biener <rguenther@suse.de>
3610 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3611 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3612 (TODO_verify_all): Adjust.
3613 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3614 TODO_verify_stmts and TODO_verify_rtl_sharing.
3615 * bb-reorder.c: Likewise.
3616 * cfgexpand.c: Likewise.
3617 * cprop.c: Likewise.
3619 * function.c: Likewise.
3620 * fwprop.c: Likewise.
3622 * gimple-ssa-isolate-paths.c: Likewise.
3623 * gimple-ssa-strength-reduction.c: Likewise.
3624 * ipa-split.c: Likewise.
3625 * loop-init.c: Likewise.
3626 * loop-unroll.c: Likewise.
3627 * lower-subreg.c: Likewise.
3628 * modulo-sched.c: Likewise.
3629 * postreload-gcse.c: Likewise.
3630 * predict.c: Likewise.
3631 * recog.c: Likewise.
3632 * sched-rgn.c: Likewise.
3633 * store-motion.c: Likewise.
3634 * tracer.c: Likewise.
3635 * trans-mem.c: Likewise.
3636 * tree-call-cdce.c: Likewise.
3637 * tree-cfg.c: Likewise.
3638 * tree-cfgcleanup.c: Likewise.
3639 * tree-complex.c: Likewise.
3640 * tree-eh.c: Likewise.
3641 * tree-emutls.c: Likewise.
3642 * tree-if-conv.c: Likewise.
3643 * tree-into-ssa.c: Likewise.
3644 * tree-loop-distribution.c: Likewise.
3645 * tree-object-size.c: Likewise.
3646 * tree-parloops.c: Likewise.
3647 * tree-pass.h: Likewise.
3648 * tree-sra.c: Likewise.
3649 * tree-ssa-ccp.c: Likewise.
3650 * tree-ssa-copy.c: Likewise.
3651 * tree-ssa-copyrename.c: Likewise.
3652 * tree-ssa-dce.c: Likewise.
3653 * tree-ssa-dom.c: Likewise.
3654 * tree-ssa-dse.c: Likewise.
3655 * tree-ssa-forwprop.c: Likewise.
3656 * tree-ssa-ifcombine.c: Likewise.
3657 * tree-ssa-loop-ch.c: Likewise.
3658 * tree-ssa-loop-ivcanon.c: Likewise.
3659 * tree-ssa-loop.c: Likewise.
3660 * tree-ssa-math-opts.c: Likewise.
3661 * tree-ssa-phiopt.c: Likewise.
3662 * tree-ssa-phiprop.c: Likewise.
3663 * tree-ssa-pre.c: Likewise.
3664 * tree-ssa-reassoc.c: Likewise.
3665 * tree-ssa-sink.c: Likewise.
3666 * tree-ssa-strlen.c: Likewise.
3667 * tree-ssa-tail-merge.c: Likewise.
3668 * tree-ssa-uncprop.c: Likewise.
3669 * tree-switch-conversion.c: Likewise.
3670 * tree-tailcall.c: Likewise.
3671 * tree-vect-generic.c: Likewise.
3672 * tree-vectorizer.c: Likewise.
3673 * tree-vrp.c: Likewise.
3675 * var-tracking.c: Likewise.
3676 * bt-load.c: Likewise.
3677 * cfgcleanup.c: Likewise.
3678 * combine-stack-adj.c: Likewise.
3679 * combine.c: Likewise.
3680 * compare-elim.c: Likewise.
3681 * config/epiphany/resolve-sw-modes.c: Likewise.
3682 * config/i386/i386.c: Likewise.
3683 * config/mips/mips.c: Likewise.
3684 * config/s390/s390.c: Likewise.
3685 * config/sh/sh_treg_combine.cc: Likewise.
3686 * config/sparc/sparc.c: Likewise.
3689 * final.c: Likewise.
3690 * ifcvt.c: Likewise.
3691 * mode-switching.c: Likewise.
3692 * passes.c: Likewise.
3693 * postreload.c: Likewise.
3695 * reg-stack.c: Likewise.
3696 * regcprop.c: Likewise.
3697 * regrename.c: Likewise.
3700 2014-05-06 Richard Biener <rguenther@suse.de>
3703 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3704 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3706 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3709 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3711 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3712 Tom de Vries <tom@codesourcery.com>
3714 * target.def (call_fusage_contains_non_callee_clobbers): New
3716 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3718 (@node Miscellaneous Register Hooks): New node.
3719 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3720 * doc/tm.texi: Regenerate.
3722 2014-05-05 Marek Polacek <polacek@redhat.com>
3725 * opts.c (common_handle_option): Call error_at instead of warning_at.
3727 2014-05-05 Richard Biener <rguenther@suse.de>
3729 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3730 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3731 under the TODO_verify_il umbrella.
3733 2014-05-05 Richard Biener <rguenther@suse.de>
3735 * passes.c (execute_function_todo): Move TODO_verify_flow under
3736 the TODO_verify_ul umbrella.
3738 2014-05-05 Richard Biener <rguenther@suse.de>
3741 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3742 X & CST away from a CST that is the mask of a mode.
3744 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3746 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3747 int argument to enum machine_mode.
3748 (picochip_class_max_nregs): Ditto.
3749 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3750 (picochip_class_max_nregs): Ditto.
3752 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3754 * target.def: Add new target hook.
3755 * doc/tm.texi: Regenerate.
3756 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3757 * targhooks.c (default_keep_leaf_when_profiled): New function.
3759 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3760 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3762 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3764 PR tree-optimization/60363
3765 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3766 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3767 (update_destination_phis): New parameter.
3768 (create_edge_and_update_destination_phis): Ditto.
3769 (ssa_fix_duplicate_block_edges): Pass new arguments.
3770 (thread_single_edge): Ditto.
3772 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3774 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3775 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3776 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3777 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3778 Use RS6000_BTM_HARD_FLOAT.
3779 (BU_MISC_2): Likewise.
3780 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3781 RS6000_BTM_HARD_FLOAT.
3782 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3784 (rs6000_invalid_builtin): Add hard floating builtin support.
3785 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3786 hard float builtins.
3787 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3789 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3791 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3792 Add missing function* argument.
3794 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3796 * lra-constraints.c (valid_address_p): Move earlier in file.
3797 Add a constraint argument to the address_info version.
3798 (satisfies_memory_constraint_p): New function.
3799 (satisfies_address_constraint_p): Likewise.
3800 (process_alt_operands, curr_insn_transform): Use them.
3801 (process_address): Pass the constraint to valid_address_p when
3802 checking address operands.
3804 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3806 * config/mips/mips.c (mips_isa_rev): New variable.
3807 (mips_set_architecture): Set it.
3808 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3810 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3811 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3812 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3813 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3814 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3815 conditions in terms of mips_isa_rev.
3816 (mips_isa_rev): Declare.
3818 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3820 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3821 (prob_unlikely, prob_likely): Make variables const.
3823 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3825 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3827 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3829 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3831 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3833 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3834 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3835 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3837 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3838 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3840 Replace usage of ROUND_REG with sh_round_reg.
3841 Use CEIL instead of ROUND_ADVANCE.
3843 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3846 * config/sh/sh.c: Include stdlib headers before everything else.
3848 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3850 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3851 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3852 (gimplify_adjust_omp_clauses): Simd region is never
3853 directly nested in combined parallel. Instead, for linear
3854 with copyin/copyout, if in combined for simd loop, make decl
3855 firstprivate/lastprivate on OMP_FOR.
3856 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3857 expand_omp_for_static_chunk): When setting endvar, also set
3858 fd->loop.v to the same value.
3860 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3862 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3864 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3866 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3869 2014-05-02 Marek Polacek <polacek@redhat.com>
3871 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3873 2014-05-02 Kito Cheng <kito@0xlab.org>
3875 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3876 to a C expression marco.
3877 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3878 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3879 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3880 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3881 HONOR_REG_ALLOC_ORDER.
3882 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3884 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3886 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3888 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3890 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3892 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3894 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3895 (convert_scalar_cond_reduction): Likewise.
3896 (predicate_scalar_phi): Add recognition and transformation
3897 of simple conditioanl reduction to be vectorizable.
3899 2014-05-01 Marek Polacek <polacek@redhat.com>
3902 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3904 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3906 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3907 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3908 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3909 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3910 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3911 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3912 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3913 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3915 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3917 * config/arc/arc.opt (mlra): Move comment above option name
3918 to avoid mis-parsing as language options.
3920 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3922 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3923 * config/sol2.h: ... here.
3924 * config/sol2-10.h: Remove.
3926 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3927 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3928 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3929 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3930 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3931 * config/sol2.h: ... here.
3932 (SECTION_NAME_FORMAT): Don't redefine.
3933 (STARTFILE_ARCH32_SPEC): Rename to ...
3934 (STARTFILE_ARCH_SPEC): ... this.
3935 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3936 * config/sparc/sol2.h: ... here.
3937 (SECTION_NAME_FORMAT): Don't undef.
3938 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3939 (SUBTARGET_EXTRA_SPECS): Remove.
3940 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3942 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3943 (MD_STARTFILE_PREFIX): Remove.
3944 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3945 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3946 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3947 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3948 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3949 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3950 * config/i386/sol2.h: ... here.
3951 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3952 * config/i386/sol2-bi.h: Remove.
3953 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3954 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3956 * config/i386/t-sol2-64: Rename to ...
3957 * config/i386/t-sol2: ... this.
3958 * config/sparc/t-sol2-64: Rename to ...
3959 * config/sparc/t-sol2: ... this.
3961 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3962 sol2_tm_file_head, sol2_tm_file_tail.
3963 Include ${cpu_type}/sol2.h before sol2.h.
3965 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3966 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3967 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3968 Reflect i386/t-sol2-64 renaming.
3969 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3970 Reflect sparc/t-sol2-64 renaming.
3972 2014-04-30 Richard Biener <rguenther@suse.de>
3974 * passes.c (execute_function_todo): Move TODO_verify_stmts
3975 and TODO_verify_ssa under the TODO_verify_il umbrella.
3976 * tree-ssa.h (verify_ssa): Adjust prototype.
3977 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3978 we should verify SSA operands.
3979 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3980 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3981 whether we should verify whether not throwing stmts have EH info.
3982 * graphite-scop-detection.c (create_sese_edges): Adjust.
3983 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3984 * tree-eh.c (lower_try_finally_switch): Do not add the
3985 default case label twice.
3987 2014-04-30 Marek Polacek <polacek@redhat.com>
3989 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3990 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3991 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3992 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3994 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3996 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3997 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3998 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3999 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4000 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4001 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4002 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4003 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4005 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4007 * tree-cfg.c (dump_function_to_file): Dump the return type of
4008 functions, in a line to itself before the function body, mimicking
4009 the layout of a C function.
4011 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4013 PR tree-optimization/60971
4014 * tree-tailcall.c (process_assignment): Reject conversions which
4017 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4019 * calls.c (initialize_argument_information): Always treat
4020 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4021 (expand_call): Likewise.
4022 (emit_library_call_calue_1): Likewise.
4023 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4024 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4027 2014-04-29 Nick Clifton <nickc@redhat.com>
4029 * config/msp430/msp430.md (umulsidi): Fix typo.
4030 (mulhisi3): Enable even inside interrupt handlers.
4031 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4032 bigger return address pushed in large mode.
4034 2014-04-29 Nick Clifton <nickc@redhat.com>
4036 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4037 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4039 * config/m32r/m32r.c (init_reg_tables): Likewise.
4040 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4041 enum to hold the modes.
4043 2014-04-29 Richard Biener <rguenther@suse.de>
4045 * dominance.c (free_dominance_info): Add overload with
4047 (dom_info_state): Likewise.
4048 (dom_info_available_p): Likewise.
4049 * basic-block.h (free_dominance_info, dom_info_state,
4050 dom_info_available_p): Declare overloads.
4051 * passes.c (execute_function_todo): Verify that verifiers
4052 don't change dominator info state. Drop dominator info
4053 for IPA pass invocations.
4054 * cgraph.c (release_function_body): Restore asserts that
4055 dominator information is released.
4057 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4059 * doc/invoke.texi: Fix typo.
4060 * tree-vrp.c: Fix typos.
4061 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4063 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4065 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4067 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4069 * config/aarch64/aarch64-builtins.c
4070 (aarch64_types_storestruct_lane_qualifiers): New.
4071 (TYPES_STORESTRUCT_LANE): Likewise.
4072 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4073 (st3_lane): Likewise.
4074 (st4_lane): Likewise.
4075 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4076 (vec_store_lanesci_lane<mode>): Likewise.
4077 (vec_store_lanesxi_lane<mode>): Likewise.
4078 (aarch64_st2_lane<VQ:mode>): Likewise.
4079 (aarch64_st3_lane<VQ:mode>): Likewise.
4080 (aarch64_st4_lane<VQ:mode>): Likewise.
4081 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4082 * config/aarch64/arm_neon.h
4083 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4084 use new macro arguments.
4085 (__ST3_LANE_FUNC): Likewise.
4086 (__ST4_LANE_FUNC): Likewise.
4087 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4088 (V_THREE_ELEM): Likewise.
4089 (V_FOUR_ELEM): Likewise.
4091 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4093 * doc/gimple.texi: Replace the description of the now-defunct
4094 union gimple_statement_d with a diagram showing the
4095 gimple_statement_base class hierarchy and its relationships to
4096 the GSS_ and GIMPLE_ enums.
4098 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4100 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4101 * config/aarch64/aarch64.c
4102 (aarch64_cannot_change_mode_class): Weaken conditions.
4103 (aarch64_modes_tieable_p): New.
4104 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4106 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4108 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4109 (loadsync_<mode>): Change mode.
4110 (load_quadpti, store_quadpti): New.
4111 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4112 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4114 2014-04-28 Martin Jambor <mjambor@suse.cz>
4116 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4117 same alias type as the original statement.
4118 (subreplacement_assignment_data): New type.
4119 (handle_unscalarized_data_in_subtree): New type of parameter,
4120 generate new memory accesses with same alias type as the original
4122 (load_assign_lhs_subreplacements): Likewise.
4123 (sra_modify_constructor_assign): Generate new memory accesses with
4124 same alias type as the original statement.
4126 2014-04-28 Richard Biener <rguenther@suse.de>
4128 * tree-pass.h (TODO_verify_il): Define.
4129 (TODO_verify_all): Complete properly.
4130 * passes.c (execute_function_todo): Move existing loop-closed
4131 SSA verification under TODO_verify_il.
4132 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4133 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4134 Fix tree sharing issue.
4136 2014-04-28 Richard Biener <rguenther@suse.de>
4139 * builtins.def (DEF_C11_BUILTIN): Add.
4140 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4141 * coretypes.h (enum function_class): Add function_c11_misc.
4142 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4143 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4144 (call_may_clobber_ref_p_1): Likewise.
4145 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4146 (mark_all_reaching_defs_necessary_1): Likewise.
4147 (propagate_necessity): Likewise.
4148 (eliminate_unnecessary_stmts): Likewise.
4149 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4151 2014-04-28 Richard Biener <rguenther@suse.de>
4153 * tree-vrp.c (vrp_var_may_overflow): Remove.
4154 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4155 with overflow immediately bump to one before that value and
4156 let iteration figure out overflow status.
4158 2014-04-28 Richard Biener <rguenther@suse.de>
4160 * configure.ac: Do valgrind header checks unconditionally.
4161 Add --enable-valgrind-annotations.
4162 * system.h: Guard valgrind header inclusion with
4163 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4164 * alloc-pool.c (pool_alloc, pool_free): Use
4165 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4166 to guard possibly dead code.
4167 * config.in: Regenerated.
4168 * configure: Likewise.
4170 2014-04-28 Jeff Law <law@redhat.com>
4172 PR tree-optimization/60902
4173 * tree-ssa-threadedge.c
4174 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4175 over real defs when invalidating outputs from statements that do not
4176 produce useful outputs for threading.
4178 2014-04-28 Richard Biener <rguenther@suse.de>
4180 PR tree-optimization/60979
4181 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4182 SCOPs that end in a block with a successor with abnormal
4185 2014-04-28 Richard Biener <rguenther@suse.de>
4187 * tree-pass.h (execute_pass_list): Adjust prototype.
4188 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4189 (do_per_function): Change callback signature, push all actual
4190 work to the callbals.
4191 (do_per_function_toporder): Likewise.
4192 (execute_function_dump): Adjust.
4193 (execute_function_todo): Likewise.
4194 (clear_last_verified): Likewise.
4195 (verify_curr_properties): Likewise.
4196 (update_properties_after_pass): Likewise.
4197 (execute_pass_list_1): Split out from ...
4198 (execute_pass_list): ... here. Adjust.
4199 (execute_ipa_pass_list): Likewise.
4200 * cgraphunit.c (cgraph_add_new_function): Adjust.
4201 (analyze_function): Likewise.
4202 (expand_function): Likewise.
4203 * cgraph.c (release_function_body): Free dominance info
4204 here instead of asserting it was magically freed elsewhere.
4206 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4208 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4209 * configure: Regenerate.
4210 * config/sparc/sparc.opt (muser-mode): New option.
4211 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4213 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4214 * doc/invoke.texi (SPARC options): Document -muser-mode.
4216 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4218 * cselib.c (find_slot_memmode): Delete.
4219 (cselib_hasher): Change compare_type to a struct.
4220 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4222 (preserve_constants_and_equivs): Adjust for new compare_type.
4223 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4224 (wrap_constant): Delete.
4225 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4227 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4229 * doc/install.texi (Building with profile feedback): Remove
4232 2014-04-26 Tom de Vries <tom@codesourcery.com>
4234 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4237 2014-04-25 Cary Coutant <ccoutant@google.com>
4240 * dwarf2out.c (should_move_die_to_comdat): A type definition
4241 can contain a subprogram definition, but don't move it to a
4243 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4244 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4246 (clone_tree_hash): Rename to...
4247 (clone_tree_partial): ...this; change callers. Copy
4248 DW_TAG_subprogram DIEs as declarations.
4249 (copy_decls_walk): Don't copy children of a declaration into a
4252 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4255 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4258 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4260 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4261 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4263 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4266 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4268 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4270 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4272 PR tree-optimization/60930
4273 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4274 creating a multiply candidate by folding two constant
4275 multiplicands when the result overflows.
4277 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4279 PR tree-optimization/60960
4280 * tree-vect-generic.c (expand_vector_operation): Only call
4281 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4283 2014-04-25 Tom de Vries <tom@codesourcery.com>
4285 * expr.c (clobber_reg_mode): New function.
4286 * expr.h (clobber_reg): New function.
4288 2014-04-25 Tom de Vries <tom@codesourcery.com>
4290 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4293 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4294 Tom de Vries <tom@codesourcery.com>
4296 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4298 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4299 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4300 new argument to find_all_hard_reg_sets call.
4302 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4304 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4305 Use HOST_WIDE_INT_C for mask literal.
4306 (aarch_rev16_shleft_mask_imm_p): Likewise.
4308 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4311 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4313 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4315 PR preprocessor/56540
4316 * config/i386/i386-c.c (ix86_target_macros): Define
4317 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4319 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4321 * configure.ac (tga_func): Remove.
4322 (LIB_TLS_SPEC): Remove.
4323 * configure: Regenerate.
4324 * config.in: Regenerate.
4325 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4327 2014-04-25 Richard Biener <rguenther@suse.de>
4330 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4331 call stmt use/clobber sets during stmt walk instead of
4332 walking the possibly incomplete set of caller edges.
4334 2014-04-25 Richard Biener <rguenther@suse.de>
4337 * passes.c (apply_ipa_transforms): Inline into only caller ...
4338 (execute_one_pass): ... here. Properly bring in function
4339 bodies for nodes we want to apply IPA transforms to.
4341 2014-04-24 Cong Hou <congh@google.com>
4343 PR tree-optimization/60896
4344 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4345 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4346 (vect_mark_pattern_stmts): Set the def type of all statements in
4347 PATTERN_DEF_SEQ as vect_internal_def.
4349 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4351 * doc/extend.texi (PowerPC Built-in Functions): Document new
4352 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4353 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4355 * config/rs6000/predicates.md (const_0_to_3_operand): New
4356 predicate to match 0..3 integer constants.
4358 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4359 to support adding miscellaneous builtin functions.
4360 (BU_DFP_MISC_2): Likewise.
4361 (BU_P7_MISC_1): Likewise.
4362 (BU_P7_MISC_2): Likewise.
4363 (BU_P8V_MISC_3): Likewise.
4364 (BU_MISC_1): Likewise.
4365 (BU_MISC_2): Likewise.
4366 (DIVWE): Add extended divide builtin functions.
4369 (DIVWEUO): Likewise.
4373 (DIVDEUO): Likewise.
4374 (DXEX): Add decimal floating-point builtin functions.
4377 (DDEDPDQ): Likewise.
4379 (DENBCDQ): Likewise.
4386 (CDTBCD): Add new BCD builtin functions.
4390 (BCDADD_LT): Likewise.
4391 (BCDADD_EQ): Likewise.
4392 (BCDADD_GT): Likewise.
4393 (BCDADD_OV): Likewise.
4395 (BCDSUB_LT): Likewise.
4396 (BCDSUB_EQ): Likewise.
4397 (BCDSUB_GT): Likewise.
4398 (BCDSUB_OV): Likewise.
4399 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4400 (UNPACK_TD): Likewise.
4401 (PACK_TF): Likewise.
4402 (UNPACK_TF): Likewise.
4403 (UNPACK_TF_0): Likewise.
4404 (UNPACK_TF_1): Likewise.
4405 (PACK_V1TI): Likewise.
4406 (UNPACK_V1TI): Likewise.
4408 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4409 support for decimal floating point builtin functions.
4410 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4411 functions that take constant arguments.
4412 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4413 (rs6000_init_builtins): Setup long double, _Decimal64, and
4414 _Decimal128 types for new builtin functions.
4415 (builtin_function_type): Set the unsigned flags appropriately for
4416 the new builtin functions.
4417 (rs6000_opt_masks): Add support for decimal floating point builtin
4420 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4421 floating point builtin functions.
4422 (RS6000_BTM_COMMON): Likewise.
4423 (RS6000_BTI_long_double): Likewise.
4424 (RS6000_BTI_dfloat64): Likewise.
4425 (RS6000_BTI_dfloat128): Likewise.
4426 (long_double_type_internal_node): Likewise.
4427 (dfloat64_type_internal_node): Likewise.
4428 (dfloat128_type_internal_node): Likewise.
4430 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4431 2.07 bcd arithmetic instructions.
4432 (UNSPEC_BCDSUB): Likewise.
4433 (UNSPEC_BCD_OVERFLOW): Likewise.
4434 (UNSPEC_BCD_ADD_SUB): Likewise.
4435 (bcd_add_sub): Likewise.
4436 (BCD_TEST): Likewise.
4437 (bcd<bcd_add_sub>): Likewise.
4438 (bcd<bcd_add_sub>_test): Likewise.
4439 (bcd<bcd_add_sub>_test2): Likewise.
4440 (bcd<bcd_add_sub>_<code>): Likewise.
4441 (peephole2 for combined bcd ops): Likewise.
4443 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4444 decimal floating point builtin functions.
4445 (UNSPEC_DENBCD): Likewise.
4446 (UNSPEC_DXEX): Likewise.
4447 (UNSPEC_DIEX): Likewise.
4448 (UNSPEC_DSCLI): Likewise.
4449 (UNSPEC_DSCRI): Likewise.
4450 (D64_D128): Likewise.
4451 (dfp_suffix): Likewise.
4452 (dfp_ddedpd_<mode>): Likewise.
4453 (dfp_denbcd_<mode>): Likewise.
4454 (dfp_dxex_<mode>): Likewise.
4455 (dfp_diex_<mode>): Likewise.
4456 (dfp_dscli_<mode>): Likewise.
4457 (dfp_dscri_<mode>): Likewise.
4459 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4461 (UNSPEC_CDTBCD): Likewise.
4462 (UNSPEC_CBCDTD): Likewise.
4463 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4464 (UNSPEC_DIVEO): Likewise.
4465 (UNSPEC_DIVEU): Likewise.
4466 (UNSPEC_DIVEUO): Likewise.
4467 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4468 pack/unpack 128-bit types.
4469 (UNSPEC_PACK_128BIT): Likewise.
4470 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4471 (udiv<mode>3): Use idiv_ldiv mode attribute.
4472 (div<mode>3): Likewise.
4473 (addg6s): Add new BCD builtin functions.
4476 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4477 (div_extend): Likewise.
4478 (div<div_extend>_<mode>"): Likewise.
4479 (FP128_64): Add support for new builtin functions to pack/unpack
4481 (unpack<mode>): Likewise.
4482 (unpacktf_0): Likewise.
4483 (unpacktf_1): Likewise.
4484 (unpack<mode>_dm): Likewise.
4485 (unpack<mode>_nodm): Likewise.
4486 (pack<mode>): Likewise.
4487 (unpackv1ti): Likewise.
4488 (packv1ti): Likewise.
4490 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4492 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4495 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4497 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4498 * gimplify.c (omp_is_private): Change last argument's type to int.
4499 Only diagnose lastprivate if the simd argument is 1, only diagnose
4500 linear if the simd argument is 2.
4501 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4502 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4503 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4504 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4505 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4506 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4507 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4508 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4509 * tree-nested.c (convert_nonlocal_omp_clauses,
4510 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4512 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4515 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4518 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4520 * flag-types.h (enum ivar_visibility): Add.
4522 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4524 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4525 function * argument.
4527 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4529 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4531 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4532 Tom de Vries <tom@codesourcery.com>
4534 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4535 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4537 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4538 * emit-rtl.c (try_split): Same.
4540 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4541 Tom de Vries <tom@codesourcery.com>
4543 * common.opt (fuse-caller-save): New option.
4545 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4547 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4548 elements for big-endian.
4550 2014-04-24 Richard Biener <rguenther@suse.de>
4552 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4553 during TER and instead use the sepops interface for expanding
4554 non-GIMPLE_SINGLE_RHS.
4556 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4558 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4559 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4561 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4563 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4564 assembler 64-bit option.
4565 * configure: Regenerate.
4567 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4569 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4570 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4571 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4572 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4573 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4575 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4577 * config/aarch64/aarch64-builtins.c
4578 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4579 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4580 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4581 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4583 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4584 (Vrevsuff): New mode attribute.
4586 2014-04-24 Terry Guo <terry.guo@arm.com>
4588 * config/arm/arm.h (machine_function): Define variable
4589 after_arm_reorg here.
4590 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4591 (arm_split_constant): Update the way to access variable
4594 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4596 2014-04-23 Tom de Vries <tom@codesourcery.com>
4598 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4600 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4602 * is-a.h: Update comments to reflect the following changes to the
4603 "pointerness" of the API, making the template parameter match the
4604 return type, allowing use of is-a.h with typedefs of pointers.
4605 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4606 that the return type matches the parameter to the is_a_helper.
4608 (dyn_cast): Likewise.
4610 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4611 pointer from the is-a.h API.
4613 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4614 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4616 (is_a_helper <varpool_node>::test): Likewise, convert to...
4617 (is_a_helper <varpool_node *>::test): ...this.
4619 (varpool_first_variable): Update for removal of implicit pointer
4620 from the is-a.h API.
4621 (varpool_next_variable): Likewise.
4622 (varpool_first_static_initializer): Likewise.
4623 (varpool_next_static_initializer): Likewise.
4624 (varpool_first_defined_variable): Likewise.
4625 (varpool_next_defined_variable): Likewise.
4626 (cgraph_first_defined_function): Likewise.
4627 (cgraph_next_defined_function): Likewise.
4628 (cgraph_first_function): Likewise.
4629 (cgraph_next_function): Likewise.
4630 (cgraph_first_function_with_gimple_body): Likewise.
4631 (cgraph_next_function_with_gimple_body): Likewise.
4632 (cgraph_alias_target): Likewise.
4633 (varpool_alias_target): Likewise.
4634 (cgraph_function_or_thunk_node): Likewise.
4635 (varpool_variable_node): Likewise.
4636 (symtab_real_symbol_p): Likewise.
4637 * cgraphunit.c (referred_to_p): Likewise.
4638 (analyze_functions): Likewise.
4639 (handle_alias_pairs): Likewise.
4640 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4641 * gimple-ssa.h (gimple_vuse_op): Likewise.
4642 (gimple_vdef_op): Likewise.
4643 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4644 * gimple.c (gimple_build_asm_1): Likewise.
4645 (gimple_build_try): Likewise.
4646 (gimple_build_resx): Likewise.
4647 (gimple_build_eh_dispatch): Likewise.
4648 (gimple_build_omp_for): Likewise.
4649 (gimple_omp_for_set_clauses): Likewise.
4651 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4652 (is_a_helper <gimple_statement_asm *>::test): ...this.
4653 (is_a_helper <gimple_statement_bind>::test): Convert to...
4654 (is_a_helper <gimple_statement_bind *>::test): ...this.
4655 (is_a_helper <gimple_statement_call>::test): Convert to...
4656 (is_a_helper <gimple_statement_call *>::test): ...this.
4657 (is_a_helper <gimple_statement_catch>::test): Convert to...
4658 (is_a_helper <gimple_statement_catch *>::test): ...this.
4659 (is_a_helper <gimple_statement_resx>::test): Convert to...
4660 (is_a_helper <gimple_statement_resx *>::test): ...this.
4661 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4662 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4663 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4664 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4665 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4666 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4667 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4668 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4669 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4670 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4671 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4672 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4673 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4674 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4675 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4676 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4677 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4678 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4679 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4680 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4681 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4682 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4683 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4684 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4685 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4686 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4687 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4688 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4689 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4690 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4691 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4692 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4693 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4694 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4695 (is_a_helper <gimple_statement_phi>::test): Convert to...
4696 (is_a_helper <gimple_statement_phi *>::test): ...this.
4697 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4698 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4699 (is_a_helper <gimple_statement_try>::test): Convert to...
4700 (is_a_helper <gimple_statement_try *>::test): ...this.
4701 (is_a_helper <gimple_statement_wce>::test): Convert to...
4702 (is_a_helper <gimple_statement_wce *>::test): ...this.
4703 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4704 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4705 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4706 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4707 (is_a_helper <const gimple_statement_call>::test): Convert to...
4708 (is_a_helper <const gimple_statement_call *>::test): ...this.
4709 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4710 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4711 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4712 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4713 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4714 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4715 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4716 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4717 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4719 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4721 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4723 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4725 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4726 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4727 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4729 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4730 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4732 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4733 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4734 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4735 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4736 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4737 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4739 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4740 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4741 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4742 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4744 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4745 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4746 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4747 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4748 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4749 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4750 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4751 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4752 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4753 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4754 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4755 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4756 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4757 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4758 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4759 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4761 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4763 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4764 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4766 (gimple_use_ops): Update for removal of implicit pointer from the
4768 (gimple_set_use_ops): Likewise.
4769 (gimple_vuse): Likewise.
4770 (gimple_vdef): Likewise.
4771 (gimple_vuse_ptr): Likewise.
4772 (gimple_vdef_ptr): Likewise.
4773 (gimple_set_vuse): Likewise.
4774 (gimple_set_vdef): Likewise.
4775 (gimple_omp_return_set_lhs): Likewise.
4776 (gimple_omp_return_lhs): Likewise.
4777 (gimple_omp_return_lhs_ptr): Likewise.
4778 (gimple_call_fntype): Likewise.
4779 (gimple_call_set_fntype): Likewise.
4780 (gimple_call_set_internal_fn): Likewise.
4781 (gimple_call_use_set): Likewise.
4782 (gimple_call_clobber_set): Likewise.
4783 (gimple_bind_vars): Likewise.
4784 (gimple_bind_set_vars): Likewise.
4785 (gimple_bind_body_ptr): Likewise.
4786 (gimple_bind_set_body): Likewise.
4787 (gimple_bind_add_stmt): Likewise.
4788 (gimple_bind_block): Likewise.
4789 (gimple_bind_set_block): Likewise.
4790 (gimple_asm_ninputs): Likewise.
4791 (gimple_asm_noutputs): Likewise.
4792 (gimple_asm_nclobbers): Likewise.
4793 (gimple_asm_nlabels): Likewise.
4794 (gimple_asm_input_op): Likewise.
4795 (gimple_asm_input_op_ptr): Likewise.
4796 (gimple_asm_output_op): Likewise.
4797 (gimple_asm_output_op_ptr): Likewise.
4798 (gimple_asm_set_output_op): Likewise.
4799 (gimple_asm_clobber_op): Likewise.
4800 (gimple_asm_set_clobber_op): Likewise.
4801 (gimple_asm_label_op): Likewise.
4802 (gimple_asm_set_label_op): Likewise.
4803 (gimple_asm_string): Likewise.
4804 (gimple_catch_types): Likewise.
4805 (gimple_catch_types_ptr): Likewise.
4806 (gimple_catch_handler_ptr): Likewise.
4807 (gimple_catch_set_types): Likewise.
4808 (gimple_catch_set_handler): Likewise.
4809 (gimple_eh_filter_types): Likewise.
4810 (gimple_eh_filter_types_ptr): Likewise.
4811 (gimple_eh_filter_failure_ptr): Likewise.
4812 (gimple_eh_filter_set_types): Likewise.
4813 (gimple_eh_filter_set_failure): Likewise.
4814 (gimple_eh_must_not_throw_fndecl): Likewise.
4815 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4816 (gimple_eh_else_n_body_ptr): Likewise.
4817 (gimple_eh_else_e_body_ptr): Likewise.
4818 (gimple_eh_else_set_n_body): Likewise.
4819 (gimple_eh_else_set_e_body): Likewise.
4820 (gimple_try_eval_ptr): Likewise.
4821 (gimple_try_cleanup_ptr): Likewise.
4822 (gimple_try_set_eval): Likewise.
4823 (gimple_try_set_cleanup): Likewise.
4824 (gimple_wce_cleanup_ptr): Likewise.
4825 (gimple_wce_set_cleanup): Likewise.
4826 (gimple_phi_capacity): Likewise.
4827 (gimple_phi_num_args): Likewise.
4828 (gimple_phi_result): Likewise.
4829 (gimple_phi_result_ptr): Likewise.
4830 (gimple_phi_set_result): Likewise.
4831 (gimple_phi_arg): Likewise.
4832 (gimple_phi_set_arg): Likewise.
4833 (gimple_resx_region): Likewise.
4834 (gimple_resx_set_region): Likewise.
4835 (gimple_eh_dispatch_region): Likewise.
4836 (gimple_eh_dispatch_set_region): Likewise.
4837 (gimple_omp_critical_name): Likewise.
4838 (gimple_omp_critical_name_ptr): Likewise.
4839 (gimple_omp_critical_set_name): Likewise.
4840 (gimple_omp_for_clauses): Likewise.
4841 (gimple_omp_for_clauses_ptr): Likewise.
4842 (gimple_omp_for_set_clauses): Likewise.
4843 (gimple_omp_for_collapse): Likewise.
4844 (gimple_omp_for_index): Likewise.
4845 (gimple_omp_for_index_ptr): Likewise.
4846 (gimple_omp_for_set_index): Likewise.
4847 (gimple_omp_for_initial): Likewise.
4848 (gimple_omp_for_initial_ptr): Likewise.
4849 (gimple_omp_for_set_initial): Likewise.
4850 (gimple_omp_for_final): Likewise.
4851 (gimple_omp_for_final_ptr): Likewise.
4852 (gimple_omp_for_set_final): Likewise.
4853 (gimple_omp_for_incr): Likewise.
4854 (gimple_omp_for_incr_ptr): Likewise.
4855 (gimple_omp_for_set_incr): Likewise.
4856 (gimple_omp_for_pre_body_ptr): Likewise.
4857 (gimple_omp_for_set_pre_body): Likewise.
4858 (gimple_omp_parallel_clauses): Likewise.
4859 (gimple_omp_parallel_clauses_ptr): Likewise.
4860 (gimple_omp_parallel_set_clauses): Likewise.
4861 (gimple_omp_parallel_child_fn): Likewise.
4862 (gimple_omp_parallel_child_fn_ptr): Likewise.
4863 (gimple_omp_parallel_set_child_fn): Likewise.
4864 (gimple_omp_parallel_data_arg): Likewise.
4865 (gimple_omp_parallel_data_arg_ptr): Likewise.
4866 (gimple_omp_parallel_set_data_arg): Likewise.
4867 (gimple_omp_task_clauses): Likewise.
4868 (gimple_omp_task_clauses_ptr): Likewise.
4869 (gimple_omp_task_set_clauses): Likewise.
4870 (gimple_omp_task_child_fn): Likewise.
4871 (gimple_omp_task_child_fn_ptr): Likewise.
4872 (gimple_omp_task_set_child_fn): Likewise.
4873 (gimple_omp_task_data_arg): Likewise.
4874 (gimple_omp_task_data_arg_ptr): Likewise.
4875 (gimple_omp_task_set_data_arg): Likewise.
4876 (gimple_omp_taskreg_clauses): Likewise.
4877 (gimple_omp_taskreg_clauses_ptr): Likewise.
4878 (gimple_omp_taskreg_set_clauses): Likewise.
4879 (gimple_omp_taskreg_child_fn): Likewise.
4880 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4881 (gimple_omp_taskreg_set_child_fn): Likewise.
4882 (gimple_omp_taskreg_data_arg): Likewise.
4883 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4884 (gimple_omp_taskreg_set_data_arg): Likewise.
4885 (gimple_omp_task_copy_fn): Likewise.
4886 (gimple_omp_task_copy_fn_ptr): Likewise.
4887 (gimple_omp_task_set_copy_fn): Likewise.
4888 (gimple_omp_task_arg_size): Likewise.
4889 (gimple_omp_task_arg_size_ptr): Likewise.
4890 (gimple_omp_task_set_arg_size): Likewise.
4891 (gimple_omp_task_arg_align): Likewise.
4892 (gimple_omp_task_arg_align_ptr): Likewise.
4893 (gimple_omp_task_set_arg_align): Likewise.
4894 (gimple_omp_single_clauses): Likewise.
4895 (gimple_omp_single_clauses_ptr): Likewise.
4896 (gimple_omp_single_set_clauses): Likewise.
4897 (gimple_omp_target_clauses): Likewise.
4898 (gimple_omp_target_clauses_ptr): Likewise.
4899 (gimple_omp_target_set_clauses): Likewise.
4900 (gimple_omp_target_child_fn): Likewise.
4901 (gimple_omp_target_child_fn_ptr): Likewise.
4902 (gimple_omp_target_set_child_fn): Likewise.
4903 (gimple_omp_target_data_arg): Likewise.
4904 (gimple_omp_target_data_arg_ptr): Likewise.
4905 (gimple_omp_target_set_data_arg): Likewise.
4906 (gimple_omp_teams_clauses): Likewise.
4907 (gimple_omp_teams_clauses_ptr): Likewise.
4908 (gimple_omp_teams_set_clauses): Likewise.
4909 (gimple_omp_sections_clauses): Likewise.
4910 (gimple_omp_sections_clauses_ptr): Likewise.
4911 (gimple_omp_sections_set_clauses): Likewise.
4912 (gimple_omp_sections_control): Likewise.
4913 (gimple_omp_sections_control_ptr): Likewise.
4914 (gimple_omp_sections_set_control): Likewise.
4915 (gimple_omp_for_set_cond): Likewise.
4916 (gimple_omp_for_cond): Likewise.
4917 (gimple_omp_atomic_store_set_val): Likewise.
4918 (gimple_omp_atomic_store_val): Likewise.
4919 (gimple_omp_atomic_store_val_ptr): Likewise.
4920 (gimple_omp_atomic_load_set_lhs): Likewise.
4921 (gimple_omp_atomic_load_lhs): Likewise.
4922 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4923 (gimple_omp_atomic_load_set_rhs): Likewise.
4924 (gimple_omp_atomic_load_rhs): Likewise.
4925 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4926 (gimple_omp_continue_control_def): Likewise.
4927 (gimple_omp_continue_control_def_ptr): Likewise.
4928 (gimple_omp_continue_set_control_def): Likewise.
4929 (gimple_omp_continue_control_use): Likewise.
4930 (gimple_omp_continue_control_use_ptr): Likewise.
4931 (gimple_omp_continue_set_control_use): Likewise.
4932 (gimple_transaction_body_ptr): Likewise.
4933 (gimple_transaction_label): Likewise.
4934 (gimple_transaction_label_ptr): Likewise.
4935 (gimple_transaction_set_body): Likewise.
4936 (gimple_transaction_set_label): Likewise.
4938 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4939 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4940 * ipa-ref.c (ipa_record_reference): Likewise.
4941 * ipa-reference.c (analyze_function): Likewise.
4942 (ipa_reference_write_optimization_summary): Likewise.
4943 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4944 (address_taken_from_non_vtable_p): Likewise.
4945 (comdat_can_be_unshared_p_1): Likewise.
4946 * lto-cgraph.c (lto_output_ref): Likewise.
4947 (add_references): Likewise.
4948 (compute_ltrans_boundary): Likewise.
4949 (output_symtab): Likewise.
4950 (input_ref): Likewise.
4951 (input_cgraph_1): Likewise.
4952 (output_cgraph_opt_summary): Likewise.
4953 * lto-streamer-out.c (lto_output): Likewise.
4954 (output_symbol_p): Likewise.
4955 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4956 (lsei_start_function_in_partition): Likewise.
4957 (lsei_next_variable_in_partition): Likewise.
4958 (lsei_start_variable_in_partition): Likewise.
4959 * symtab.c (insert_to_assembler_name_hash): Likewise.
4960 (unlink_from_assembler_name_hash): Likewise.
4961 (symtab_unregister_node): Likewise.
4962 (symtab_remove_node): Likewise.
4963 (dump_symtab_node): Likewise.
4964 (verify_symtab_base): Likewise.
4965 (verify_symtab_node): Likewise.
4966 (symtab_make_decl_local): Likewise.
4967 (symtab_alias_ultimate_target): Likewise.
4968 (symtab_resolve_alias): Likewise.
4969 (symtab_get_symbol_partitioning_class): Likewise.
4970 * tree-phinodes.c (allocate_phi_node): Likewise.
4971 (reserve_phi_args_for_new_edge): Likewise.
4972 (remove_phi_args): Likewise.
4973 * varpool.c (varpool_node_for_asm): Likewise.
4974 (varpool_remove_unreferenced_decls): Likewise.
4976 2014-04-23 Jeff Law <law@redhat.com>
4978 PR tree-optimization/60902
4979 * tree-ssa-threadedge.c
4980 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4981 invalidate outputs from statements that do not produce useful
4982 outputs for threading.
4984 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4986 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4987 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4988 machine descriptions for Stack Smashing Protector.
4990 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4992 * aarch64.md (<optab>_rol<mode>3): New pattern.
4993 (<optab>_rolsi3_uxtw): Likewise.
4994 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4996 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4998 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4999 (arm_cortex_a12_tune): Likewise.
5001 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5003 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5005 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5007 * config/arm/arm.md (arm_rev16si2): New pattern.
5008 (arm_rev16si2_alt): Likewise.
5009 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5011 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5013 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5014 (rev16<mode>2_alt): Likewise.
5015 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5016 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5017 (aarch_rev16_shleft_mask_imm_p): Likewise.
5018 (aarch_rev16_p_1): Likewise.
5019 (aarch_rev16_p): Likewise.
5020 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5021 (aarch_rev16_shright_mask_imm_p): Likewise.
5022 (aarch_rev16_shleft_mask_imm_p): Likewise.
5024 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5026 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5027 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5029 (cortex_a53_extra_costs): Likewise.
5030 (cortex_a57_extra_costs): Likewise.
5031 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5032 (cortexa7_extra_costs): Likewise.
5033 (cortexa8_extra_costs): Likewise.
5034 (cortexa12_extra_costs): Likewise.
5035 (cortexa15_extra_costs): Likewise.
5036 (v7m_extra_costs): Likewise.
5037 (arm_new_rtx_costs): Handle BSWAP.
5039 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5041 * config/arm/arm.c (cortexa8_extra_costs): New table.
5042 (arm_cortex_a8_tune): New tuning struct.
5043 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5045 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5047 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5049 2014-04-23 Richard Biener <rguenther@suse.de>
5051 * Makefile.in (OBJS): Remove loop-unswitch.o.
5052 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5053 * passes.def (pass_rtl_unswitch): Likewise.
5054 * loop-init.c (gate_rtl_unswitch): Likewise.
5055 (rtl_unswitch): Likewise.
5056 (pass_data_rtl_unswitch): Likewise.
5057 (pass_rtl_unswitch): Likewise.
5058 (make_pass_rtl_unswitch): Likewise.
5059 * rtl.h (reversed_condition): Likewise.
5060 (compare_and_jump_seq): Likewise.
5061 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5063 * loop-unroll.c (compare_and_jump_seq): Likewise.
5065 2014-04-23 Richard Biener <rguenther@suse.de>
5067 PR tree-optimization/60903
5068 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5069 commented code block.
5070 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5071 loop flags to newly created BBs and edges.
5073 2014-04-23 Nick Clifton <nickc@redhat.com>
5075 * config/msp430/msp430.c (msp430_handle_option): Move function
5077 (msp430_option_override): Simplify mcu and mcpu option handling.
5078 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5079 support for -mhwmult command line option.
5080 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5081 -mhwmult command line option.
5082 (msp430_hwmult_enabled): Delete.
5083 (msp43o_output_labelref): Add support for -mhwmult command line option.
5084 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5085 (umulsidi3): Likewise.
5086 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5087 (mcpu, mlarge, msmall): Likewise.
5088 (mhwmult): New option.
5089 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5091 (msp430_is_f5_mcu): Remove prototype.
5092 (msp430_use_f5_series_hwmult): Add prototype.
5093 * config/msp430/msp430-opts.h: New file.
5094 * common/config/msp430: New directory.
5095 * common/config/msp430/msp430-common.c: New file.
5096 * config.gcc (msp430): Remove target_has_targetm_common.
5097 * doc/invoke.texi: Document -mhwmult command line option.
5099 2014-04-23 Nick Clifton <nickc@redhat.com>
5101 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5102 default-manifest.o if it can be found in the search path.
5103 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5105 2014-04-23 Terry Guo <terry.guo@arm.com>
5107 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5109 2014-04-23 Richard Biener <rguenther@suse.de>
5112 * tree-inline.c (declare_return_variable): Use mark_addressable.
5114 2014-04-23 Richard Biener <rguenther@suse.de>
5117 * loop-init.c (loop_optimizer_init): Make sure to apply
5118 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5120 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5123 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5125 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5126 if flag_sanitize_undefined_trap_on_error.
5127 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5128 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5129 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5130 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5131 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5132 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5133 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5134 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5135 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5136 * ubsan.c (ubsan_instrument_unreachable): Return
5137 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5138 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5139 if flag_sanitize_undefined_trap_on_error and
5140 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5141 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5142 instrument_bool_enum_load): Emit __builtin_trap () if
5143 flag_sanitize_undefined_trap_on_error and
5144 __builtin_handle_*_abort () if !flag_sanitize_recover.
5145 * doc/invoke.texi (-fsanitize-recover,
5146 -fsanitize-undefined-trap-on-error): Document.
5148 2014-04-22 Christian Bruel <christian.bruel@st.com>
5150 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5151 Force immediates to SImode.
5153 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5155 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5157 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5158 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5159 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5160 (nios2_fpu_insn): Add entry for round.
5161 (N2FPU_NO_ERRNO_P): Define.
5162 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5164 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5166 2014-04-22 Richard Henderson <rth@redhat.com>
5168 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5169 (add<GPI>3_compare0): Remove leading * from name.
5170 (add<GPI>3_carryin): Likewise.
5171 (sub<GPI>3_compare0): Likewise.
5172 (sub<GPI>3_carryin): Likewise.
5173 (<su_optab>mulditi3): New expander.
5174 (multi3): New expander.
5175 (madd<GPI>): Remove leading * from name.
5177 2014-04-22 Martin Jambor <mjambor@suse.cz>
5179 * cgraphclones.c (cgraph_function_versioning): Copy
5180 ipa_transforms_to_apply instead of asserting it is empty.
5182 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5185 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5186 on count_exp to get mode.
5188 2014-04-22 Andrew Pinski <apinski@cavium.com>
5190 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5191 Handle TLS for ILP32.
5192 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5193 (tlsie_small_<mode>): this and handle PTR.
5194 (tlsie_small_sidi): New pattern.
5195 (tlsle_small): Change to an expand to handle ILP32.
5196 (tlsle_small_<mode>): New pattern.
5197 (tlsdesc_small): Rename to ...
5198 (tlsdesc_small_<mode>): this and handle PTR.
5200 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5202 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5204 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5206 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5207 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5208 (aarch64_types_signed_poly_qualifiers): Likewise.
5209 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5210 (aarch64_types_poly_signed_qualifiers): Likewise.
5211 (TYPES_REINTERP_SS): Type macro added.
5212 (TYPES_REINTERP_SU): Likewise.
5213 (TYPES_REINTERP_SP): Likewise.
5214 (TYPES_REINTERP_US): Likewise.
5215 (TYPES_REINTERP_PS): Likewise.
5216 (aarch64_fold_builtin): New expression folding added.
5217 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5218 Declarations removed.
5219 (REINTERP_SS): Declarations added.
5220 (REINTERP_US): Likewise.
5221 (REINTERP_PS): Likewise.
5222 (REINTERP_SU): Likewise.
5223 (REINTERP_SP): Likewise.
5224 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5225 (vreinterpretq_p8_f64): Likewise.
5226 (vreinterpret_p16_f64): Likewise.
5227 (vreinterpretq_p16_f64): Likewise.
5228 (vreinterpret_f32_f64): Likewise.
5229 (vreinterpretq_f32_f64): Likewise.
5230 (vreinterpret_f64_f32): Likewise.
5231 (vreinterpret_f64_p8): Likewise.
5232 (vreinterpret_f64_p16): Likewise.
5233 (vreinterpret_f64_s8): Likewise.
5234 (vreinterpret_f64_s16): Likewise.
5235 (vreinterpret_f64_s32): Likewise.
5236 (vreinterpret_f64_s64): Likewise.
5237 (vreinterpret_f64_u8): Likewise.
5238 (vreinterpret_f64_u16): Likewise.
5239 (vreinterpret_f64_u32): Likewise.
5240 (vreinterpret_f64_u64): Likewise.
5241 (vreinterpretq_f64_f32): Likewise.
5242 (vreinterpretq_f64_p8): Likewise.
5243 (vreinterpretq_f64_p16): Likewise.
5244 (vreinterpretq_f64_s8): Likewise.
5245 (vreinterpretq_f64_s16): Likewise.
5246 (vreinterpretq_f64_s32): Likewise.
5247 (vreinterpretq_f64_s64): Likewise.
5248 (vreinterpretq_f64_u8): Likewise.
5249 (vreinterpretq_f64_u16): Likewise.
5250 (vreinterpretq_f64_u32): Likewise.
5251 (vreinterpretq_f64_u64): Likewise.
5252 (vreinterpret_s64_f64): Likewise.
5253 (vreinterpretq_s64_f64): Likewise.
5254 (vreinterpret_u64_f64): Likewise.
5255 (vreinterpretq_u64_f64): Likewise.
5256 (vreinterpret_s8_f64): Likewise.
5257 (vreinterpretq_s8_f64): Likewise.
5258 (vreinterpret_s16_f64): Likewise.
5259 (vreinterpretq_s16_f64): Likewise.
5260 (vreinterpret_s32_f64): Likewise.
5261 (vreinterpretq_s32_f64): Likewise.
5262 (vreinterpret_u8_f64): Likewise.
5263 (vreinterpretq_u8_f64): Likewise.
5264 (vreinterpret_u16_f64): Likewise.
5265 (vreinterpretq_u16_f64): Likewise.
5266 (vreinterpret_u32_f64): Likewise.
5267 (vreinterpretq_u32_f64): Likewise.
5269 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5271 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5272 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5273 (vreinterpret_p8_s8): Likewise.
5274 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5275 (vreinterpret_p8_s16): Likewise.
5276 (vreinterpret_p8_s32): Likewise.
5277 (vreinterpret_p8_s64): Likewise.
5278 (vreinterpret_p8_f32): Likewise.
5279 (vreinterpret_p8_u8): Likewise.
5280 (vreinterpret_p8_u16): Likewise.
5281 (vreinterpret_p8_u32): Likewise.
5282 (vreinterpret_p8_u64): Likewise.
5283 (vreinterpret_p8_p16): Likewise.
5284 (vreinterpretq_p8_s8): Likewise.
5285 (vreinterpretq_p8_s16): Likewise.
5286 (vreinterpretq_p8_s32): Likewise.
5287 (vreinterpretq_p8_s64): Likewise.
5288 (vreinterpretq_p8_f32): Likewise.
5289 (vreinterpretq_p8_u8): Likewise.
5290 (vreinterpretq_p8_u16): Likewise.
5291 (vreinterpretq_p8_u32): Likewise.
5292 (vreinterpretq_p8_u64): Likewise.
5293 (vreinterpretq_p8_p16): Likewise.
5294 (vreinterpret_p16_s8): Likewise.
5295 (vreinterpret_p16_s16): Likewise.
5296 (vreinterpret_p16_s32): Likewise.
5297 (vreinterpret_p16_s64): Likewise.
5298 (vreinterpret_p16_f32): Likewise.
5299 (vreinterpret_p16_u8): Likewise.
5300 (vreinterpret_p16_u16): Likewise.
5301 (vreinterpret_p16_u32): Likewise.
5302 (vreinterpret_p16_u64): Likewise.
5303 (vreinterpret_p16_p8): Likewise.
5304 (vreinterpretq_p16_s8): Likewise.
5305 (vreinterpretq_p16_s16): Likewise.
5306 (vreinterpretq_p16_s32): Likewise.
5307 (vreinterpretq_p16_s64): Likewise.
5308 (vreinterpretq_p16_f32): Likewise.
5309 (vreinterpretq_p16_u8): Likewise.
5310 (vreinterpretq_p16_u16): Likewise.
5311 (vreinterpretq_p16_u32): Likewise.
5312 (vreinterpretq_p16_u64): Likewise.
5313 (vreinterpretq_p16_p8): Likewise.
5314 (vreinterpret_f32_s8): Likewise.
5315 (vreinterpret_f32_s16): Likewise.
5316 (vreinterpret_f32_s32): Likewise.
5317 (vreinterpret_f32_s64): Likewise.
5318 (vreinterpret_f32_u8): Likewise.
5319 (vreinterpret_f32_u16): Likewise.
5320 (vreinterpret_f32_u32): Likewise.
5321 (vreinterpret_f32_u64): Likewise.
5322 (vreinterpret_f32_p8): Likewise.
5323 (vreinterpret_f32_p16): Likewise.
5324 (vreinterpretq_f32_s8): Likewise.
5325 (vreinterpretq_f32_s16): Likewise.
5326 (vreinterpretq_f32_s32): Likewise.
5327 (vreinterpretq_f32_s64): Likewise.
5328 (vreinterpretq_f32_u8): Likewise.
5329 (vreinterpretq_f32_u16): Likewise.
5330 (vreinterpretq_f32_u32): Likewise.
5331 (vreinterpretq_f32_u64): Likewise.
5332 (vreinterpretq_f32_p8): Likewise.
5333 (vreinterpretq_f32_p16): Likewise.
5334 (vreinterpret_s64_s8): Likewise.
5335 (vreinterpret_s64_s16): Likewise.
5336 (vreinterpret_s64_s32): Likewise.
5337 (vreinterpret_s64_f32): Likewise.
5338 (vreinterpret_s64_u8): Likewise.
5339 (vreinterpret_s64_u16): Likewise.
5340 (vreinterpret_s64_u32): Likewise.
5341 (vreinterpret_s64_u64): Likewise.
5342 (vreinterpret_s64_p8): Likewise.
5343 (vreinterpret_s64_p16): Likewise.
5344 (vreinterpretq_s64_s8): Likewise.
5345 (vreinterpretq_s64_s16): Likewise.
5346 (vreinterpretq_s64_s32): Likewise.
5347 (vreinterpretq_s64_f32): Likewise.
5348 (vreinterpretq_s64_u8): Likewise.
5349 (vreinterpretq_s64_u16): Likewise.
5350 (vreinterpretq_s64_u32): Likewise.
5351 (vreinterpretq_s64_u64): Likewise.
5352 (vreinterpretq_s64_p8): Likewise.
5353 (vreinterpretq_s64_p16): Likewise.
5354 (vreinterpret_u64_s8): Likewise.
5355 (vreinterpret_u64_s16): Likewise.
5356 (vreinterpret_u64_s32): Likewise.
5357 (vreinterpret_u64_s64): Likewise.
5358 (vreinterpret_u64_f32): Likewise.
5359 (vreinterpret_u64_u8): Likewise.
5360 (vreinterpret_u64_u16): Likewise.
5361 (vreinterpret_u64_u32): Likewise.
5362 (vreinterpret_u64_p8): Likewise.
5363 (vreinterpret_u64_p16): Likewise.
5364 (vreinterpretq_u64_s8): Likewise.
5365 (vreinterpretq_u64_s16): Likewise.
5366 (vreinterpretq_u64_s32): Likewise.
5367 (vreinterpretq_u64_s64): Likewise.
5368 (vreinterpretq_u64_f32): Likewise.
5369 (vreinterpretq_u64_u8): Likewise.
5370 (vreinterpretq_u64_u16): Likewise.
5371 (vreinterpretq_u64_u32): Likewise.
5372 (vreinterpretq_u64_p8): Likewise.
5373 (vreinterpretq_u64_p16): Likewise.
5374 (vreinterpret_s8_s16): Likewise.
5375 (vreinterpret_s8_s32): Likewise.
5376 (vreinterpret_s8_s64): Likewise.
5377 (vreinterpret_s8_f32): Likewise.
5378 (vreinterpret_s8_u8): Likewise.
5379 (vreinterpret_s8_u16): Likewise.
5380 (vreinterpret_s8_u32): Likewise.
5381 (vreinterpret_s8_u64): Likewise.
5382 (vreinterpret_s8_p8): Likewise.
5383 (vreinterpret_s8_p16): Likewise.
5384 (vreinterpretq_s8_s16): Likewise.
5385 (vreinterpretq_s8_s32): Likewise.
5386 (vreinterpretq_s8_s64): Likewise.
5387 (vreinterpretq_s8_f32): Likewise.
5388 (vreinterpretq_s8_u8): Likewise.
5389 (vreinterpretq_s8_u16): Likewise.
5390 (vreinterpretq_s8_u32): Likewise.
5391 (vreinterpretq_s8_u64): Likewise.
5392 (vreinterpretq_s8_p8): Likewise.
5393 (vreinterpretq_s8_p16): Likewise.
5394 (vreinterpret_s16_s8): Likewise.
5395 (vreinterpret_s16_s32): Likewise.
5396 (vreinterpret_s16_s64): Likewise.
5397 (vreinterpret_s16_f32): Likewise.
5398 (vreinterpret_s16_u8): Likewise.
5399 (vreinterpret_s16_u16): Likewise.
5400 (vreinterpret_s16_u32): Likewise.
5401 (vreinterpret_s16_u64): Likewise.
5402 (vreinterpret_s16_p8): Likewise.
5403 (vreinterpret_s16_p16): Likewise.
5404 (vreinterpretq_s16_s8): Likewise.
5405 (vreinterpretq_s16_s32): Likewise.
5406 (vreinterpretq_s16_s64): Likewise.
5407 (vreinterpretq_s16_f32): Likewise.
5408 (vreinterpretq_s16_u8): Likewise.
5409 (vreinterpretq_s16_u16): Likewise.
5410 (vreinterpretq_s16_u32): Likewise.
5411 (vreinterpretq_s16_u64): Likewise.
5412 (vreinterpretq_s16_p8): Likewise.
5413 (vreinterpretq_s16_p16): Likewise.
5414 (vreinterpret_s32_s8): Likewise.
5415 (vreinterpret_s32_s16): Likewise.
5416 (vreinterpret_s32_s64): Likewise.
5417 (vreinterpret_s32_f32): Likewise.
5418 (vreinterpret_s32_u8): Likewise.
5419 (vreinterpret_s32_u16): Likewise.
5420 (vreinterpret_s32_u32): Likewise.
5421 (vreinterpret_s32_u64): Likewise.
5422 (vreinterpret_s32_p8): Likewise.
5423 (vreinterpret_s32_p16): Likewise.
5424 (vreinterpretq_s32_s8): Likewise.
5425 (vreinterpretq_s32_s16): Likewise.
5426 (vreinterpretq_s32_s64): Likewise.
5427 (vreinterpretq_s32_f32): Likewise.
5428 (vreinterpretq_s32_u8): Likewise.
5429 (vreinterpretq_s32_u16): Likewise.
5430 (vreinterpretq_s32_u32): Likewise.
5431 (vreinterpretq_s32_u64): Likewise.
5432 (vreinterpretq_s32_p8): Likewise.
5433 (vreinterpretq_s32_p16): Likewise.
5434 (vreinterpret_u8_s8): Likewise.
5435 (vreinterpret_u8_s16): Likewise.
5436 (vreinterpret_u8_s32): Likewise.
5437 (vreinterpret_u8_s64): Likewise.
5438 (vreinterpret_u8_f32): Likewise.
5439 (vreinterpret_u8_u16): Likewise.
5440 (vreinterpret_u8_u32): Likewise.
5441 (vreinterpret_u8_u64): Likewise.
5442 (vreinterpret_u8_p8): Likewise.
5443 (vreinterpret_u8_p16): Likewise.
5444 (vreinterpretq_u8_s8): Likewise.
5445 (vreinterpretq_u8_s16): Likewise.
5446 (vreinterpretq_u8_s32): Likewise.
5447 (vreinterpretq_u8_s64): Likewise.
5448 (vreinterpretq_u8_f32): Likewise.
5449 (vreinterpretq_u8_u16): Likewise.
5450 (vreinterpretq_u8_u32): Likewise.
5451 (vreinterpretq_u8_u64): Likewise.
5452 (vreinterpretq_u8_p8): Likewise.
5453 (vreinterpretq_u8_p16): Likewise.
5454 (vreinterpret_u16_s8): Likewise.
5455 (vreinterpret_u16_s16): Likewise.
5456 (vreinterpret_u16_s32): Likewise.
5457 (vreinterpret_u16_s64): Likewise.
5458 (vreinterpret_u16_f32): Likewise.
5459 (vreinterpret_u16_u8): Likewise.
5460 (vreinterpret_u16_u32): Likewise.
5461 (vreinterpret_u16_u64): Likewise.
5462 (vreinterpret_u16_p8): Likewise.
5463 (vreinterpret_u16_p16): Likewise.
5464 (vreinterpretq_u16_s8): Likewise.
5465 (vreinterpretq_u16_s16): Likewise.
5466 (vreinterpretq_u16_s32): Likewise.
5467 (vreinterpretq_u16_s64): Likewise.
5468 (vreinterpretq_u16_f32): Likewise.
5469 (vreinterpretq_u16_u8): Likewise.
5470 (vreinterpretq_u16_u32): Likewise.
5471 (vreinterpretq_u16_u64): Likewise.
5472 (vreinterpretq_u16_p8): Likewise.
5473 (vreinterpretq_u16_p16): Likewise.
5474 (vreinterpret_u32_s8): Likewise.
5475 (vreinterpret_u32_s16): Likewise.
5476 (vreinterpret_u32_s32): Likewise.
5477 (vreinterpret_u32_s64): Likewise.
5478 (vreinterpret_u32_f32): Likewise.
5479 (vreinterpret_u32_u8): Likewise.
5480 (vreinterpret_u32_u16): Likewise.
5481 (vreinterpret_u32_u64): Likewise.
5482 (vreinterpret_u32_p8): Likewise.
5483 (vreinterpret_u32_p16): Likewise.
5484 (vreinterpretq_u32_s8): Likewise.
5485 (vreinterpretq_u32_s16): Likewise.
5486 (vreinterpretq_u32_s32): Likewise.
5487 (vreinterpretq_u32_s64): Likewise.
5488 (vreinterpretq_u32_f32): Likewise.
5489 (vreinterpretq_u32_u8): Likewise.
5490 (vreinterpretq_u32_u16): Likewise.
5491 (vreinterpretq_u32_u64): Likewise.
5492 (vreinterpretq_u32_p8): Likewise.
5493 (vreinterpretq_u32_p16): Likewise.
5495 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5497 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5499 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5501 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5502 (vqnegd_s64): Likewise.
5503 (vqabs_s64): Likewise.
5504 (vqabsd_s64): Likewise.
5506 2014-04-22 Richard Henderson <rth@redhat.com>
5508 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5509 computation to the top of the loop.
5511 2014-04-22 Renlin <renlin.li@arm.com>
5512 Jiong Wang <jiong.wang@arm.com>
5514 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5515 * config/aarch64/aarch64.c (aarch64_layout_frame)
5516 (aarch64_initial_elimination_offset): Likewise.
5518 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5520 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5523 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5525 * machmode.h (bitwise_mode_for_mode): Declare.
5526 * stor-layout.h (bitwise_type_for_mode): Likewise.
5527 * stor-layout.c (bitwise_mode_for_mode): New function.
5528 (bitwise_type_for_mode): Likewise.
5529 * builtins.c (fold_builtin_memory_op): Use it instead of
5530 int_mode_for_mode and build_nonstandard_integer_type.
5532 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5534 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5535 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5536 (*-*-solaris2*): Simplify.
5537 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5538 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5539 *-*-solaris2.9* handling.
5541 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5543 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5544 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5546 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5547 * configure: Regenerate.
5549 * config/i386/sol2-9.h: Remove.
5551 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5552 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5553 Remove Solaris 9 references.
5555 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5557 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5558 (floatuns<GPI:mode><GPF:mode>2): Remove.
5559 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5560 and floatuns conversions.
5561 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5562 and floatuns conversions.
5563 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5564 (w1,w2): New mode attributes for inequal width conversions.
5566 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5568 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5569 the output asm format.
5571 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5573 * config/aarch64/aarch64-simd.md
5574 (aarch64_cm<optab>di): Always split.
5575 (*aarch64_cm<optab>di): New.
5576 (aarch64_cmtstdi): Always split.
5577 (*aarch64_cmtstdi): New.
5579 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5581 PR tree-optimization/60823
5582 * omp-low.c (ipa_simd_modify_function_body): Go through
5583 all SSA_NAMEs and for those refering to vector arguments
5584 which are going to be replaced adjust SSA_NAME_VAR and,
5585 if it is a default definition, change it into a non-default
5586 definition assigned at the beginning of function from new_decl.
5587 (ipa_simd_modify_stmt_ops): Rewritten.
5588 * tree-dfa.c (set_ssa_default_def): When removing default def,
5589 check for NULL loc instead of NULL *loc.
5591 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5593 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5594 restrictions on core registers for DImode values in Thumb2.
5596 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5598 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5599 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5601 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5603 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5604 (*iordi_notzesidi_di): Likewise.
5605 (*iordi_notsesidi_di): Likewise.
5607 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5609 * config/arm/arm-protos.h (tune_params): New struct members.
5610 * config/arm/arm.c: Initialise tune_params per processor.
5611 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5612 for speed, based on new tune_params.
5614 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5616 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5617 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5618 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5619 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5620 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5621 (vrnda_f64): Likewise.
5622 (vrndi_f64): Likewise.
5623 (vrndm_f64): Likewise.
5624 (vrndn_f64): Likewise.
5625 (vrndp_f64): Likewise.
5626 (vrndx_f64): Likewise.
5628 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5630 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5631 GET_MODE_SIZE argument is enum machine_mode.
5633 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5636 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5637 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5639 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5642 * asan.c (asan_emit_stack_protection): Force the base to align to
5643 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5644 appropriate bits if STRICT_ALIGNMENT.
5645 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5647 (expand_used_vars): Leave a space in the stack frame for alignment
5648 if STRICT_ALIGNMENT.
5650 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5652 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5654 (gimple_store_p): Likewise.
5655 (gimple_assign_load_p): Likewise.
5656 (gimple_assign_cast_p): Likewise.
5657 (gimple_clobber_p): Likewise.
5659 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5660 rather than a gimple.
5661 (gimple_assign_cast_p): Likewise.
5663 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5666 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5667 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5669 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5670 more debug information for E500 if -mdebug=reg.
5672 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5675 * config/i386/i386.c (ix86_expand_builtin)
5676 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5677 register for target RTX.
5678 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5680 2014-04-18 Cong Hou <congh@google.com>
5682 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5683 the widen-mult pattern by handling two operands with different sizes,
5684 and operands whose size is smaller than half of the result type.
5686 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5688 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5689 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5690 (do_estimate_edge_time): Compute it.
5691 * ipa-inline.c (want_inline_small_function_p): Bypass
5692 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5694 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5696 * ipa-inline.c (spec_rem): New static variable.
5697 (dump_overall_stats): New function.
5698 (dump_inline_stats): New function.
5700 2014-04-18 Richard Henderson <rth@redhat.com>
5702 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5703 to GET_MODE_SIZE, not a reg_class_t.
5705 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5707 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5708 (vsx_xxmrglw_<mode>): Likewise.
5710 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5713 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5714 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5715 (rs6000_init_hard_regno_mode_ok): Likewise.
5717 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5719 * ipa-inline.c (inline_small_functions): Account only non-cold
5721 * doc/invoke.texi (inline-unit-growth): Update documentation.
5723 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5725 * config/rs6000/rs6000.md (addti3, subti3): New.
5727 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5730 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5731 comment. Check optimize_insn_for_size_p instead of
5732 optimize_insn_for_speed_p.
5734 2014-04-17 Martin Jambor <mjambor@suse.cz>
5736 * gimple-iterator.c (gsi_start_edge): New function.
5737 * gimple-iterator.h (gsi_start_edge): Declare.
5738 * tree-sra.c (single_non_eh_succ): New function.
5739 (disqualify_ops_if_throwing_stmt): Renamed to
5740 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5741 having one non-EH successor BB.
5742 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5743 generate loads into replacements.
5744 (sra_modify_assign): Likewise and and also use the simple path for
5746 (sra_modify_function_body): Commit statements on edges.
5748 2014-04-17 Richard Biener <rguenther@suse.de>
5751 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5752 comparison results and add clarifying comment.
5754 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5756 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5757 (blank_mode): Initialize it.
5758 (emit_mode_size_inline, emit_mode_nunits_inline,
5759 emit_mode_inner_inline): New functions.
5760 (emit_insn_modes_h): Call them and surround their output with
5761 #if GCC_VERSION >= 4001 ... #endif.
5762 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5763 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5764 mode_* arrays if the argument is __builtin_constant_p.
5765 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5766 is enum machine_mode.
5768 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5770 * passes.c (opt_pass::execute): Adjust.
5771 (pass_manager::execute_pass_mode_switching): Likewise.
5772 (early_local_passes::execute): Likewise.
5773 (execute_one_pass): Pass cfun to the pass's execute method.
5774 * tree-pass.h (opt_pass::execute): Add function * argument.
5775 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5776 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5777 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5778 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5779 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5780 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5781 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5782 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5783 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5784 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5785 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5786 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5787 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5788 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5789 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5790 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5791 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5792 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5793 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5794 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5795 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5796 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5797 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5798 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5799 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5800 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5801 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5802 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5803 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5806 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5808 * passes.c (opt_pass::gate): Take function * argument.
5809 (gate_all_early_local_passes): Merge into
5810 (early_local_passes::gate): this.
5811 (gate_all_early_optimizations): Merge into
5812 (all_early_optimizations::gate): this.
5813 (gate_all_optimizations): Mege into
5814 (all_optimizations::gate): this.
5815 (gate_all_optimizations_g): Merge into
5816 (all_optimizations_g::gate): this.
5817 (gate_rest_of_compilation): Mege into
5818 (rest_of_compilation::gate): this.
5819 (gate_postreload): Merge into
5820 (postreload::gate): this.
5821 (dump_one_pass): Pass cfun to the pass's gate method.
5822 (execute_ipa_summary_passes): Likewise.
5823 (execute_one_pass): Likewise.
5824 (ipa_write_summaries_2): Likewise.
5825 (ipa_write_optimization_summaries_1): Likewise.
5826 (ipa_read_summaries_1): Likewise.
5827 (ipa_read_optimization_summaries_1): Likewise.
5828 (execute_ipa_stmt_fixups): Likewise.
5829 * tree-pass.h (opt_pass::gate): Add function * argument.
5830 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5831 combine-stack-adj.c, combine.c, compare-elim.c,
5832 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5833 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5834 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5835 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5836 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5837 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5838 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5839 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5840 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5841 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5842 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5843 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5844 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5845 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5846 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5847 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5848 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5849 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5850 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5851 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5852 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5853 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5854 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5855 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5856 var-tracking.c, vtable-verify.c, web.c: Adjust.
5858 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5860 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5861 * configure: Regenerate.
5863 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5865 * passes.c (dump_one_pass): don't check pass->has_gate.
5866 (execute_ipa_summary_passes): Likewise.
5867 (execute_one_pass): Likewise.
5868 (ipa_write_summaries_2): Likewise.
5869 (ipa_write_optimization_summaries_1): Likewise.
5870 (ipa_read_optimization_summaries_1): Likewise.
5871 (execute_ipa_stmt_fixups): Likewise.
5872 * tree-pass.h (pass_data::has_gate): Remove.
5873 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5874 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5875 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5876 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5877 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5878 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5879 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5880 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5881 gimple-low.c, gimple-ssa-isolate-paths.c,
5882 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5883 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5884 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5885 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5886 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5887 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5888 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5889 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5890 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5891 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5892 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5893 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5894 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5895 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5896 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5897 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5898 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5899 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5900 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5901 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5902 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5903 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5906 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5908 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5910 * passes.c (pass_manager::register_dump_files_1): Merge into
5911 (pass_manager::register_dump_files): this, and remove its handling of
5912 properties since the pass always has the properties anyway.
5913 (pass_manager::pass_manager): Adjust.
5915 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5917 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5918 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5919 dealing with properties.
5920 (pass_manager::register_dump_files): Adjust.
5922 2014-03-20 Mark Wielaard <mjw@redhat.com>
5924 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5925 then represent the bound as normal constant value.
5927 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5930 Forward port from 4.8 branch
5931 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5933 * config/i386/bmiintrin.h (_blsi_u32): New.
5937 (_blsmsk_u32): Ditto.
5938 (_blsmsk_u64): Ditto.
5939 (_tzcnt_u32): Ditto.
5940 (_tzcnt_u64): Ditto.
5942 2014-04-17 Kito Cheng <kito@0xlab.org>
5944 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5946 2014-04-17 Richard Biener <rguenther@suse.de>
5949 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5950 boolean results for comparisons.
5952 2014-04-17 Richard Biener <rguenther@suse.de>
5954 PR tree-optimization/60836
5955 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5956 initial PHI args to be gimple values.
5958 2014-04-17 Richard Biener <rguenther@suse.de>
5960 PR tree-optimization/60841
5961 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5962 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5963 of stmts to SLP build.
5964 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5965 (vect_analyze_slp): Likewise.
5966 (vect_analyze_slp_instance): Likewise.
5967 (vect_build_slp_tree): Limit overall SLP tree growth.
5968 * tree-vectorizer.h (vect_analyze_data_refs,
5969 vect_analyze_slp): Adjust prototypes.
5971 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5973 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5976 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5978 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5979 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5980 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5981 for TARGET_SLOW_PSHUFB
5983 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5985 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5986 * config/i386/i386.c (intel_cost): Ditto.
5988 2014-04-17 Joey Ye <joey.ye@arm.com>
5990 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5992 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5994 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5997 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5999 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6000 (type_all_derivations_known_p): New predicate.
6001 (type_all_ctors_visible_p): New predicate.
6002 (type_possibly_instantiated_p): New predicate.
6003 (get_odr_type): Compute all_derivations_known.
6004 (dump_odr_type): Dump the flag.
6005 (maybe_record_type): Cleanup.
6006 (record_target_from_binfo): Add bases_to_consider array;
6007 record bases for types w/o instances and skip CXX destructor.
6008 (possible_polymorphic_call_targets_1): Add bases_to_consider
6009 and consider_construction parameters; check if type may have instance.
6010 (get_polymorphic_call_info): Set maybe_in_construction to true
6011 when we know nothing.
6012 (record_targets_from_bases): Skip CXX destructors; they are
6013 never called for types in construction.
6014 (possible_polymorphic_call_targets): Do not record target when
6015 type may not have instance.
6017 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6020 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6021 external aliases alive, too.
6023 2014-04-16 Andrew Pinski <apinski@cavium.com>
6025 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6028 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6030 * final.c (compute_alignments): Do not apply loop alignment to a block
6031 falling through to the exit.
6033 2014-04-16 Catherine Moore <clm@codesourcery.com>
6035 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6036 Adjust constraints for microMIPS store patterns.
6038 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6040 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6042 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6044 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6045 (append_use): Run at -O0.
6046 (append_vdef): Likewise.
6047 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6048 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6050 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6052 PR tree-optimization/60844
6053 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6054 (propagate_op_to_single_use, remove_visited_stmt_chain,
6055 linearize_expr, repropagate_negates, reassociate_bb): Use it
6056 instead of gsi_remove.
6058 2014-04-16 Martin Jambor <mjambor@suse.cz>
6060 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6061 ipa_transforms_to_apply.
6062 (cgraph_function_versioning): Assert that old_node has empty
6063 ipa_transforms_to_apply.
6064 * trans-mem.c (ipa_tm_create_version): Likewise.
6065 * tree-inline.c (tree_function_versioning): Do not duplicate
6066 ipa_transforms_to_apply.
6068 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6071 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6073 Pass necessary as flags on 64-bit Solaris/x86.
6074 Use lowercase relocs for x86_64-*-*.
6075 * configure: Regenerate.
6077 2014-04-15 Jan Hubicka <jh@suse.cz>
6079 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6080 (maybe_record_node, likely_target_p): Use it.
6082 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6085 Revert following patch
6087 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6090 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6091 software floating point or no floating point registers, do not
6092 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6093 in GPRs that occurs after we tested for GPRs that would never be
6096 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6097 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6098 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6099 specifically allow DDmode, since that does not use the SPE SIMD
6102 2014-03-21 Mark Wielaard <mjw@redhat.com>
6104 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6105 as unsigned or int depending on type and value used.
6107 2014-04-15 Richard Biener <rguenther@suse.de>
6109 PR rtl-optimization/56965
6110 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6111 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6113 * alias.c (true_dependence_1): Do not call
6114 nonoverlapping_component_refs_p.
6115 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6116 nonoverlapping_component_refs_p.
6117 (indirect_refs_may_alias_p): Likewise.
6119 2014-04-15 Teresa Johnson <tejohnson@google.com>
6121 * cfg.c (dump_bb_info): Fix flags check.
6122 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6124 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6126 PR rtl-optimization/60663
6127 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6130 2014-04-15 Richard Biener <rguenther@suse.de>
6132 * lto-streamer.h (LTO_major_version): Bump to 4.
6134 2014-04-15 Richard Biener <rguenther@suse.de>
6136 * common.opt (lto_partition_model): New enum.
6137 (flto-partition=): Merge separate options with a single with argument,
6138 add -flto-partition=one support.
6139 * flag-types.h (enum lto_partition_model): Declare.
6140 * opts.c (finish_options): Remove duplicate -flto-partition=
6142 * lto-wrapper.c (run_gcc): Adjust.
6144 2014-04-15 Richard Biener <rguenther@suse.de>
6146 * alias.c (ncr_compar): New function.
6147 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6149 2014-04-15 Richard Biener <rguenther@suse.de>
6151 * alias.c (record_component_aliases): Do not walk BINFOs.
6153 2014-04-15 Richard Biener <rguenther@suse.de>
6155 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6156 Add struct function argument and adjust.
6157 (find_func_aliases_for_call): Likewise.
6158 (find_func_aliases): Likewise.
6159 (find_func_clobbers): Likewise.
6160 (intra_create_variable_infos): Likewise.
6161 (compute_points_to_sets): Likewise.
6162 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6164 2014-04-15 Richard Biener <rguenther@suse.de>
6166 * tree.c (iterative_hash_expr): Use enum tree_code_class
6167 to store TREE_CODE_CLASS.
6168 (tree_block): Likewise.
6169 (tree_set_block): Likewise.
6170 * tree.h (fold_build_pointer_plus_loc): Use
6171 convert_to_ptrofftype_loc.
6173 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6176 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6179 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6181 * cfgloop.h (struct loop): Move force_vectorize down.
6182 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6183 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6184 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6185 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6186 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6187 * tree-core.h (enum annot_expr_kind): Add new kind values.
6188 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6189 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6191 * tree.def (ANNOTATE_EXPR): Tweak comment.
6193 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6195 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6198 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6200 * tree.h (TYPE_IDENTIFIER): Declare.
6201 * tree.c (subrange_type_for_debug_p): Use it.
6202 * godump.c (go_format_type): Likewise.
6203 * dwarf2out.c (is_cxx_auto, modified_type_die,
6204 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6205 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6207 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6210 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6212 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6214 * config/i386/i386.c (examine_argument): Return bool. Return true if
6215 parameter should be passed in memory.
6216 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6217 (construct_container): Update calls to examine_argument.
6218 (function_arg_advance_64): Ditto.
6219 (return_in_memory_32): Merge with ix86_return_in_memory.
6220 (return_in_memory_64): Ditto.
6221 (return_in_memory_ms_64): Ditto.
6223 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6225 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6226 * coverage.c (coverage_compute_profile_id): Handle externally visible
6229 2014-04-14 Martin Jambor <mjambor@suse.cz>
6231 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6232 DECL_DISREGARD_INLINE_LIMITS functions.
6234 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6237 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6239 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6242 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6243 optimize_insn_for_speed_p instead of
6244 optimize_function_for_speed_p.
6246 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6248 * doc/invoke.texi (free): Document AArch64.
6250 2014-04-14 Richard Biener <rguenther@suse.de>
6252 PR tree-optimization/60042
6253 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6254 (insert_into_preds_of_block): Do not prevent PHI insertion
6255 for REFERENCE exprs here ...
6256 (eliminate_dom_walker::before_dom_children): ... but prevent
6257 their use here under similar conditions when applied to the
6258 IL after PRE optimizations.
6260 2014-04-14 Richard Biener <rguenther@suse.de>
6262 * passes.def: Move early points-to after early SRA.
6264 2014-04-14 Richard Biener <rguenther@suse.de>
6266 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6267 check for which sign-changes we allow when forwarding
6268 a converted value into a switch.
6270 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6272 * stor-layout.c (place_field): Finalize non-constant offset for the
6275 2014-04-14 Richard Biener <rguenther@suse.de>
6277 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6279 (expand_switch_using_bit_tests_p): Likewise.
6280 (process_switch): Compute and pass on speed_p based on the
6282 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6283 optimize_bb_for_speed_p.
6285 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6287 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6288 * function.h (struct function): Rename has_force_vect_loops into
6289 has_force_vectorize_loops.
6290 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6291 (input_struct_function_base): Likewise.
6292 * lto-streamer-out.c (output_cfg): Likewise.
6293 (output_struct_function_base): Likewise.
6294 * omp-low.c (expand_omp_simd): Likewise.
6295 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6296 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6297 (version_loop_for_if_conversion): Likewise.
6298 (tree_if_conversion): Likewise.
6299 (main_tree_if_conversion): Likewise.
6300 (gate_tree_if_conversion): Likewise.
6301 * tree-inline.c (copy_loops): Likewise.
6302 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6303 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6304 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6305 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6306 * tree-vectorizer.c (vectorize_loops): Likewise.
6307 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6309 2014-04-14 Richard Biener <rguenther@suse.de>
6312 * lto-streamer-out.c (wrap_refs): New function.
6313 (lto_output): Wrap symbol references in global initializes in
6314 type-preserving MEM_REFs.
6316 2014-04-14 Christian Bruel <christian.bruel@st.com>
6318 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6320 2014-04-14 Christian Bruel <christian.bruel@st.com>
6322 * config/sh/sh.md (setmemqi): New expand pattern.
6323 * config/sh/sh.h (CLEAR_RATIO): Define.
6324 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6325 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6327 2014-04-14 Richard Biener <rguenther@suse.de>
6330 * fold-const.c (negate_expr_p): Don't negate directional rounding
6332 (fold_negate_expr): Likewise.
6334 2014-04-14 Richard Biener <rguenther@suse.de>
6336 PR tree-optimization/59817
6337 PR tree-optimization/60453
6338 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6339 recursion to catch all CHRECs in the scalar evolution and restrict
6340 the predicate for the remains appropriately.
6342 2014-04-12 Catherine Moore <clm@codesourcery.com>
6344 * config/mips/constraints.md: Add new register constraint "kb".
6345 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6346 (*movhi_internal): Likewise.
6347 (*movqi_internal): Likewise.
6348 * config/mips/mips.h (M16_STORE_REGS): New register class.
6349 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6350 (REG_CLASS_CONTENTS): Likewise.
6351 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6353 2014-04-11 Tobias Burnus <burnus@net-b.de>
6356 * doc/invoke.texi (-Wformat-signedness): Document it.
6357 (Wformat=2): Mention that this enables -Wformat-signedness.
6359 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6361 * common/config/epiphany/epiphany-common.c
6362 (epiphany_option_optimization_table): Enable section anchors by
6363 default at -O1 or higher.
6364 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6365 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6366 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6367 carries no extra cost.
6368 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6369 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6370 * config/epiphany/predicates.md (memclob_operand): New predicate.
6371 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6372 Use memclob_operand predicate and X constraint for operand 3.
6374 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6376 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6377 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6380 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6382 PR rtl-optimization/60651
6383 * mode-switching.c (optimize_mode_switching): Make sure to emit
6384 sets of a lower numbered entity before sets of a higher numbered
6385 entity to a mode of the same or lower priority.
6386 When creating a seginfo for a basic block that starts with a code
6387 label, move the insertion point past the code label.
6388 (new_seginfo): Document and enforce requirement that
6389 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6390 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6391 * doc/tm.texi: Regenerate.
6393 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6396 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6398 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6400 * BASE-VER: Set to 4.10.0.
6402 2014-04-11 Tobias Burnus <burnus@net-b.de>
6405 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6406 * doc/gcc.texi (Service): Update description in the @menu
6407 * doc/invoke.texi (Option Summary): Remove misplaced and
6410 2014-04-11 Steve Ellcey <sellcey@mips.com>
6411 Jakub Jelinek <jakub@redhat.com>
6414 * expr.c (convert_move): Use emit_store_flag_force instead of
6415 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6418 2014-04-11 Richard Biener <rguenther@suse.de>
6421 * varasm.c (assemble_alias): Avoid endless error reporting
6422 recursion by setting TREE_ASM_WRITTEN.
6424 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6426 * config/s390/s390.md: Add a splitter for NOT rtx.
6428 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6430 PR rtl-optimization/60663
6431 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6433 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6434 Jakub Jelinek <jakub@redhat.com>
6437 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6438 flag from decl_node to node.
6440 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6443 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6444 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6445 ameliorating the cases where it can be.
6447 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6450 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6452 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6453 (loadsync_<mode>): Change mode.
6454 (load_quadpti, store_quadpti): New.
6455 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6456 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6457 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6459 2014-04-09 Cong Hou <congh@google.com>
6462 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6465 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6467 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6468 instead of vnor to exploit possible fusion opportunity in the
6470 (altivec_expand_vec_perm_const_le): Likewise.
6472 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6474 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6475 (loadsync_<mode>): Change mode.
6476 (load_quadpti, store_quadpti): New.
6477 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6478 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6480 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6483 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6484 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6485 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6487 2014-04-08 Richard Biener <rguenther@suse.de>
6490 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6491 a 64bit widest int print double-int similar to on HWI64 hosts.
6493 2014-04-08 Richard Biener <rguenther@suse.de>
6495 PR tree-optimization/60785
6496 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6497 default defs properly.
6499 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6501 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6502 (Weffc++): Likewise.
6504 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6506 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6507 set completep to false rather than true.
6509 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6512 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6513 ARM_TARGET2_DWARF_FORMAT.
6515 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6518 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6519 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6522 2014-04-07 Richard Biener <rguenther@suse.de>
6524 PR tree-optimization/60766
6525 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6526 (may_eliminate_iv): Convert cand_value_at result to desired type.
6528 2014-04-07 Jason Merrill <jason@redhat.com>
6531 * common.opt (-fno-gnu-unique): Add.
6532 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6534 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6536 * haifa-sched.c: Fix outdated function reference and minor
6537 grammar errors in introductory comment.
6539 2014-04-07 Richard Biener <rguenther@suse.de>
6542 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6544 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6546 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6549 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6550 size accounting for thunks.
6551 (pa_asm_output_mi_thunk): Use final_start_function() and
6552 final_end_function() to output function start and end directives.
6554 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6556 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6557 device specific ISA/ feature information. Remove short_sp and
6558 errata_skip ds. Add avr_device_specific_features enum to have device
6560 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6561 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6562 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6563 updated device specific info.
6564 * config/avr/avr-mcus.def: Merge device specific details to
6565 dev_attribute field.
6566 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6568 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6569 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6570 assembler if RMW isa supported by current device.
6571 * config/avr/genmultilib.awk: Update as device info structure changed.
6572 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6574 2014-04-04 Cong Hou <congh@google.com>
6576 PR tree-optimization/60656
6577 * tree-vect-stmts.c (supportable_widening_operation):
6578 Fix a bug that elements in a vector with vect_used_by_reduction
6579 property are incorrectly reordered when the operation on it is not
6580 consistant with the one in reduction operation.
6582 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6584 PR rtl-optimization/60155
6585 * gcse.c (record_set_data): New function.
6586 (single_set_gcse): New function.
6587 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6588 (hoist_code): Likewise.
6589 (get_pressure_class_and_nregs): Likewise.
6591 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6593 * explow.c (probe_stack_range): Emit a final optimization blockage.
6595 2014-04-04 Anthony Green <green@moxielogic.com>
6597 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6600 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6603 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6604 flags are set only during streaming.
6605 * ipa.c (process_references, walk_polymorphic_call_targets,
6606 symtab_remove_unreachable_nodes): Drop bodies of always inline
6607 after early inlining.
6608 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6610 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6611 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6614 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6617 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6620 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6622 (cortex_a53_fdivd): Likewise.
6624 2014-04-04 Martin Jambor <mjambor@suse.cz>
6627 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6629 * cgraph.c (clone_of_p): Also return true if thunks match.
6630 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6631 cgraph_function_or_thunk_node and an obsolete comment.
6632 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6634 (build_function_decl_skip_args): Likewise.
6635 (set_new_clone_decl_and_node_flags): New function.
6636 (duplicate_thunk_for_node): Likewise.
6637 (redirect_edge_duplicating_thunks): Likewise.
6638 (cgraph_clone_node): New parameter args_to_skip, pass it to
6639 redirect_edge_duplicating_thunks which is called instead of
6640 cgraph_redirect_edge_callee.
6641 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6642 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6644 2014-04-04 Jeff Law <law@redhat.com>
6647 * config/arm/predicates.md (const_int_I_operand): New predicate.
6648 (const_int_M_operand): Similarly.
6649 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6651 (insv_t2, extv_reg, extzv_t2): Likewise.
6652 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6653 (pop_multiple_with_writeback_and_return): Likewise.
6654 (vfp_pop_multiple_with_writeback): Likewise
6656 2014-04-04 Richard Biener <rguenther@suse.de>
6659 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6660 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6662 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6663 * gimplify.c (gimple_add_tmp_var_fn): New function.
6664 * gimple-expr.h (create_tmp_reg_fn): Declare.
6665 * gimple-expr.c (create_tmp_reg_fn): New function.
6666 * gimple-low.c (record_vars_into): Don't change cfun.
6667 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6668 code generation without cfun.
6670 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6673 * Makefile.in (install-driver): Fix shell scripting.
6675 2014-04-03 Cong Hou <congh@google.com>
6677 PR tree-optimization/60505
6678 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6679 threshold of number of iterations below which no vectorization
6681 * tree-vect-loop.c (new_loop_vec_info):
6682 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6683 * tree-vect-loop.c (vect_analyze_loop_operations):
6684 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6685 * tree-vect-loop.c (vect_transform_loop):
6686 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6687 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6688 of iterations of the loop and see if we should build the epilogue.
6690 2014-04-03 Richard Biener <rguenther@suse.de>
6692 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6693 (streamer_tree_cache_create): Adjust.
6694 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6695 to allow optional nodes array.
6696 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6697 (streamer_tree_cache_append): Likewise.
6698 (streamer_tree_cache_create): Create nodes array optionally
6699 as specified by parameter.
6700 * lto-streamer-out.c (create_output_block): Avoid maintaining
6701 the node array in the writer cache.
6702 (DFS_write_tree): Remove assertion.
6703 (produce_asm_for_decls): Free the out decl state hash table early.
6704 * lto-streamer-in.c (lto_data_in_create): Adjust for
6705 streamer_tree_cache_create prototype change.
6707 2014-04-03 Richard Biener <rguenther@suse.de>
6709 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6710 set TREE_CHAIN to NULL_TREE.
6712 2014-04-03 Richard Biener <rguenther@suse.de>
6714 PR tree-optimization/60740
6715 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6716 over all GIMPLE_COND operands.
6718 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6720 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6721 (Weffc++): Remove Scott's numbering, merge lists and reference
6724 2014-04-03 Nick Clifton <nickc@redhat.com>
6726 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6729 2014-04-03 Martin Jambor <mjambor@suse.cz>
6731 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6732 mention gcc_unreachable before failing.
6733 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6736 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6739 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6740 inconsistent code and instead mark the context inconsistent.
6741 (possible_polymorphic_call_targets): For inconsistent contexts
6742 return empty complete list.
6744 2014-04-02 Anthony Green <green@moxielogic.com>
6746 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6747 (extendqisi2, extendhisi2): Define.
6748 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6749 (WCHAR_TYPE): Change to unsigned int.
6751 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6753 PR tree-optimization/60733
6754 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6755 insertion point for PHI candidates to be the end of the feeding
6756 block for the PHI argument.
6758 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6760 PR rtl-optimization/60650
6761 * lra-constraints.c (process_alt_operands): Decrease reject for
6762 earlyclobber matching.
6764 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6766 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6768 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6770 * config/spu/spu.c (pad_bb): Do not crash when the last
6771 insn is CODE_FOR_blockage.
6773 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6775 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6776 lies outside the target mode.
6778 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6781 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6782 software floating point or no floating point registers, do not
6783 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6784 in GPRs that occurs after we tested for GPRs that would never be
6787 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6788 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6789 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6790 specifically allow DDmode, since that does not use the SPE SIMD
6793 2014-04-02 Richard Biener <rguenther@suse.de>
6796 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6797 MODE_INTs. Properly use negv_optab.
6798 (expand_abs): Likewise.
6800 2014-04-02 Richard Biener <rguenther@suse.de>
6803 * Makefile.in (install-driver): Guard extra installs with special
6806 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6808 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6811 2014-04-01 Richard Henderson <rth@redhat.com>
6814 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6815 alternative enabled before register allocation.
6817 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6819 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6820 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6822 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6823 (nios2_got_address): Update nios2_large_got_address call site.
6824 (nios2_delegitimize_address): New function.
6825 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6826 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6827 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6829 2014-04-01 Martin Husemann <martin@duskware.de>
6831 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6834 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6836 PR rtl-optimization/60604
6837 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6838 check from register_operand.
6839 (register_operand): Redefine in terms of general_operand.
6840 (nonmemory_operand): Use register_operand for the non-constant cases.
6842 2014-04-01 Richard Biener <rguenther@suse.de>
6844 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6846 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6848 * doc/invoke.texi (mapp-regs): Clarify.
6850 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6852 * config/i386/avx512fintrin.h (__v32hi): Define type.
6853 (__v64qi): Likewise.
6854 (_mm512_set1_epi8): Define.
6855 (_mm512_set1_epi16): Define.
6856 (_mm512_set4_epi32): Define.
6857 (_mm512_set4_epi64): Define.
6858 (_mm512_set4_pd): Define.
6859 (_mm512_set4_ps): Define.
6860 (_mm512_setr4_epi64): Define.
6861 (_mm512_setr4_epi32): Define.
6862 (_mm512_setr4_pd): Define.
6863 (_mm512_setr4_ps): Define.
6864 (_mm512_setzero_epi32): Define.
6866 2014-03-31 Martin Jambor <mjambor@suse.cz>
6869 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6870 callsite_arguments_match_p. Updated all callers. Also check types of
6871 corresponding formal parameters and actual arguments.
6872 (not_all_callers_have_enough_arguments_p) Renamed to
6873 some_callers_have_mismatched_arguments_p.
6875 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6877 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6879 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6882 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6885 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6887 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6888 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6890 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6891 Use FMAMODE_NOVF512 mode iterator.
6892 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6893 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6894 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6895 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6897 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6898 Use VF_128_256 mode iterator.
6899 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6902 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6904 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6905 static chain if needed.
6907 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6910 * lra-constraints.c (index_part_to_reg): New.
6911 (process_address): Use it.
6913 2014-03-27 Jeff Law <law@redhat.com>
6914 Jakub Jelinek <jakub@redhat.com>
6917 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6918 gen_rtx_{PLUS,MULT} to build up the address expression.
6920 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6921 creating non-canonical RTL.
6923 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6926 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6927 functions; reorganize to make cheap checks first.
6928 (inline_small_functions): Do not estimate growth when dumping;
6930 * ipa-inline.h (inline_summary): Add min_size.
6931 (growth_likely_positive): New function.
6932 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6933 (set_cond_stmt_execution_predicate): Cleanup.
6934 (estimate_edge_size_and_time): Compute min_size.
6935 (estimate_calls_size_and_time): Likewise.
6936 (estimate_node_size_and_time): Likewise.
6937 (inline_update_overall_summary): Update min_size.
6938 (do_estimate_edge_time): Likewise.
6939 (do_estimate_edge_size): Update.
6940 (do_estimate_edge_hints): Update.
6941 (growth_likely_positive): New function.
6943 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6946 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6947 also if addr has VOIDmode.
6949 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6951 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6952 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6954 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6955 instructions as well as AdvancedSIMD loads.
6957 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6959 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6960 Use crypto_aese type.
6961 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6962 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6963 crypto_aese, crypto_aesmc. Move to types.md.
6964 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6966 * config/arm/iterators.md (crypto_type): Likewise.
6968 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6970 * cgraph.c: Include expr.h and tree-dfa.h.
6971 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6974 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6977 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6978 regs from checking multi-reg pseudos.
6980 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6982 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6984 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6986 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6987 if it would clobber the stack pointer, even temporarily.
6989 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6991 * mode-switching.c: Make small adjustments to the top comment.
6993 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6995 * config/rs6000/constraints.md (wD constraint): New constraint to
6996 match the constant integer to get the top DImode/DFmode out of a
6997 vector in a VSX register.
6999 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7000 match the constant integer to get the top DImode/DFmode out of a
7001 vector in a VSX register.
7003 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7006 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7009 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7010 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7012 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7013 Optimize vec_extract of 64-bit values, where the value being
7014 extracted is in the top word, where we can use scalar
7015 instructions. Add direct move and store support. Combine the big
7016 endian/little endian vector select load support into a single insn.
7017 (vsx_extract_<mode>_internal1): Likewise.
7018 (vsx_extract_<mode>_internal2): Likewise.
7019 (vsx_extract_<mode>_load): Likewise.
7020 (vsx_extract_<mode>_store): Likewise.
7021 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7022 combined into vsx_extract_<mode>_load.
7023 (vsx_extract_<mode>_one_le): Likewise.
7025 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7026 define the top 64-bit vector element.
7028 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7031 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7032 Document vec_vbpermq builtin.
7035 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7036 enable use of xxsldwi and xxpermdi builtin functions.
7037 (vec_xxpermdi): Likewise.
7039 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7040 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7042 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7044 PR rtl-optimization/60650
7045 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7047 (find_spills_for): New.
7048 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7049 Spill all pseudos on the second iteration.
7051 2014-03-27 Marek Polacek <polacek@redhat.com>
7054 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7057 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7059 * config/s390/s390.c (s390_can_use_return_insn): Check for
7060 call-saved FPRs on 31 bit.
7062 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7065 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7066 if they need regimplification, just drop them instead of
7067 calling gimple_regimplify_operands on them.
7069 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7072 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7073 (aarch64_frame_pointer_required): Adjust logic.
7074 (aarch64_can_eliminate): Adjust logic.
7075 (aarch64_override_options_after_change): Adjust logic.
7077 2014-03-27 Dehao Chen <dehao@google.com>
7079 * ipa-inline.c (early_inliner): Update node's inline info.
7081 2014-03-26 Dehao Chen <dehao@google.com>
7083 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7084 compiler inserted conditional jumps for NAN float check.
7086 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7088 * ubsan.h (ubsan_create_data): Change second argument's type
7089 to const location_t *.
7090 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7092 (ubsan_create_data): Change second argument to const location_t *PLOC.
7093 Create Loc field whenever PLOC is non-NULL.
7094 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7095 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7099 * real.c (real_to_integer2): Change type of low to UHWI.
7101 2014-03-26 Tobias Burnus <burnus@net-b.de>
7103 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7104 (CILK_SELF_SPECS): New define.
7105 (driver_self_specs): Use it.
7107 2014-03-26 Richard Biener <rguenther@suse.de>
7109 * tree-pretty-print.c (percent_K_format): Implement special
7110 case for LTO and its stripped down BLOCK tree.
7112 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7115 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7117 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7118 one range is range_int_cst_p, but not both, at least optimize
7119 addition/subtraction of 0 and multiplication by 0 or 1.
7120 * gimple-fold.c (gimple_fold_call): Fold
7121 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7122 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7123 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7125 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7127 PR rtl-optimization/60452
7128 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7129 <case REG>: Return 1 for invalid offsets from the frame pointer.
7131 2014-03-26 Marek Polacek <polacek@redhat.com>
7134 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7137 2014-03-26 Marek Polacek <polacek@redhat.com>
7140 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7143 2014-03-26 Marek Polacek <polacek@redhat.com>
7146 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7147 multiplication in unsigned type.
7149 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7151 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7153 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7155 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7157 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7160 * cif-code.def (UNREACHABLE) New code.
7161 * ipa-inline.c (inline_small_functions): Skip edges to
7162 __builtlin_unreachable.
7163 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7164 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7165 predicate to __bulitin_unreachable.
7166 (set_cond_stmt_execution_predicate): Fix issue when
7167 invert_tree_comparison returns ERROR_MARK.
7168 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7169 propagate to inline clones.
7170 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7172 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7173 * cgraphclones.c (cgraph_clone_node): If call destination is already
7174 ureachable, do not redirect it back.
7175 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7178 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7180 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7181 Do not modify inline clones.
7183 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7185 * config/i386/i386.md (general_sext_operand): New mode attr.
7186 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7187 don't generate (sign_extend (const_int)).
7188 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7189 operands[2]. Use We constraint instead of <i> and
7190 <general_sext_operand> predicate instead of <general_operand>.
7191 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7192 * config/i386/constraints.md (We): New constraint.
7193 * config/i386/predicates.md (x86_64_sext_operand,
7194 sext_operand): New predicates.
7196 2014-03-25 Martin Jambor <mjambor@suse.cz>
7199 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7200 inconsistent devirtualizations to __builtin_unreachable.
7202 2014-03-25 Marek Polacek <polacek@redhat.com>
7205 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7207 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7209 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7210 order of elements for big-endian.
7212 2014-03-25 Richard Biener <rguenther@suse.de>
7215 * gimplify-me.c (gimple_regimplify_operands): Update the
7218 2014-03-25 Martin Jambor <mjambor@suse.cz>
7221 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7222 (lto_output_varpool_node): Likewise.
7223 (input_overwrite_node): Likewise.
7224 (input_varpool_node): Likewise.
7226 2014-03-25 Richard Biener <rguenther@suse.de>
7228 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7229 (run_gcc): Likewise.
7231 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7233 * combine.c (simplify_compare_const): Add MODE argument.
7234 Handle mode_width 0 as very large mode_width.
7235 (try_combine, simplify_comparison): Adjust callers.
7237 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7238 type to avoid signed integer overflow.
7239 * explow.c (plus_constant): Likewise.
7241 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7243 * doc/generic.texi: Correct typos.
7245 2014-03-24 Tobias Burnus <burnus@net-b.de>
7247 * doc/invoke.texi (-flto): Expand section about
7248 using static libraries with LTO.
7250 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7252 PR rtl-optimization/60501
7253 * optabs.def (addptr3_optab): New optab.
7254 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7255 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7256 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7258 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7260 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7262 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7264 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7267 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7268 (_mm256_undefined_ps): Define.
7269 (_mm256_undefined_pd): Define.
7270 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7271 (_mm_undefined_pd): Define.
7272 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7273 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7274 (_mm512_undefined_ps): Define.
7275 (_mm512_undefined_pd): Define.
7276 Use _mm*_undefined_*.
7277 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7279 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7281 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7282 (lshr_simd): DI mode added.
7283 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7284 (aarch64_ushr_simddi): Likewise.
7285 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7286 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7287 (vshrd_n_u64): Likewise.
7289 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7291 * Makefile.in (s-macro_list): Depend on cc1.
7293 2014-03-23 Teresa Johnson <tejohnson@google.com>
7295 * ipa-utils.c (ipa_print_order): Use specified dump file.
7297 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7299 PR rtl-optimization/60601
7300 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7302 * gcc.c (eval_spec_function): Initialize save_growing_value.
7304 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7307 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7308 code == MINUS_EXPR, never swap op0 with op1.
7310 * toplev.c (init_local_tick): Avoid signed integer multiplication
7312 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7313 shift by first operand's bitsize.
7315 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7318 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7320 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7321 TARGET_ISA_64BIT_P(x).
7323 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7325 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7326 pattern for vector nor instead of subtract from splat(-1).
7327 (altivec_expand_vec_perm_const_le): Likewise.
7329 2014-03-21 Richard Henderson <rth@twiddle.net>
7332 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7333 related insns after epilogue_completed.
7335 2014-03-21 Martin Jambor <mjambor@suse.cz>
7338 * cgraph.h (symtab_node): New flag body_removed.
7339 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7340 when removing bodies.
7341 * symtab.c (dump_symtab_base): Dump body_removed flag.
7342 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7343 had their bodies removed.
7345 2014-03-21 Martin Jambor <mjambor@suse.cz>
7348 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7351 2014-03-21 Richard Biener <rguenther@suse.de>
7353 PR tree-optimization/60577
7354 * tree-core.h (struct tree_base): Document nothrow_flag use
7355 in VAR_DECL_NONALIASED.
7356 * tree.h (VAR_DECL_NONALIASED): New.
7357 (may_be_aliased): Adjust.
7358 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7360 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7362 * expr.c (expand_expr_real_1): Remove outdated comment.
7364 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7367 * ira.c (adjust_cleared_regs): Call copy_rtx on
7368 *reg_equiv[REGNO (loc)].src_p before passing it to
7369 simplify_replace_fn_rtx.
7372 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7373 into CONST, put pic register as first operand of PLUS. Use
7374 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7376 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7378 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7380 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7382 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7383 around for store forwarding issue in the FPU on the UT699.
7384 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7385 loads and operations if -mfix-ut699 is specified.
7386 (divtf3_hq): Tweak attribute.
7387 (sqrttf2_hq): Likewise.
7389 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7391 * calls.c (store_one_arg): Remove incorrect const qualification on the
7392 type of the temporary.
7393 * cfgexpand.c (expand_return): Likewise.
7394 * expr.c (expand_constructor): Likewise.
7395 (expand_expr_real_1): Likewise.
7397 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7399 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7402 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7405 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7407 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7409 * config/arm/aarch-common-protos.h
7410 (alu_cost_table): Fix spelling of "extend".
7411 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7413 2014-03-19 Richard Biener <rguenther@suse.de>
7416 * tree-core.h (tree_type_common): Re-order pointer members
7417 to reduce recursion depth during GC walks.
7419 2014-03-19 Marek Polacek <polacek@redhat.com>
7422 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7423 before accessing it.
7425 2014-03-19 Richard Biener <rguenther@suse.de>
7428 * lto-streamer-in.c (input_function): In WPA stage do not drop
7431 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7433 PR tree-optimization/60559
7434 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7435 with build_zero_cst assignment.
7437 2014-03-18 Kai Tietz <ktietz@redhat.com>
7439 PR rtl-optimization/56356
7440 * sdbout.c (sdbout_parms): Verify that parms'
7441 incoming argument is valid.
7442 (sdbout_reg_parms): Likewise.
7444 2014-03-18 Richard Henderson <rth@redhat.com>
7447 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7448 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7449 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7451 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7453 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7454 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7455 Italicize plugin event names in description. Explain that
7456 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7457 Remind that no GCC functions should be called after PLUGIN_FINISH.
7458 Explain what pragmas with expansion are.
7460 2014-03-18 Martin Liska <mliska@suse.cz>
7462 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7463 gimple call statement is update.
7464 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7465 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7467 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7470 * ubsan.c (ubsan_instrument_unreachable): Call
7471 initialize_sanitizer_builtins.
7472 (ubsan_pass): Likewise.
7475 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7476 varpool_finalize_decl instead of rest_of_decl_compilation.
7478 2014-03-18 Richard Biener <rguenther@suse.de>
7480 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7481 by using bitmap_and_compl instead of bitmap_and_compl_into.
7482 (df_rd_transfer_function): Likewise.
7484 2014-03-18 Richard Biener <rguenther@suse.de>
7486 * doc/lto.texi (fresolution): Fix typo.
7488 2014-03-18 Richard Biener <rguenther@suse.de>
7490 * doc/invoke.texi (flto): Update for changes in 4.9.
7492 2014-03-18 Richard Biener <rguenther@suse.de>
7494 * doc/loop.texi: Remove section on the removed lambda framework.
7495 Update loop docs with recent changes in preserving loop structure.
7497 2014-03-18 Richard Biener <rguenther@suse.de>
7499 * doc/lto.texi (-fresolution): Document.
7501 2014-03-18 Richard Biener <rguenther@suse.de>
7503 * doc/contrib.texi: Adjust my name.
7505 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7508 * internal-fn.c: Include diagnostic-core.h.
7509 (expand_BUILTIN_EXPECT): New function.
7510 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7511 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7512 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7513 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7515 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7516 Revert 3 argument __builtin_expect code.
7517 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7518 * gimple-fold.c (gimple_fold_call): Likewise.
7519 * tree.h (fold_builtin_expect): New prototype.
7520 * builtins.c (build_builtin_expect_predicate): Add predictor
7521 argument, if non-NULL, create 3 argument __builtin_expect.
7522 (fold_builtin_expect): No longer static. Add ARG2 argument,
7523 pass it through to build_builtin_expect_predicate.
7524 (fold_builtin_2): Adjust caller.
7525 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7526 * internal-fn.def (BUILTIN_EXPECT): New.
7528 2014-03-18 Tobias Burnus <burnus@net-b.de>
7531 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7532 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7533 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7535 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7538 * predict.c (combine_predictions_for_bb): Fix up formatting.
7539 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7540 fill what it points to if non-NULL.
7541 (tree_predict_by_opcode): Adjust caller, use the predictor.
7542 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7544 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7546 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7547 proper constant for the store mode.
7549 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7551 * symtab.c (change_decl_assembler_name): Fix transparent alias
7554 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7556 * config/aarch64/aarch64.c: Correct the comments about the
7557 aarch64 stack layout.
7559 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7561 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7562 check for GF_OMP_FOR_KIND_FOR.
7564 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7566 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7567 ymm and zmm register names.
7569 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7572 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7573 note creation for the 2010-08-31 changes.
7575 2014-03-17 Marek Polacek <polacek@redhat.com>
7578 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7579 as -fno-tree-loop-vectorize.
7580 (expand_omp_simd): Likewise.
7582 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7584 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7585 (eligible_for_call_delay): New prototype.
7586 * config/sparc/sparc.c (tls_call_delay): Rename into...
7587 (eligible_for_call_delay): ...this. Return false if the instruction
7588 cannot be put in the delay slot of a branch.
7589 (eligible_for_restore_insn): Simplify.
7590 (eligible_for_return_delay): Return false if the instruction cannot be
7591 put in the delay slot of a branch and simplify.
7592 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7593 put in the delay slot of a branch.
7594 * config/sparc/sparc.md (fix_ut699): New attribute.
7595 (tls_call_delay): Delete.
7596 (in_call_delay): Reimplement.
7597 (eligible_for_sibcall_delay): Rename into...
7598 (in_sibcall_delay): ...this.
7599 (eligible_for_return_delay): Rename into...
7600 (in_return_delay): ...this.
7601 (in_branch_delay): Reimplement.
7602 (in_uncond_branch_delay): Delete.
7603 (in_annul_branch_delay): Delete.
7605 2014-03-14 Richard Henderson <rth@redhat.com>
7608 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7609 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7610 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7611 (floathi splitters): Remove.
7612 (float<SWI48x>xf2): New pattern.
7613 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7614 code that tried to handle DImode for 32-bit, but which was excluded
7615 by the pattern's condition. Drop allocation of stack temporary.
7616 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7617 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7618 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7619 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7620 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7621 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7622 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7623 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7624 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7625 (*float<SWI48x><X87MODEF>2_i387): Remove.
7626 (all float _with_temp splitters): Remove.
7627 (*float<SWI48x><MODEF>2_i387): New pattern.
7628 (*float<SWI48><MODEF>2_sse): New pattern.
7629 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7630 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7632 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7633 Marek Polacek <polacek@redhat.com>
7636 * common.opt (dump_base_name_prefixed): New Variable.
7637 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7638 if x_dump_base_name_prefixed is already set, set it at the end.
7640 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7642 PR rtl-optimization/60508
7643 * lra-constraints.c (get_reload_reg): Add new parameter
7645 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7646 Pass the new parameter values.
7648 2014-03-14 Richard Biener <rguenther@suse.de>
7650 * common.opt: Revert unintented changes from r205065.
7653 2014-03-14 Richard Biener <rguenther@suse.de>
7656 * cfghooks.c (split_block): Properly adjust all loops the
7657 block was a latch of.
7659 2014-03-14 Martin Jambor <mjambor@suse.cz>
7662 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7665 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7668 * config/avr/avr.c (avr_set_current_function): Pass function name
7669 through default_strip_name_encoding before sanity checking instead
7670 of skipping the first char of the assembler name.
7672 2014-03-13 Richard Henderson <rth@redhat.com>
7675 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7676 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7677 * config/i386/i386-protos.h: Likewise.
7678 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7679 in the expander instead of a splitter.
7680 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7681 any possibility of requiring a memory.
7682 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7683 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7684 (fp branch splitters): Update for ix86_split_fp_branch.
7685 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7686 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7687 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7688 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7689 (*fop_<MODEF>_3_i387): Likewise.
7690 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7691 (splitters for the fop_* register patterns): Remove.
7692 (fscalexf4_i387): Rename from *fscalexf4_i387.
7693 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7695 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7697 PR tree-optimization/59779
7698 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7699 type for bitsize and maxsize instead of HOST_WIDE_INT.
7701 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7703 PR rtl-optimization/57320
7704 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7705 the CFG after thread_prologue_and_epilogue_insns.
7707 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7709 PR rtl-optimization/57189
7710 * lra-constraints.c (process_alt_operands): Disfavor spilling
7713 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7715 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7717 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7719 PR tree-optimization/59025
7721 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7722 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7724 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7727 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7728 calls of avr_out_plus_1.
7730 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7732 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7733 BB's single pred and update the father loop's latch info later.
7735 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7737 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7741 (VEC_base): Likewise.
7742 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7743 registers, we need to swap double words in little endian mode.
7745 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7746 to be a container mode for 128-bit integer operations added in ISA
7747 2.07. Unlike TImode and PTImode, the preferred register set is
7748 the Altivec/VMX registers for the 128-bit operations.
7750 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7752 (rs6000_split_128bit_ok_p): Likewise.
7754 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7755 macros for creating ISA 2.07 normal and overloaded builtin
7756 functions with 3 arguments.
7757 (BU_P8V_OVERLOAD_3): Likewise.
7758 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7759 for use as overloaded functions.
7760 (VPERM_1TI_UNS): Likewise.
7761 (VSEL_1TI): Likewise.
7762 (VSEL_1TI_UNS): Likewise.
7763 (ST_INTERNAL_1ti): Likewise.
7764 (LD_INTERNAL_1ti): Likewise.
7765 (XXSEL_1TI): Likewise.
7766 (XXSEL_1TI_UNS): Likewise.
7767 (VPERM_1TI): Likewise.
7768 (VPERM_1TI_UNS): Likewise.
7769 (XXPERMDI_1TI): Likewise.
7770 (SET_1TI): Likewise.
7771 (LXVD2X_V1TI): Likewise.
7772 (STXVD2X_V1TI): Likewise.
7773 (VEC_INIT_V1TI): Likewise.
7774 (VEC_SET_V1TI): Likewise.
7775 (VEC_EXT_V1TI): Likewise.
7776 (EQV_V1TI): Likewise.
7777 (NAND_V1TI): Likewise.
7778 (ORC_V1TI): Likewise.
7779 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7780 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7782 (VADDUQM): Likewise.
7783 (VSUBCUQ): Likewise.
7784 (VADDEUQM): Likewise.
7785 (VADDECUQ): Likewise.
7786 (VSUBEUQM): Likewise.
7787 (VSUBECUQ): Likewise.
7789 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7790 __int128_t and __uint128_t types.
7791 (__uint128_type): Likewise.
7792 (altivec_categorize_keyword): Add support for vector __int128_t,
7793 vector __uint128_t, vector __int128, and vector unsigned __int128
7794 as a container type for TImode operations that need to be done in
7795 VSX/Altivec registers.
7796 (rs6000_macro_to_expand): Likewise.
7797 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7798 to support 128-bit integer instructions vaddcuq, vadduqm,
7799 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7800 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7802 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7803 for V1TImode, and set up preferences to use VSX/Altivec registers.
7804 Setup VSX reload handlers.
7805 (rs6000_debug_reg_global): Likewise.
7806 (rs6000_init_hard_regno_mode_ok): Likewise.
7807 (rs6000_preferred_simd_mode): Likewise.
7808 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7809 (easy_altivec_constant): Likewise.
7810 (output_vec_const_move): Likewise.
7811 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7813 (rs6000_expand_vector_extract): Likewise.
7814 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7816 (rs6000_const_vec): Add support for V1TImode.
7817 (rs6000_emit_le_vsx_load): Swap double words when loading or
7818 storing TImode/V1TImode.
7819 (rs6000_emit_le_vsx_store): Likewise.
7820 (rs6000_emit_le_vsx_move): Likewise.
7821 (rs6000_emit_move): Add support for V1TImode.
7822 (altivec_expand_ld_builtin): Likewise.
7823 (altivec_expand_st_builtin): Likewise.
7824 (altivec_expand_vec_init_builtin): Likewise.
7825 (altivec_expand_builtin): Likewise.
7826 (rs6000_init_builtins): Add support for V1TImode type. Add
7827 support for ISA 2.07 128-bit integer builtins. Define type names
7828 for the VSX/Altivec vector types.
7829 (altivec_init_builtins): Add support for overloaded vector
7830 functions with V1TImode type.
7831 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7832 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7834 (rs6000_split_128bit_ok_p): Likewise.
7835 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7836 __int128_t and vector __uint128_t.
7838 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7839 and mode attributes.
7846 (VS_scalar): Likewise.
7847 (VS_double): Likewise.
7848 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7850 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7851 we support the ISA 2.07 128-bit integer arithmetic instructions.
7852 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7853 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7854 and TImode types for use with the builtin functions.
7855 (V1TI_type_node): Likewise.
7856 (unsigned_V1TI_type_node): Likewise.
7857 (intTI_type_internal_node): Likewise.
7858 (uintTI_type_internal_node): Likewise.
7860 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7861 128-bit builtin functions.
7862 (UNSPEC_VADDEUQM): Likewise.
7863 (UNSPEC_VADDECUQ): Likewise.
7864 (UNSPEC_VSUBCUQ): Likewise.
7865 (UNSPEC_VSUBEUQM): Likewise.
7866 (UNSPEC_VSUBECUQ): Likewise.
7867 (VM): Add V1TImode to vector mode iterators.
7869 (VI_unit): Likewise.
7870 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7871 (altivec_vaddcuq): Likewise.
7872 (altivec_vsubuqm): Likewise.
7873 (altivec_vsubcuq): Likewise.
7874 (altivec_vaddeuqm): Likewise.
7875 (altivec_vaddecuq): Likewise.
7876 (altivec_vsubeuqm): Likewise.
7877 (altivec_vsubecuq): Likewise.
7879 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7881 (BOOL_128): Likewise.
7882 (BOOL_REGS_OUTPUT): Likewise.
7883 (BOOL_REGS_OP1): Likewise.
7884 (BOOL_REGS_OP2): Likewise.
7885 (BOOL_REGS_UNARY): Likewise.
7886 (BOOL_REGS_AND_CR0): Likewise.
7888 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7889 128-bit integer builtin support.
7890 (vec_vadduqm): Likewise.
7891 (vec_vaddecuq): Likewise.
7892 (vec_vaddeuqm): Likewise.
7893 (vec_vsubecuq): Likewise.
7894 (vec_vsubeuqm): Likewise.
7895 (vec_vsubcuq): Likewise.
7896 (vec_vsubuqm): Likewise.
7898 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7899 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7900 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7901 128-bit integer add/subtract to ISA 2.07.
7903 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7905 * config/arc/arc.c (arc_predicate_delay_insns):
7906 Fix third argument passed to conditionalize_nonjump.
7908 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7910 * config/aarch64/aarch64-builtins.c
7911 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7912 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7913 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7914 instead of __builtin_lfloor.
7915 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7917 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7919 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7920 (tree_ssa_ifcombine_bb_1): New function.
7921 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7922 is an empty forwarder block to then_bb or vice versa and then_bb
7923 and else_bb are effectively swapped.
7925 2014-03-12 Christian Bruel <christian.bruel@st.com>
7928 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7929 REG_CFA_DEF_CFA note.
7930 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7931 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7933 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7935 PR tree-optimization/60454
7936 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7938 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7940 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7941 Do not define target_cpu_default2 to generic.
7942 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7943 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7944 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7946 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7947 Marc Glisse <marc.glisse@inria.fr>
7949 PR tree-optimization/60502
7950 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7951 instead of build_low_bits_mask.
7953 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7956 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7957 if there are multiple uses, but op doesn't live on E edge.
7958 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7959 clobber stmts before __builtin_unreachable.
7961 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7963 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7964 hard_frame_pointer_rtx.
7965 * cse.c (cse_insn): Remove volatile check.
7966 * cselib.c (cselib_process_insn): Likewise.
7967 * dse.c (scan_insn): Likewise.
7969 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7971 * config/arc/arc.c (conditionalize_nonjump): New function,
7973 (arc_ifcvt): ... this.
7974 (arc_predicate_delay_insns): Use it.
7976 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7978 * config/arc/predicates.md (extend_operand): During/after reload,
7979 allow const_int_operand.
7980 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7981 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7982 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7984 (umulsi3_highpart_i): Likewise.
7986 2014-03-11 Richard Biener <rguenther@suse.de>
7988 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7989 Add asserts to guard possible wrong-code bugs.
7991 2014-03-11 Richard Biener <rguenther@suse.de>
7993 PR tree-optimization/60429
7994 PR tree-optimization/60485
7995 * tree-ssa-structalias.c (set_union_with_increment): Properly
7996 take into account all fields that overlap the shifted vars.
7997 (do_sd_constraint): Likewise.
7998 (do_ds_constraint): Likewise.
7999 (get_constraint_for_ptr_offset): Likewise.
8001 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8003 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8004 (nios2_compute_frame_layout):
8005 Add calculation of cfun->machine->fp_save_offset.
8006 (nios2_expand_prologue): Correct setting of frame pointer register
8008 (nios2_expand_epilogue): Update recovery of stack pointer from
8009 frame pointer accordingly.
8010 (nios2_initial_elimination_offset): Update calculation of offset
8011 for eliminating to HARD_FRAME_POINTER_REGNUM.
8013 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8016 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8017 cgraph_get_create_node on VAR_DECLs.
8019 2014-03-10 Richard Biener <rguenther@suse.de>
8022 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8024 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8026 * config/vms/vms.opt (vms_float_format): New variable.
8028 2014-03-08 Tobias Burnus <burnus@net-b.de>
8030 * doc/invoke.texi (-fcilkplus): Update implementation status.
8032 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8033 Richard Biener <rguenther@suse.de>
8035 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8036 consistently accross all TUs.
8037 (run_gcc): Enable -fshort-double automatically at link at link-time
8038 and disallow override.
8040 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8043 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8044 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8045 if they can't be used.
8047 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8049 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8050 for Solaris 11/x86 ld.
8051 * configure: Regenerate.
8053 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8055 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8056 (LIB_TLS_SPEC): Save as ld_tls_libs.
8057 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8058 (HAVE_AS_IX86_TLSLDM): New test.
8059 * configure, config.in: Regenerate.
8060 * config/i386/i386.c (legitimize_tls_address): Fall back to
8061 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8062 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8063 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8064 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8066 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8068 * common.opt (fira-loop-pressure): Mark as optimization.
8070 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8072 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8073 an OpenMP mappable type.
8075 2014-03-06 Matthias Klose <doko@ubuntu.com>
8077 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8078 MULTILIB_OSDIRNAMES is not defined.
8080 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8081 Meador Inge <meadori@codesourcery.com>
8084 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8085 (arm_legitimize_address): Call legitimize_tls_address for any
8086 arm_tls_referenced_p expression, handle constant addend. Call it
8087 before testing for !TARGET_ARM.
8088 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8090 2014-03-06 Richard Biener <rguenther@suse.de>
8096 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8098 * tree-streamer.c (record_common_node): Assert we don't record
8099 nodes with type double.
8100 (preload_common_node): Skip type double, complex double and double
8101 pointer since it is now frontend dependent due to fshort-double option.
8103 2014-03-06 Richard Biener <rguenther@suse.de>
8105 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8106 or -fno-lto is specified and the linker has full plugin support.
8107 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8108 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8109 * lto-wrapper.c (merge_and_complain): Merge compile-time
8110 optimization levels.
8111 (run_gcc): And pass it through to the link options.
8113 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8117 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8119 * cselib.c (remove_useless_values): Skip to avoid quadratic
8120 behavior if the condition moved from...
8121 (cselib_process_insn): ... here holds.
8123 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8126 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8127 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8130 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8131 (TM_H): Add x86-tune.def.
8133 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8135 * config/aarch64/aarch64.c (generic_tunings):
8136 Use cortexa57_extra_costs.
8138 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8141 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8142 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8143 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8146 2014-03-04 Heiher <r@hev.cc>
8148 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8149 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8151 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8153 * config/i386/predicates.md (const2356_operand): Change to ...
8154 (const2367_operand): ... this.
8155 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8157 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8158 (*avx512pf_scatterpf<mode>sf): Ditto.
8159 (avx512pf_scatterpf<mode>df): Ditto.
8160 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8161 (*avx512pf_scatterpf<mode>df): Ditto.
8162 * config/i386/i386.c (ix86_expand_builtin): Update
8163 incorrect hint operand error message.
8165 2014-03-04 Richard Biener <rguenther@suse.de>
8167 * lto-section-in.c (lto_get_section_data): Fix const cast.
8169 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8171 * tree-streamer.c (record_common_node): Assert we don't record
8172 nodes with type double.
8173 (preload_common_node): Skip type double, complex double and double
8174 pointer since it is now frontend dependent due to fshort-double option.
8176 2014-03-04 Richard Biener <rguenther@suse.de>
8179 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8180 (lto_input_toplevel_asms): Likewise.
8181 * lto-section-in.c (lto_get_section_data): Instead do it here
8184 2014-03-04 Richard Biener <rguenther@suse.de>
8186 PR tree-optimization/60382
8187 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8188 dead PHIs a reduction.
8190 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8192 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8194 (_mm_prefetch): Move out of GCC target("sse") pragma.
8195 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8196 GCC target("prfchw") pragma.
8197 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8199 * config/i386/i386.c (ix86_option_override_internal): Enable
8200 -mprfchw with -mprefetchwt1.
8202 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8204 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8207 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8209 * opts.h (CL_PCH_IGNORE): Define.
8210 * targhooks.c (option_affects_pch_p):
8211 Return false for options that have CL_PCH_IGNORE set.
8212 * opt-functions.awk: Process PchIgnore.
8213 * doc/options.texi: Document PchIgnore.
8215 * config/arc/arc.opt (misize): Add PchIgnore property.
8217 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8219 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8220 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8221 constraint on constants to permit them being loaded into
8222 GENERAL_REGS or BASE_REGS.
8224 2014-03-03 Nick Clifton <nickc@redhat.com>
8226 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8227 anti-cacnonical alternatives.
8228 (negandhi3_real): New pattern.
8229 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8231 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8233 * config/avr/avr-mcus.def: Remove atxmega16x1.
8234 * config/avr/avr-tables.opt: Regenerate.
8235 * config/avr/t-multilib: Regenerate.
8236 * doc/avr-mmcu.texi: Regenerate.
8238 2014-03-03 Tobias Grosser <tobias@grosser.es>
8239 Mircea Namolaru <mircea.namolaru@inria.fr>
8241 PR tree-optimization/58028
8242 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8245 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8247 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8248 not handled by recognizers.
8250 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8253 * function.c (expand_function_end): Don't emit
8254 clobber_return_register sequence if clobber_after is a BARRIER.
8255 * cfgexpand.c (construct_exit_block): Append instructions before
8256 return_label to prev_bb.
8258 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8260 * config/rs6000/constraints.md: Document reserved use of "wc".
8262 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8265 * ipa.c (function_and_variable_visibility): When dissolving comdat
8266 group, also set all symbols to local.
8268 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8273 2013-12-14 Jan Hubicka <jh@suse.cz>
8275 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8277 2014-03-02 Jon Beniston <jon@beniston.com>
8283 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8284 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8285 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8286 (simple_return, *simple_return): New patterns
8287 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8288 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8290 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8292 * dwarf2out.c (gen_subprogram_die): Tidy.
8294 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8297 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8298 (*mov_t_msb_neg_negc): ... this new insn.
8300 2014-02-28 Jason Merrill <jason@redhat.com>
8303 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8306 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8309 * dwarf2out.c (decltype_auto_die): New static.
8310 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8311 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8312 (is_cxx_auto): Likewise.
8314 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8316 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8317 we are not using general regs only.
8319 2014-02-28 Richard Biener <rguenther@suse.de>
8322 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8323 previous fix and only allow to remove trivial pre-headers
8324 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8325 (remove_forwarder_block): Properly update the latch of a loop.
8327 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8330 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8331 (cselib_preserved_hash_table): New.
8332 (preserve_constants_and_equivs): Move preserved vals to it.
8333 (cselib_find_slot): Look it up first.
8334 (cselib_init): Initialize it.
8335 (cselib_finish): Release it.
8336 (dump_cselib_table): Dump it.
8338 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8341 * cselib.c (remove_useless_values): Skip to avoid quadratic
8342 behavior if the condition moved from...
8343 (cselib_process_insn): ... here holds.
8345 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8348 * var-tracking.c (vt_initialize): Apply the same condition to
8349 preserve the CFA base value.
8351 2014-02-28 Joey Ye <joey.ye@arm.com>
8354 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8355 if reload in progress or completed.
8357 2014-02-28 Tobias Burnus <burnus@net-b.de>
8360 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8363 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8365 * doc/tm.texi.in (Condition Code Status): Update documention for
8366 relative locations of cc0-setter and cc0-user.
8368 2014-02-27 Jeff Law <law@redhat.com>
8370 PR rtl-optimization/52714
8371 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8372 into two independent simple sets, if I3 is a jump, ensure the
8373 pattern we place into I3 is a (set (pc) ...).
8375 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8376 Jeff Law <law@redhat.com>
8378 PR rtl-optimization/49847
8379 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8380 are in different blocks.
8381 * doc/tm.texi (Condition Code Status): Update documention for
8382 relative locations of cc0-setter and cc0-user.
8384 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8387 * lra.c (lra_emit_add): Check SUBREG too.
8389 2014-02-27 Andreas Schwab <schwab@suse.de>
8391 * config/m68k/m68k.c (m68k_option_override): Disable
8392 -flive-range-shrinkage for classic m68k.
8393 (m68k_override_options_after_change): Likewise.
8395 2014-02-27 Marek Polacek <polacek@redhat.com>
8398 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8399 -Wmaybe-uninitialized.
8401 2014-02-27 Alan Modra <amodra@gmail.com>
8404 * reload1.c (emit_input_reload_insns): When reload_override_in,
8405 set old to rl->in_reg when rl->in_reg is a subreg.
8407 2014-02-26 Richard Biener <rguenther@suse.de>
8410 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8412 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8414 * common/config/i386/predicates.md (const1256_operand): Remove.
8415 (const2356_operand): New.
8416 (const_1_to_2_operand): Remove.
8417 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8418 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8419 (*avx512pf_gatherpf<mode>sf): Ditto.
8420 (avx512pf_gatherpf<mode>df): Ditto.
8421 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8422 (*avx512pf_gatherpf<mode>df): Ditto.
8423 (avx512pf_scatterpf<mode>sf): Ditto.
8424 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8425 (*avx512pf_scatterpf<mode>sf): Ditto.
8426 (avx512pf_scatterpf<mode>df): Ditto.
8427 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8428 (*avx512pf_scatterpf<mode>df): Ditto.
8429 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8431 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8433 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8434 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8435 (_mm512_mask_testn_epi64_mask): Move to ...
8436 * config/i386/avx512cdintrin.h: Here.
8437 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8438 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8439 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8440 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8441 TARGET_AVX512F from TARGET_AVX512CD.
8443 2014-02-26 Richard Biener <rguenther@suse.de>
8446 * ipa.c (walk_polymorphic_call_targets): Properly guard
8447 call to inline_update_overall_summary.
8449 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8452 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8453 and latches only if requested. Fix latch if it is removed.
8454 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8455 LOOPS_HAVE_PREHEADERS.
8457 2014-02-25 Andrew Pinski <apinski@cavium.com>
8459 * builtins.c (expand_builtin_thread_pointer): Create a new target
8460 when the target is NULL.
8462 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8464 PR rtl-optimization/60317
8465 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8466 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8467 * lra-assigns.c: Include params.h.
8468 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8469 other reload pseudos considerations.
8471 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8473 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8474 to use canonical form for nor<mode>3.
8476 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8479 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8482 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8484 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8485 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8486 (ix86_handle_option): Handle OPT_mprefetchwt1.
8487 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8488 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8490 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8491 OPTION_MASK_ISA_PREFETCHWT1.
8492 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8493 (PTA_PREFETCHWT1): New.
8494 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8495 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8496 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8497 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8498 (*prefetch_avx512pf_<mode>_: Change into ...
8499 (*prefetch_prefetchwt1_<mode>: This.
8500 * config/i386/i386.opt (mprefetchwt1): New.
8501 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8502 (_mm_prefetch): Handle intent to write.
8503 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8505 2014-02-25 Richard Biener <rguenther@suse.de>
8508 * emit-rtl.c (mem_attrs_htab): Remove.
8509 (mem_attrs_htab_hash): Likewise.
8510 (mem_attrs_htab_eq): Likewise.
8511 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8512 (init_emit_once): Do not allocate mem_attrs_htab.
8514 2014-02-25 Richard Biener <rguenther@suse.de>
8517 * lto-opts.c (lto_write_options): Output non-explicit conservative
8518 -fwrapv, -fno-trapv and -fno-strict-overflow.
8519 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8520 (run_gcc): And pass them through.
8522 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8524 * sel-sched.c (calculate_new_fences): New parameter ptime.
8525 Calculate it as a maximum over all fence cycles.
8526 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8527 Print the final schedule timing when sched_verbose.
8529 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8531 PR rtl-optimization/60292
8532 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8533 bit fot the fence instruction.
8535 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8537 * calls.h: Fix typo in comment.
8539 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8541 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8542 adjusting offsetable addresses.
8544 2014-02-24 Guozhi Wei <carrot@google.com>
8546 * sparseset.h (sparseset_pop): Fix the wrong index.
8548 2014-02-24 Walter Lee <walt@tilera.com>
8550 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8551 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8553 * common/config/tilegx/tilegx-common.c
8554 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8555 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8557 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8558 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8559 (tilegx_gimplify_va_arg_expr): Handle big endian.
8560 (tilegx_expand_unaligned_load): Ditto.
8561 (tilegx_expand_unaligned_store): Ditto.
8562 (TARGET_RETURN_IN_MSB): New.
8563 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8564 (TARGET_ENDIAN_DEFAULT): New.
8565 (TARGET_BIG_ENDIAN): Handle big endian.
8566 (BYTES_BIG_ENDIAN): Ditto.
8567 (WORDS_BIG_ENDIAN): Ditto.
8568 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8571 * config/tilegx/tilegx.md (extv): Handle big endian.
8573 (insn_st<n>): Ditto.
8574 (insn_st<n>_add<bitsuffix>): Ditto.
8575 (insn_stnt<n>): Ditto.
8576 (insn_stnt<n>_add<bitsuffix>):Ditto.
8577 (vec_interleave_highv8qi): Handle big endian.
8578 (vec_interleave_highv8qi_be): New.
8579 (vec_interleave_highv8qi_le): New.
8580 (insn_v1int_h): Handle big endian.
8581 (vec_interleave_lowv8qi): Handle big endian.
8582 (vec_interleave_lowv8qi_be): New.
8583 (vec_interleave_lowv8qi_le): New.
8584 (insn_v1int_l): Handle big endian.
8585 (vec_interleave_highv4hi): Handle big endian.
8586 (vec_interleave_highv4hi_be): New.
8587 (vec_interleave_highv4hi_le): New.
8588 (insn_v2int_h): Handle big endian.
8589 (vec_interleave_lowv4hi): Handle big endian.
8590 (vec_interleave_lowv4hi_be): New.
8591 (vec_interleave_lowv4hi_le): New.
8592 (insn_v2int_l): Handle big endian.
8593 (vec_interleave_highv2si): Handle big endian.
8594 (vec_interleave_highv2si_be): New.
8595 (vec_interleave_highv2si_le): New.
8596 (insn_v4int_h): Handle big endian.
8597 (vec_interleave_lowv2si): Handle big endian.
8598 (vec_interleave_lowv2si_be): New.
8599 (vec_interleave_lowv2si_le): New.
8600 (insn_v4int_l): Handle big endian.
8601 * config/tilegx/tilegx.opt (mbig-endian): New option.
8602 (mlittle-endian): New option.
8603 * doc/install.texi: Document tilegxbe-linux.
8604 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8606 2014-02-24 Martin Jambor <mjambor@suse.cz>
8609 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8610 there are no parameter descriptors.
8612 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8614 PR rtl-optimization/60268
8615 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8616 initialization to ...
8617 (sched_rgn_init): ... here.
8618 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8620 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8622 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8625 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8627 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8630 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8632 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8633 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8635 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8637 * config/microblaze/predicates.md: Add cmp_op predicate.
8638 * config/microblaze/microblaze.md: Add branch_compare instruction
8639 which uses cmp_op predicate and emits cmp insn before branch.
8640 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8641 to microblaze_expand_conditional_branch and consolidate logic.
8642 (microblaze_expand_conditional_branch): emit branch_compare
8643 insn instead of handling cmp op separate from branch insn.
8645 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8647 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8650 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8652 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8653 define_insn with define_expand and new define_insn
8654 *altivec_lve<VI_char>x_internal.
8655 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8656 and new define_insn *altivec_stve<VI_char>x_internal.
8657 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8659 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8661 (altivec_expand_stvex_be): New function.
8663 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8665 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8666 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8667 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8668 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8670 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8673 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8674 instead of emit_move_insn.
8676 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8678 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8680 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8681 gen_altivec_vsumsws.
8683 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8685 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8686 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8687 (altivec_lvxl_<mode>): New define_expand incorporating
8688 -maltivec=be semantics where needed.
8689 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8690 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8691 semantics where needed.
8692 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8693 (altivec_stvx_<mode>): New define_expand incorporating
8694 -maltivec=be semantics where needed.
8695 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8696 VM2 iterator instead of V4SI.
8697 (altivec_stvxl_<mode>): New define_expand incorporating
8698 -maltivec=be semantics where needed.
8699 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8700 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8701 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8702 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8703 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8704 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8705 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8706 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8707 ALTIVEC_BUILTIN_STVXL.
8708 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8709 (altivec_expand_stvx_be): Likewise.
8710 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8711 (altivec_expand_lvx_be): Likewise.
8712 (altivec_expand_stvx_be): Likewise.
8713 (altivec_expand_builtin): Add cases for
8714 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8715 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8716 (altivec_init_builtins): Add definitions for
8717 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8718 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8720 2014-02-21 Catherine Moore <clm@codesourcery.com>
8722 * doc/invoke.texi (mvirt, mno-virt): Document.
8723 * config/mips/mips.opt (mvirt): New option.
8724 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8726 2014-02-21 Richard Biener <rguenther@suse.de>
8728 PR tree-optimization/60276
8729 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8730 (STMT_VINFO_MIN_NEG_DIST): New macro.
8731 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8732 STMT_VINFO_MIN_NEG_DIST.
8733 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8734 made for negative dependence distances still hold.
8736 2014-02-21 Richard Biener <rguenther@suse.de>
8739 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8740 DECL_INITIAL for globals not in the current function context.
8742 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8744 PR tree-optimization/56490
8745 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8746 * tree-ssa-uninit.c: Include params.h.
8747 (compute_control_dep_chain): Add num_calls argument, return false
8748 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8749 num_calls to recursive call.
8750 (find_predicates): Change dep_chain into normal array,
8751 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8752 variable and adjust compute_control_dep_chain caller.
8753 (find_def_preds): Likewise.
8755 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8757 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8758 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8760 2014-02-21 Nick Clifton <nickc@redhat.com>
8762 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8763 (pushhi1): Likewise.
8764 (popqi1): Add mode to pre_dec.
8767 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8769 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8770 mode for mask of V8SFmode permutation.
8772 2014-02-20 Richard Henderson <rth@redhat.com>
8775 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8776 a new pseudo for OLDVAL.
8778 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8781 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8782 gen_reg_rtx if d->testing_p.
8783 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8784 if d->testing_p and we will certainly return true.
8785 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8788 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8790 * emit-rtl.c (gen_reg_rtx): Assert that
8791 crtl->emit.regno_pointer_align_length is non-zero.
8793 2014-02-20 Richard Henderson <rth@redhat.com>
8796 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8797 on failure the store back into EXPECT.
8799 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8800 Sandra Loosemore <sandra@codesourcery.com>
8802 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8803 * config/nios2/nios2.c (nios2_function_profiler): Add
8804 -fPIC (flag_pic == 2) support.
8805 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8806 (nios2_large_offset_p): New function.
8807 (nios2_unspec_reloc_p): Move up position, update to use
8808 nios2_large_offset_p.
8809 (nios2_unspec_address): Remove function.
8810 (nios2_unspec_offset): New function.
8811 (nios2_large_got_address): New function.
8812 (nios2_got_address): Add large offset support.
8813 (nios2_legitimize_tls_address): Update usage of removed and new
8815 (nios2_symbol_binds_local_p): New function.
8816 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8817 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8818 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8819 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8820 processing for (const (unspec ...)).
8821 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8823 2014-02-20 Richard Biener <rguenther@suse.de>
8825 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8826 doing the substitution.
8827 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8829 2014-02-20 Martin Jambor <mjambor@suse.cz>
8832 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8833 info when checking whether lattices are bottom.
8835 2014-02-20 Richard Biener <rguenther@suse.de>
8838 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8841 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8844 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8845 parameter specifying the scaling.
8846 (inline_call): Update.
8847 (want_inline_recursively): Guard division by zero.
8848 (recursive_inlining): Update.
8849 * ipa-inline.h (clone_inlined_nodes): Update.
8851 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8854 * config/i386/i386.c (classify_argument): Pass structures of size
8855 64 bytes or less in register.
8857 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8858 Kirill Yukhin <kirill.yukhin@intel.com>
8860 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8861 (_mm_rcp28_round_ss): Ditto.
8862 (_mm_rsqrt28_round_sd): Ditto.
8863 (_mm_rsqrt28_round_ss): Ditto.
8864 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8865 (_mm_rcp14_round_ss): Ditto.
8866 (_mm_rsqrt14_round_sd): Ditto.
8867 (_mm_rsqrt14_round_ss): Ditto.
8868 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8869 the first input operand, get rid of match_dup.
8870 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8872 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8874 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8875 operand as the first input operand, set type attribute.
8876 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8878 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8879 operand as the first input operand, set type attribute.
8881 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8883 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8886 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8889 * config/i386/i386.c (construct_container): Remove TFmode check
8890 for X86_64_INTEGER_CLASS.
8892 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8895 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8896 only when -Wpsabi is enabled.
8898 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8901 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8902 passing arrays in registers are the same as for structs, so remove the
8903 special case for them.
8905 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8907 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8908 destination type, extract only the valid bits if the source type is not
8909 integral and has a different mode.
8911 2014-02-19 Richard Biener <rguenther@suse.de>
8914 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8917 2014-02-19 Richard Biener <rguenther@suse.de>
8920 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8921 (ipa_modify_call_arguments): Emit an argument load explicitely and
8922 preserve virtual SSA form there and for the replacement call.
8923 Do not update SSA form nor free dominance info.
8925 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8927 * ipa.c (function_and_variable_visibility): Also clear WEAK
8928 flag when disolving COMDAT_GROUP.
8930 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8932 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8933 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8935 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8936 do more sanity checks.
8937 (detect_type_change): Return true when giving up early.
8938 (compute_complex_assign_jump_func): Fix type parameter of
8939 ipa_set_ancestor_jf.
8940 (compute_complex_ancestor_jump_func): Likewise.
8941 (update_jump_functions_after_inlining): Fix updating of
8943 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8945 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8947 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8948 inline clones when edge disappears.
8950 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8953 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8954 Split 64-bit moves into 2 patterns. Do not allow the use of
8955 direct move for TDmode in little endian, since the decimal value
8956 has little endian bytes within a word, but the 64-bit pieces are
8957 ordered in a big endian fashion, and normal subreg's of TDmode are
8959 (mov<mode>_64bit_dm): Likewise.
8960 (movtd_64bit_nodm): Likewise.
8962 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8964 PR tree-optimization/60174
8965 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8966 statement of an SSA_NAME that occurs in an abnormal PHI node.
8968 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8971 * final.c (SEEN_BB): Remove.
8972 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8973 (final_scan_insn): Don't force_source_line on second
8974 NOTE_INSN_BASIC_BLOCK.
8976 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8979 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8980 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8981 (type_natural_mode): Warn ABI change when %zmm register is not
8982 available for AVX512F vector value passing.
8984 2014-02-18 Kai Tietz <ktietz@redhat.com>
8987 * config/i386/i386.c (ix86_expand_prologue): Use value in
8988 rax register as displacement when restoring %r10 or %rax.
8989 Fix wrong offset when restoring both registers.
8991 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8993 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8994 assertion with conditional return.
8996 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8997 Uros Bizjak <ubizjak@gmail.com>
9000 * config/i386/driver-i386.c (host_detect_local_cpu): If
9001 YMM state is not saved by the OS, also clear has_f16c. Move
9002 CPUID 0x80000001 handling before YMM state saving checking.
9004 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9006 PR rtl-optimization/58960
9007 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9008 factored out from ...
9009 (sched_init): ... here.
9010 (free_global_sched_pressure_data): New, factored out from ...
9011 (sched_finish): ... here.
9012 * sched-int.h (free_global_sched_pressure_data): Declare.
9013 * sched-rgn.c (nr_regions_initial): New static global.
9014 (haifa_find_rgns): Initialize it.
9015 (schedule_region): Disable sched-pressure for the newly
9018 2014-02-17 Richard Biener <rguenther@suse.de>
9020 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9021 release SSA defs of pattern stmts.
9023 2014-02-17 Richard Biener <rguenther@suse.de>
9025 * tree-inline.c (expand_call_inline): Release the virtual
9026 operand defined by the call we are about to inline.
9028 2014-02-17 Richard Biener <rguenther@suse.de>
9030 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9032 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9033 Ilya Tocar <ilya.tocar@intel.com>
9035 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9036 arguments order in builtin.
9037 (_mm512_permutexvar_epi64): Ditto.
9038 (_mm512_mask_permutexvar_epi64): Ditto
9039 (_mm512_maskz_permutexvar_epi32): Ditto
9040 (_mm512_permutexvar_epi32): Ditto
9041 (_mm512_mask_permutexvar_epi32): Ditto
9043 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9045 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9046 (p8_vmrgow): Likewise.
9048 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9050 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9053 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9056 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9057 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9058 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9059 using direct move instructions on ISA 2.07. Also adjust
9060 instruction length for 64-bit.
9061 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9062 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9064 2014-02-15 Alan Modra <amodra@gmail.com>
9068 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9069 find_replacement on parts of insn rtl that might be reloaded.
9071 2014-02-15 Richard Biener <rguenther@suse.de>
9073 PR tree-optimization/60183
9074 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9075 (tree_ssa_phiprop): Calculate and free post-dominators.
9077 2014-02-14 Jeff Law <law@redhat.com>
9079 PR rtl-optimization/60131
9080 * ree.c (get_extended_src_reg): New function.
9081 (combine_reaching_defs): Use it rather than assuming location of REG.
9082 (find_and_remove_re): Verify first operand of extension is
9083 a REG before adding the insns to the copy list.
9085 2014-02-14 Roland McGrath <mcgrathr@google.com>
9087 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9088 * configure: Regenerated.
9089 * config.in: Regenerated.
9090 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9091 instead of ASM_SHORT.
9093 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9094 Richard Earnshaw <rearnsha@arm.com>
9096 PR rtl-optimization/59535
9097 * lra-constraints.c (process_alt_operands): Encourage alternative
9098 when unassigned pseudo class is superset of the alternative class.
9099 (inherit_reload_reg): Don't inherit when optimizing for code size.
9100 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9101 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9102 modes not less than 4 for Thumb1.
9104 2014-02-14 Kyle McMartin <kyle@redhat.com>
9107 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9109 2014-02-14 Richard Biener <rguenther@suse.de>
9111 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9112 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9113 Do not strip INDIRECT_REFs.
9115 2014-02-14 Richard Biener <rguenther@suse.de>
9118 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9119 DECL_FUNCTION_SPECIFIC_TARGET.
9120 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9121 * tree-streamer-out.c (pack_ts_target_option): Remove.
9122 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9123 (write_ts_function_decl_tree_pointers): Do not stream
9124 DECL_FUNCTION_SPECIFIC_TARGET.
9125 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9126 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9127 (lto_input_ts_function_decl_tree_pointers): Do not stream
9128 DECL_FUNCTION_SPECIFIC_TARGET.
9130 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9132 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9133 (get_initial_def_for_induction, vectorizable_induction): Ignore
9134 debug stmts when looking for exit_phi.
9135 (vectorizable_live_operation): Fix up condition.
9137 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9139 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9140 nreverse() because it changes the content of original tree list.
9142 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9144 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9145 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9147 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9149 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9150 GNU coding standards.
9152 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9155 * dwarf2out.c (gen_subprogram_die): Don't call
9156 add_calling_convention_attribute if subr_die is old_die.
9158 2014-02-13 Sharad Singhai <singhai@google.com>
9160 * doc/optinfo.texi: Fix order of nodes.
9162 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9164 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9165 operands[2], not operands[3].
9167 2014-02-13 Richard Biener <rguenther@suse.de>
9170 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9171 mention the possibility of an in-tree build.
9172 (CLooG): Update recommended version to 0.18.1, mention the
9173 possibility of an in-tree build and clarify that the ISL
9174 bundled with CLooG does not work.
9176 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9179 * expr.c (compress_float_constant): If x is a hard register,
9180 extend into a pseudo and then move to x.
9182 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9184 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9185 caused by bad second argument to warning_at() with -mhotpatch and
9186 nested functions (e.g. with gfortran).
9188 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9190 * opts.c (option_name): Remove "enabled by default" rider.
9192 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9194 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9196 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9197 Uros Bizjak <ubizjak@gmail.com>
9200 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9201 * configure: Regenerated.
9203 2014-02-12 Richard Biener <rguenther@suse.de>
9205 * vec.c (vec_prefix::calculate_allocation): Move as
9206 inline variant to vec.h.
9207 (vec_prefix::calculate_allocation_1): New out-of-line version.
9208 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9209 (vec_prefix::m_has_auto_buf): Rename to ...
9210 (vec_prefix::m_using_auto_storage): ... this.
9211 (vec_prefix::calculate_allocation): Inline the easy cases
9212 and dispatch to calculate_allocation_1 which doesn't need the
9214 (va_heap::reserve): Use gcc_checking_assert.
9215 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9216 m_using_auto_storage.
9217 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9219 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9220 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9221 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9223 2014-02-12 Richard Biener <rguenther@suse.de>
9225 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9226 when we found a dependence.
9228 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9230 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9232 (maybe_fold_stmt): ... into this new function.
9233 * omp-low.c (lower_omp): Update comment.
9235 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9238 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9241 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9243 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9244 identifiers in comments.
9245 (cortexa53_extra_costs): Likewise.
9246 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9247 (cortexa7_extra_costs): Likewise.
9248 (cortexa12_extra_costs): Likewise.
9249 (cortexa15_extra_costs): Likewise.
9250 (v7m_extra_costs): Likewise.
9252 2014-02-12 Richard Biener <rguenther@suse.de>
9255 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9256 of posix_memalign being successful.
9257 (lower_stmt): Restrict lowering of posix_memalign to when
9258 -ftree-bit-ccp is enabled.
9260 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9262 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9264 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9266 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9268 PR rtl-optimization/60116
9269 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9270 other_insn once the combination has been validated.
9272 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9275 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9277 * ipa-devirt.c: Include demangle.h
9278 (odr_violation_reported): New static variable.
9279 (add_type_duplicate): Update odr_violations.
9280 (maybe_record_node): Add completep parameter; update it.
9281 (record_target_from_binfo): Add COMPLETEP parameter;
9282 update it as needed.
9283 (possible_polymorphic_call_targets_1): Likewise.
9284 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9285 rename FINAL to COMPLETE.
9286 (record_targets_from_bases): Sanity check we found the binfo;
9287 fix COMPLETEP updating.
9288 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9289 parameter, fix computing of COMPLETEP.
9290 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9291 at LTO time do demangling.
9292 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9293 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9295 (gimple_get_virt_method_for_binfo): Likewise.
9296 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9297 gimple_get_virt_method_for_vtable): Update prototypes.
9299 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9302 * genautomata.c (add_presence_absence): Fix typo with
9303 {final_}presence_list.
9305 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9308 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9309 for VSX/Altivec vectors that land in GPR registers.
9311 2014-02-11 Richard Henderson <rth@redhat.com>
9312 Jakub Jelinek <jakub@redhat.com>
9315 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9316 around drhs if type conversion to lacc->type is not useless.
9318 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9320 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9322 (cortex-a57.cortex-a53): Likewise.
9323 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9325 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9327 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9330 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9332 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9333 add_options_for_arm_vfp3.
9335 2014-02-11 Jeff Law <law@redhat.com>
9338 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9339 object with an undesirable mode.
9341 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9344 * config/i386/sol2-9.h: New file.
9345 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9346 *-*-solaris2.9*): Use it.
9348 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9350 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9351 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9353 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9355 * config/microblaze/microblaze.c: Extend mcpu version format
9357 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9359 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9361 2014-02-10 Richard Henderson <rth@redhat.com>
9364 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9365 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9366 ms-abi vs -mno-accumulate-outgoing-args.
9367 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9368 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9371 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9374 * cfgexpand.c (expand_asm_operands): Attach source location to
9375 ASM_INPUT rtx objects.
9376 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9378 2014-02-10 Nick Clifton <nickc@redhat.com>
9380 * config/mn10300/mn10300.c (popcount): New function.
9381 (mn10300_expand_prologue): Include saved registers in stack usage
9384 2014-02-10 Jeff Law <law@redhat.com>
9387 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9388 when changing the SET_DEST of a prior insn to avoid an input reload.
9390 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9392 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9393 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9394 -mcall-openbsd, or -mcall-linux.
9395 (CC1_ENDIAN_BIG_SPEC): Remove.
9396 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9397 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9398 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9399 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9400 and %cc1_endian_default.
9401 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9403 2014-02-10 Richard Biener <rguenther@suse.de>
9405 PR tree-optimization/60115
9406 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9407 MEM_REF handling. Properly verify that the accesses are not
9408 out of the objects bound.
9410 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9412 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9415 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9417 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9418 proper constants and fix formatting.
9419 (possible_polymorphic_call_targets): Fix formatting.
9421 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9422 Ilya Tocar <ilya.tocar@intel.com>
9424 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9425 (_mm512_loadu_epi32): Renamed into...
9426 (_mm512_loadu_si512): This.
9427 (_mm512_storeu_epi32): Renamed into...
9428 (_mm512_storeu_si512): This.
9429 (_mm512_maskz_ceil_ps): Removed.
9430 (_mm512_maskz_ceil_pd): Ditto.
9431 (_mm512_maskz_floor_ps): Ditto.
9432 (_mm512_maskz_floor_pd): Ditto.
9433 (_mm512_floor_round_ps): Ditto.
9434 (_mm512_floor_round_pd): Ditto.
9435 (_mm512_ceil_round_ps): Ditto.
9436 (_mm512_ceil_round_pd): Ditto.
9437 (_mm512_mask_floor_round_ps): Ditto.
9438 (_mm512_mask_floor_round_pd): Ditto.
9439 (_mm512_mask_ceil_round_ps): Ditto.
9440 (_mm512_mask_ceil_round_pd): Ditto.
9441 (_mm512_maskz_floor_round_ps): Ditto.
9442 (_mm512_maskz_floor_round_pd): Ditto.
9443 (_mm512_maskz_ceil_round_ps): Ditto.
9444 (_mm512_maskz_ceil_round_pd): Ditto.
9445 (_mm512_expand_pd): Ditto.
9446 (_mm512_expand_ps): Ditto.
9447 * config/i386/i386.c (ix86_builtins): Remove
9448 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9449 (bdesc_args): Ditto.
9450 * config/i386/predicates.md (const1256_operand): New.
9451 (const_1_to_2_operand): Ditto.
9452 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9453 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9454 (*avx512pf_gatherpf<mode>sf): Ditto.
9455 (avx512pf_gatherpf<mode>df): Ditto.
9456 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9457 (*avx512pf_gatherpf<mode>df): Ditto.
9458 (avx512pf_scatterpf<mode>sf): Ditto.
9459 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9460 (*avx512pf_scatterpf<mode>sf): Ditto.
9461 (avx512pf_scatterpf<mode>df): Ditto.
9462 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9463 (*avx512pf_scatterpf<mode>df): Ditto.
9464 (avx512f_expand<mode>): Removed.
9465 (<shift_insn><mode>3<mask_name>): Change predicate type.
9467 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9469 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9470 not at the end of datarefs vector use ordered_remove to avoid
9471 reordering datarefs vector.
9474 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9475 mark local addressable non-static vars as GOVD_PRIVATE
9476 instead of GOVD_LOCAL.
9477 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9478 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9482 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9483 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9484 assume_aligned or alloc_align attributes.
9485 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9486 arguments. Handle also assume_aligned and alloc_align attributes.
9487 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9488 calls to functions with assume_aligned or alloc_align attributes.
9489 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9491 2014-02-08 Terry Guo <terry.guo@arm.com>
9493 * doc/invoke.texi: Document ARM -march=armv7e-m.
9495 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9497 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9498 flag on __cilkrts_rethrow builtin.
9501 * ipa-cp.c (determine_versionability): Fail at -O0
9502 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9503 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9506 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9509 * tree-inline.c (copy_forbidden): Fail for
9510 __attribute__((optimize (0))) functions.
9512 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9514 * varpool.c: Include pointer-set.h.
9515 (varpool_remove_unreferenced_decls): Variables in other partitions
9516 will not be output; be however careful to not lose information
9519 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9521 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9522 lookup in the vtable constructor.
9524 2014-02-07 Jeff Law <law@redhat.com>
9527 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9528 define_insn_and_split.
9530 * ipa-inline.c (inline_small_functions): Fix typos.
9532 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9534 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9535 (s390_can_use_return_insn): Declare.
9536 * config/s390/s390.h (EPILOGUE_USES): Define.
9537 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9539 (s390_chunkify_start): Handle return JUMP_LABELs.
9540 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9541 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9542 (s390_can_use_return_insn): New functions.
9543 (s390_fix_long_loop_prediction): Handle conditional returns.
9544 (TARGET_SET_UP_BY_PROLOGUE): Define.
9545 * config/s390/s390.md (ANY_RETURN): New code iterator.
9546 (*creturn, *csimple_return, return, simple_return): New patterns.
9548 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9550 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9551 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9552 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9553 REG_CFA_RESTORE list when deciding not to restore a register.
9555 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9557 * config/s390/s390.c: Include tree-pass.h and context.h.
9558 (s390_early_mach): New function, split out from...
9559 (s390_emit_prologue): ...here.
9560 (pass_data_s390_early_mach): New pass structure.
9561 (pass_s390_early_mach): New class.
9562 (s390_option_override): Create and register early_mach pass.
9563 Move to end of file.
9565 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9567 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9568 to match for the exit block.
9570 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9572 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9573 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9574 Reject misaligned operands.
9576 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9578 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9580 2014-02-07 Richard Biener <rguenther@suse.de>
9583 * gimple-low.c (lower_builtin_posix_memalign): New function.
9584 (lower_stmt): Call it to lower posix_memalign in a way
9585 to make alignment info accessible.
9587 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9590 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9591 __builtin_setjmp_receiver.
9593 2014-02-07 Richard Biener <rguenther@suse.de>
9596 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9597 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9598 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9599 Handle BUILT_IN_POSIX_MEMALIGN.
9600 (find_func_clobbers): Likewise.
9601 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9602 (call_may_clobber_ref_p_1): Likewise.
9604 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9607 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9610 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9613 * lto-cgraph.c (lto_output_node): Use
9614 symtab_get_symbol_partitioning_class.
9615 (lto_output_varpool_node): likewise.
9616 (symtab_get_symbol_partitioning_class): Move here from
9618 * cgraph.h (symbol_partitioning_class): Likewise.
9619 (symtab_get_symbol_partitioning_class): Declare.
9621 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9623 * ggc.h (ggc_internal_cleared_alloc): New macro.
9624 * vec.h (vec_safe_copy): Handle memory stats.
9625 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9626 * target-globals.c (save_target_globals): Likewise.
9628 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9631 * expr.c (emit_move_resolve_push): Export; be bit more selective
9632 on when to clear alias set.
9633 * expr.h (emit_move_resolve_push): Declare.
9634 * function.h (struct function): Add tail_call_marked.
9635 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9636 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9637 * config/i386/i386.md (TImode move expander): De not call
9639 (FP push expanders): Preserve memory attributes.
9640 * config/i386/sse.md (push<mode>1): Remove.
9641 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9642 (ix86_expand_push): Remove.
9643 * config/i386/mmx.md (push<mode>1): Remove.
9645 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9647 PR rtl-optimization/60030
9648 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9649 lopart with paradoxical subreg before shifting it up by hprec.
9651 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9653 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9654 Remove extra newline at end of file.
9655 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9656 (arm_issue_rate): Handle cortexa57.
9657 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9658 (cortex-a57.cortex-a53): Likewise.
9660 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9663 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9664 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9666 (arm_expand_prologue): Adjust all callers.
9667 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9668 info, registers also at the lowest numbered registers side. Use
9669 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9673 * var-tracking.c (adjust_mems): Before adding a SET to
9674 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9676 2014-02-06 Alan Modra <amodra@gmail.com>
9679 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9680 change SDmode to DDmode when lra_in_progress.
9682 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9685 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9686 free_data_ref on the dr first, and before goto again also set dr
9687 to the next dr. For simd_lane_access, free old datarefs[i] before
9688 overwriting it. For get_vectype_for_scalar_type failure, don't
9689 free_data_ref if simd_lane_access.
9691 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9694 * tree.h (opts_for_fn): New inline function.
9695 (opt_for_fn): Define.
9696 * config/i386/i386.c (ix86_function_regparm): Use
9697 opt_for_fn (decl, optimize) instead of optimize.
9699 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9701 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9702 for SYMBOL_REF in large memory model.
9704 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9706 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9708 (cortex-a57): Likewise.
9709 (cortex-a57.cortex-a53): Likewise.
9711 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9712 Kugan Vivekanandarajah <kuganv@linaro.org>
9714 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9716 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9718 2014-02-06 Richard Biener <rguenther@suse.de>
9720 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9721 set_loop_copy and initialize_original_copy_tables.
9723 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9725 * config/aarch64/aarch64-simd.md
9726 (aarch64_ashr_simddi): Change QI to SI.
9728 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9729 Jakub Jelinek <jakub@redhat.com>
9732 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9735 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9737 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9738 CODE_FOR_altivec_vpku[hw]um to
9739 CODE_FOR_altivec_vpku[hw]um_direct.
9740 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9741 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9742 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9743 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9745 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9747 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9748 generation for -maltivec=be.
9749 (altivec_vsumsws): Simplify redundant test.
9751 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9753 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9754 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9755 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9756 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9757 gen_altivec_vpkuwum.
9758 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9760 (altivec_vpks<VI_char>ss): Likewise.
9761 (altivec_vpks<VI_char>us): Likewise.
9762 (altivec_vpku<VI_char>us): Likewise.
9763 (altivec_vpku<VI_char>um): Likewise.
9764 (altivec_vpku<VI_char>um_direct): New (copy of
9765 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9767 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9768 target is little endian and -maltivec=be is not specified.
9769 (*altivec_vupkhs<VU_char>_direct): New (copy of
9770 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9771 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9772 target is little endian and -maltivec=be is not specified.
9773 (*altivec_vupkls<VU_char>_direct): New (copy of
9774 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9775 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9776 little endian and -maltivec=be is not specified.
9777 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9778 little endian and -maltivec=be is not specified.
9780 2014-02-05 Richard Henderson <rth@redhat.com>
9783 * combine-stack-adj.c: Revert r206943.
9784 * sched-int.h (struct deps_desc): Add last_args_size.
9785 * sched-deps.c (init_deps): Initialize it.
9786 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9787 contain REG_ARGS_SIZE notes.
9789 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9791 * lto-cgraph.c (asm_nodes_output): Make global.
9792 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9793 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9794 (driver_handle_option): Handle OPT_fwpa.
9796 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9799 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9800 a comment typo and formatting issue. If odr_hash hasn't been
9801 created, return vNULL and set *completep to false.
9804 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9805 bb with no successors.
9807 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9810 * doc/invoke.texi (-march): Clarify documentation for ARM.
9814 2014-02-05 Richard Biener <rguenther@suse.de>
9816 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9817 when not vectorizing because of too many alias checks.
9818 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9819 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9821 2014-02-05 Nick Clifton <nickc@redhat.com>
9823 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9824 accept extended registers in any mode when compiling for the MN10300.
9826 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9828 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9829 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9830 sanitization attributes.
9831 (can_inline_edge_p): Likewise.
9832 (sanitize_attrs_match_for_inline_p): New function.
9834 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9836 * ipa-prop.c (detect_type_change): Shor circuit testing of
9837 type changes on THIS pointer.
9839 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9842 * config/pa/pa.c (legitimize_tls_address): Return original address
9843 if not passed a SYMBOL_REF rtx.
9844 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9846 (pa_emit_move_sequence): Simplify TLS source operands.
9847 (pa_legitimate_constant_p): Reject all TLS constants.
9848 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9849 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9851 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9853 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9854 groups when we know they are controlled by LTO.
9855 * varasm.c (default_binds_local_p_1): If object is in other partition,
9856 it will be resolved locally.
9858 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9860 * config/host-linux.c (linux_gt_pch_use_address): Don't
9861 use SSIZE_MAX because it is not always defined.
9863 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9866 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9867 threshold for pseudo splitting.
9868 (update_ebb_live_info): Process call argument hard registers and
9869 hard registers from insn definition too.
9870 (max_small_class_regs_num): New constant.
9871 (inherit_in_ebb): Update live hard regs through EBBs. Update
9872 reloads_num only for small register classes. Don't split for
9875 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9878 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9881 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9883 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9886 2014-02-04 Marek Polacek <polacek@redhat.com>
9888 * gdbinit.in (pel): Define.
9890 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9892 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9895 2014-02-04 Richard Biener <rguenther@suse.de>
9898 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9899 in function context local.
9900 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9901 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9902 similar to LTO_imported_decl_ref.
9904 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9906 PR tree-optimization/60002
9907 * cgraphclones.c (build_function_decl_skip_args): Clear
9910 PR tree-optimization/60023
9911 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9912 false to gsi_replace.
9913 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9914 has been in some EH region and vec_stmt could throw, add
9915 vec_stmt into the same EH region.
9916 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9917 has no lhs, ignore it.
9918 * internal-fn.c (expand_MASK_LOAD): Likewise.
9921 * tree-inline.c (copy_forbidden): Fail for
9922 __attribute__((optimize (0))) functions.
9925 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9926 is set, copy one less argument.
9927 (expand_simd_clones): Don't subtract clone_info->inbranch
9928 from simd_clone_struct_alloc argument.
9930 PR rtl-optimization/57915
9931 * recog.c (simplify_while_replacing): If all unary/binary/relational
9932 operation arguments are constant, attempt to simplify those.
9935 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9936 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9938 2014-02-04 Richard Biener <rguenther@suse.de>
9940 PR tree-optimization/60012
9941 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9942 TBAA disambiguation to all DDRs.
9944 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9947 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9948 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9950 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9953 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9955 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9957 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9958 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9960 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9963 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9964 to figure out targets of polymorphic calls with known decl.
9965 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9966 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9967 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9968 (get_polymorphic_call_info): ... here.
9969 (get_polymorphic_call_info_from_invariant): New function.
9971 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9973 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9974 lookup via vtable pointer; check for type consistency
9975 and turn inconsitent facts into UNREACHABLE.
9976 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9977 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9978 type inconsistent querries; return UNREACHABLE instead.
9980 2014-02-03 Richard Henderson <rth@twiddle.net>
9983 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9984 already processed this node.
9985 (normalize_one_pred_1): Pass along mark_set.
9986 (normalize_one_pred): Create and destroy a pointer_set_t.
9987 (normalize_one_pred_chain): Likewise.
9989 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9991 PR gcov-profile/58602
9992 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9994 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9997 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9998 -fno-devirtualize; try to devirtualize by the knowledge of
9999 virtual table pointer given by aggregate propagation.
10000 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10001 (ipa_print_node_jump_functions): Dump also offset that
10002 is relevant for polymorphic calls.
10003 (determine_known_aggregate_parts): Add arg_type parameter; use it
10004 instead of determining the type from pointer type.
10005 (ipa_compute_jump_functions_for_edge): Update call of
10006 determine_known_aggregate_parts.
10007 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10008 (gimple_get_virt_method_for_binfo): ... here; simplify using
10009 vtable_pointer_value_to_vtable.
10010 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10011 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10012 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10013 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10015 (vtable_pointer_value_to_binfo): ... here.
10016 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10018 2014-02-03 Teresa Johnson <tejohnson@google.com>
10020 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10021 redef of outer loop index variable.
10023 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10027 * doc/extend.texi (Function Attributes): Typo.
10029 2014-02-03 Cong Hou <congh@google.com>
10031 PR tree-optimization/60000
10032 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10033 if the vectorized statement is a store. A store statement can only
10034 appear at the end of pattern statements.
10036 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10038 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10039 (ix86_option_override_internal): Default long double to 64-bit for
10040 32-bit Bionic and to 128-bit for 64-bit Bionic.
10042 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10043 TARGET_LONG_DOUBLE_128 is true.
10044 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10046 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10047 (mlong-double-64): Negate -mlong-double-128.
10048 (mlong-double-128): New option.
10050 * config/i386/i386-c.c (ix86_target_macros): Define
10051 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10053 * doc/invoke.texi: Document -mlong-double-128.
10055 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10057 PR rtl-optimization/60024
10058 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10060 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10062 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10064 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10066 PR rtl-optimization/57662
10067 * sel-sched.c (code_motion_path_driver): Do not mark already not
10068 existing blocks in the visiting bitmap.
10070 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10072 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10073 on the insn being emitted.
10075 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10076 Will Deacon <will.deacon@arm.com>
10078 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10080 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10082 * config/arm/arm-tables.opt: Regenerate.
10084 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10086 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10087 for vector types other than V16QImode.
10088 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10089 define_expand, and call altivec_expand_vec_perm_le when producing
10090 code with little endian element order.
10091 (*altivec_vperm_<mode>_internal): New insn having previous
10092 behavior of altivec_vperm_<mode>.
10093 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10094 altivec_expand_vec_perm_le when producing code with little endian
10096 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10097 behavior of altivec_vperm_<mode>_uns.
10099 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10101 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10102 (altivec_vsumsws): Add handling for -maltivec=be with a little
10104 (altivec_vsumsws_direct): New.
10105 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10106 gen_altivec_vsumsws.
10108 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10110 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10111 vtable_pointer_value_to_binfo): New functions.
10112 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10113 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10115 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10117 * config/nios2/nios2.md (load_got_register): Initialize GOT
10118 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10119 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10121 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10123 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10124 preserverd by passthrough, do not propagate the type.
10126 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10128 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10129 (mips_atomic_assign_expand_fenv): New function.
10130 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10132 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10134 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10135 (__builtin_mips_set_fcsr): Likewise.
10136 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10137 MIPS_USI_FTYPE_VOID.
10138 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10139 (mips16_expand_set_fcsr): Likewise.
10140 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10141 (mips16_set_fcsr_stub): Likewise.
10142 (mips16_get_fcsr_one_only_stub): New class.
10143 (mips16_set_fcsr_one_only_stub): Likewise.
10144 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10145 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10146 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10147 (hard_float): New availability predicate.
10148 (mips_builtins): Add get_fcsr and set_fcsr.
10149 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10150 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10151 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10152 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10153 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10156 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10158 * config/mips/mips.c (mips_one_only_stub): New class.
10159 (mips_need_mips16_rdhwr_p): Replace with...
10160 (mips16_rdhwr_stub): ...this new variable.
10161 (mips16_stub_call_address): New function.
10162 (mips16_rdhwr_one_only_stub): New class.
10163 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10164 (mips_output_mips16_rdhwr): Delete.
10165 (mips_finish_stub): New function.
10166 (mips_code_end): Use it to handle rdhwr stubs.
10168 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10171 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10172 when calculating size of integer atomic types.
10174 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10176 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10178 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10180 PR tree-optimization/60003
10181 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10182 * profile.c (branch_prob): Use gimple_call_builtin_p
10183 to check for BUILT_IN_SETJMP_RECEIVER.
10184 * tree-inline.c (copy_bb): Call notice_special_calls.
10186 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10189 * lra-constraints.c (process_alt_operands): Update reload_sum only
10192 2014-01-31 Richard Henderson <rth@redhat.com>
10194 PR middle-end/60004
10195 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10196 until after else_eh is processed.
10198 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10200 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10201 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10202 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10203 in smmintrin.h, remove them.
10204 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10205 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10206 * config/i386/i386.md (ROUND_SAE): Fix value.
10207 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10208 (const48_operand): New.
10209 * config/i386/subst.md (round), (round_expand): Use
10210 const_4_or_8_to_11_operand.
10211 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10213 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10215 * config/i386/constraints.md (Yk): Swap meaning with k.
10216 * config/i386/i386.md (movhi_internal): Change Yk to k.
10217 (movqi_internal): Ditto.
10218 (*k<logic><mode>): Ditto.
10221 (kandn<mode>): Ditto.
10222 (*<code>hi_1): Ditto.
10223 (*<code>qi_1): Ditto.
10224 (kxnor<mode>): Ditto.
10225 (kortestzhi): Ditto.
10226 (kortestchi): Ditto.
10228 (*one_cmplhi2_1): Ditto.
10229 (*one_cmplqi2_1): Ditto.
10230 * config/i386/sse.md (): Change k to Yk.
10231 (avx512f_load<mode>_mask): Ditto.
10232 (avx512f_blendm<mode>): Ditto.
10233 (avx512f_store<mode>_mask): Ditto.
10234 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10235 (avx512f_storedqu<mode>_mask): Ditto.
10236 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10238 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10239 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10240 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10241 (avx512f_maskcmp<mode>3): Ditto.
10242 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10243 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10244 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10245 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10246 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10247 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10248 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10249 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10250 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10251 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10252 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10253 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10254 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10255 (vec_extract_lo_<mode>_maskm): Ditto.
10256 (vec_extract_hi_<mode>_maskm): Ditto.
10257 (avx512f_vternlog<mode>_mask): Ditto.
10258 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10259 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10260 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10261 (avx512f_<code>v8div16qi2_mask): Ditto.
10262 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10263 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10264 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10265 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10266 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10267 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10268 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10269 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10270 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10271 (avx512cd_maskb_vec_dupv8di): Ditto.
10272 (avx512cd_maskw_vec_dupv16si): Ditto.
10273 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10274 (avx512f_vpermi2var<mode>3_mask): Ditto.
10275 (avx512f_vpermi2var<mode>3_mask): Ditto.
10276 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10277 (*avx512f_gathersi<mode>): Ditto.
10278 (*avx512f_gathersi<mode>_2): Ditto.
10279 (*avx512f_gatherdi<mode>): Ditto.
10280 (*avx512f_gatherdi<mode>_2): Ditto.
10281 (*avx512f_scattersi<mode>): Ditto.
10282 (*avx512f_scatterdi<mode>): Ditto.
10283 (avx512f_compress<mode>_mask): Ditto.
10284 (avx512f_compressstore<mode>_mask): Ditto.
10285 (avx512f_expand<mode>_mask): Ditto.
10286 * config/i386/subst.md (mask): Change k to Yk.
10287 (mask_scalar_merge): Ditto.
10290 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10292 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10294 2014-01-31 Richard Biener <rguenther@suse.de>
10296 PR middle-end/59990
10297 * builtins.c (fold_builtin_memory_op): Make sure to not
10298 use a floating-point mode or a boolean or enumeral type for
10299 the copy operation.
10301 2014-01-30 DJ Delorie <dj@redhat.com>
10303 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10304 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10305 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10306 whenever main() has an epilogue.
10308 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10310 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10311 unused variable "field".
10312 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10313 (vsx_mergeh_<mode>): Likewise.
10314 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10315 (altivec_vmrghh): Likewise.
10316 (altivec_vmrghw): Likewise.
10317 (altivec_vmrglb): Likewise.
10318 (altivec_vmrglh): Likewise.
10319 (altivec_vmrglw): Likewise.
10320 (altivec_vspltb): Add missing uses.
10321 (altivec_vsplth): Likewise.
10322 (altivec_vspltw): Likewise.
10323 (altivec_vspltsf): Likewise.
10325 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10328 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10329 frame related instructions.
10331 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10333 PR rtl-optimization/59959
10334 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10335 any reload of register whose subreg is invalid.
10337 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10339 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10340 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10341 Add missing return type - void.
10343 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10345 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10346 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10347 remove element index adjustment for endian (now handled in vsx.md
10349 (altivec_expand_vec_perm_const): Use
10350 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10351 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10352 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10353 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10354 define_expand and a new define_insn *altivec_vspltb_internal;
10355 adjust for -maltivec=be on a little endian target.
10356 (altivec_vspltb_direct): New.
10357 (altivec_vsplth): Divide into a define_expand and a new
10358 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10359 little endian target.
10360 (altivec_vsplth_direct): New.
10361 (altivec_vspltw): Divide into a define_expand and a new
10362 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10363 little endian target.
10364 (altivec_vspltw_direct): New.
10365 (altivec_vspltsf): Divide into a define_expand and a new
10366 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10367 a little endian target.
10369 2014-01-30 Richard Biener <rguenther@suse.de>
10371 PR tree-optimization/59993
10372 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10373 can propagate form the earlier stmt and avoid the transform
10374 when the intermediate result is needed.
10376 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10378 * README.Portability: Fix typo.
10380 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10382 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10383 comparison_operator with ordered_comparison_operator.
10385 2014-01-30 Nick Clifton <nickc@redhat.com>
10387 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10388 Rename to mn10300_store_multiple_regs.
10389 * config/mn10300/mn10300.c: Likewise.
10390 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10391 store_multiple_regs.
10392 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10393 Call mn10300_store_multiple_regs.
10395 2014-01-30 Nick Clifton <nickc@redhat.com>
10396 DJ Delorie <dj@redhat.com>
10398 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10399 %fp 2 to keep registers after it properly word-aligned.
10400 (rl78_alloc_physical_registers_umul): Handle the case where both
10401 input operands are the same.
10403 2014-01-30 Richard Biener <rguenther@suse.de>
10405 PR tree-optimization/59903
10406 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10409 2014-01-30 Jason Merrill <jason@redhat.com>
10412 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10415 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10417 2014-01-30 Richard Biener <rguenther@suse.de>
10419 PR tree-optimization/59951
10420 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10422 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10425 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10426 SFmode to DFmode case.
10428 2014-01-29 DJ Delorie <dj@redhat.com>
10430 * config/msp430/msp430.opt (-minrt): New.
10431 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10433 (ENDFILE_SPEC): Likewise.
10435 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10437 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10438 (estimate_function_body_sizes): Use it.
10440 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10443 * dwarf2out.c (is_cxx_auto): New.
10444 (is_base_type): Use it.
10445 (gen_type_die_with_usage): Likewise.
10447 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10449 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10450 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10451 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10452 -maltivec=be with LE targets.
10453 (vsx_mergeh_<mode>): Likewise.
10454 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10455 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10456 (altivec_vmrghb): Replace with define_expand and new
10457 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10458 (altivec_vmrghb_direct): New define_insn.
10459 (altivec_vmrghh): Replace with define_expand and new
10460 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10461 (altivec_vmrghh_direct): New define_insn.
10462 (altivec_vmrghw): Replace with define_expand and new
10463 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10464 (altivec_vmrghw_direct): New define_insn.
10465 (*altivec_vmrghsf): Adjust for endianness.
10466 (altivec_vmrglb): Replace with define_expand and new
10467 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10468 (altivec_vmrglb_direct): New define_insn.
10469 (altivec_vmrglh): Replace with define_expand and new
10470 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10471 (altivec_vmrglh_direct): New define_insn.
10472 (altivec_vmrglw): Replace with define_expand and new
10473 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10474 (altivec_vmrglw_direct): New define_insn.
10475 (*altivec_vmrglsf): Adjust for endianness.
10476 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10477 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10478 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10479 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10480 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10481 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10482 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10483 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10485 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10487 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10488 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10491 2014-01-29 Richard Biener <rguenther@suse.de>
10493 PR tree-optimization/58742
10494 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10495 associate_pointerplus_align.
10496 (associate_pointerplus_diff): New function.
10497 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10498 and associate_pointerplus_diff.
10500 2014-01-29 Richard Biener <rguenther@suse.de>
10502 * lto-streamer.h (LTO_major_version): Bump to 3.
10503 (LTO_minor_version): Reset to 0.
10505 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10507 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10508 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10509 (arm_file_start): Generate correct asm header for armv7ve.
10510 * config/arm/bpabi.h: Add multilib support for armv7ve.
10511 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10512 and cortex-a15 to armv7ve.
10513 * config/arm/t-aprofile: Add multilib support for armv7ve.
10514 * doc/invoke.texi: Document -march=armv7ve.
10516 2014-01-29 Richard Biener <rguenther@suse.de>
10518 PR tree-optimization/58742
10519 * tree-ssa-forwprop.c (associate_plusminus): Return true
10520 if we changed sth, defer EH cleanup to ...
10521 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10522 (simplify_mult): New function.
10524 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10526 PR middle-end/59917
10527 PR tree-optimization/59920
10528 * tree.c (build_common_builtin_nodes): Remove
10529 __builtin_setjmp_dispatcher initialization.
10530 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10531 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10532 instead of gsi_after_labels + manually skipping debug stmts.
10533 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10534 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10535 * tree-inline.c (copy_edges_for_bb): Remove
10536 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10537 argument. Ignore computed_goto_p stmts. Don't call
10538 make_abnormal_goto_edges. If a call might need abnormal edges
10539 for non-local gotos, see if it already has an edge to
10540 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10541 with true argument, don't do anything then, otherwise add
10542 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10543 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10545 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10546 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10547 (lower_stmt): Don't set data->calls_builtin_setjmp.
10548 (lower_builtin_setjmp): Adjust comment.
10549 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10550 * tree-cfg.c (found_computed_goto): Remove.
10551 (factor_computed_gotos): Remove.
10552 (make_goto_expr_edges): Return bool, true for computed gotos.
10553 Don't call make_abnormal_goto_edges.
10554 (build_gimple_cfg): Don't set found_computed_goto, don't call
10555 factor_computed_gotos.
10556 (computed_goto_p): No longer static.
10557 (make_blocks): Don't set found_computed_goto.
10558 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10559 (make_edges): If make_goto_expr_edges returns true, push bb
10560 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10561 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10562 vector. Record mapping between bbs and OpenMP regions if there
10563 are any, adjust make_gimple_omp_edges caller. Call
10564 handle_abnormal_edges.
10565 (make_abnormal_goto_edges): Remove.
10566 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10567 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10568 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10569 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10570 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10571 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10572 filling *region also set *region_idx to (*region)->entry->index.
10575 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10576 For REGs set ORIGINAL_REGNO.
10578 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10580 * doc/md.texi: Mention that a target shouldn't implement
10581 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10584 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10586 PR tree-optimization/59594
10587 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10588 a copy of the datarefs vector rather than the vector itself.
10590 2014-01-28 Jason Merrill <jason@redhat.com>
10593 * dwarf2out.c (auto_die): New static.
10594 (gen_type_die_with_usage): Handle C++1y 'auto'.
10595 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10598 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10601 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10602 (SPEC_X32): Likewise.
10603 (SPEC_64): Likewise.
10604 * config/i386/i386.c (ix86_option_override_internal): Turn off
10605 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10607 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10608 * config/i386/i386.h (TARGET_16BIT): New macro.
10609 (TARGET_16BIT_P): Likewise.
10610 * config/i386/i386.opt: Add m16.
10611 * doc/invoke.texi: Document -m16.
10613 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10615 PR preprocessor/59935
10616 * input.c (location_get_source_line): Bail out on when line number
10617 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10619 2014-01-28 Richard Biener <rguenther@suse.de>
10621 PR tree-optimization/58742
10622 * tree-ssa-forwprop.c (associate_plusminus): Handle
10623 pointer subtraction of the form (T)(P + A) - (T)P.
10625 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10627 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10630 2014-01-28 Richard Biener <rguenther@suse.de>
10633 2014-01-28 Richard Biener <rguenther@suse.de>
10635 PR rtl-optimization/45364
10636 PR rtl-optimization/59890
10637 * var-tracking.c (local_get_addr_clear_given_value): Handle
10638 already cleared slot.
10639 (val_reset): Handle not allocated local_get_addr_cache.
10640 (vt_find_locations): Use post-order on the inverted CFG.
10642 2014-01-28 Richard Biener <rguenther@suse.de>
10644 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10646 2014-01-28 Richard Biener <rguenther@suse.de>
10648 PR rtl-optimization/45364
10649 PR rtl-optimization/59890
10650 * var-tracking.c (local_get_addr_clear_given_value): Handle
10651 already cleared slot.
10652 (val_reset): Handle not allocated local_get_addr_cache.
10653 (vt_find_locations): Use post-order on the inverted CFG.
10655 2014-01-28 Alan Modra <amodra@gmail.com>
10657 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10658 * configure.ac <recursive call for build != host>: Define
10659 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10660 and LD_FOR_BUILD too.
10661 * configure: Regenerate.
10663 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10665 * config/i386/i386.c (get_builtin_code_for_version): Separate
10666 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10667 Broadwell from Haswell.
10669 2014-01-27 Steve Ellcey <sellcey@mips.com>
10671 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10672 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10673 * config/mips/mips.c (mips_option_override): Change setting
10675 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10676 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10677 Change from Mask to Var.
10679 2014-01-27 Jeff Law <law@redhat.com>
10681 * ipa-inline.c (inline_small_functions): Fix typo.
10683 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10685 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10686 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10687 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10688 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10689 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10690 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10691 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10692 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10693 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10694 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10695 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10696 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10697 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10698 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10699 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10700 (_mm512_storeu_epi64): Ditto.
10701 (_mm512_cmpge_epi32_mask): Ditto.
10702 (_mm512_cmpge_epu32_mask): Ditto.
10703 (_mm512_cmpge_epi64_mask): Ditto.
10704 (_mm512_cmpge_epu64_mask): Ditto.
10705 (_mm512_cmple_epi32_mask): Ditto.
10706 (_mm512_cmple_epu32_mask): Ditto.
10707 (_mm512_cmple_epi64_mask): Ditto.
10708 (_mm512_cmple_epu64_mask): Ditto.
10709 (_mm512_cmplt_epi32_mask): Ditto.
10710 (_mm512_cmplt_epu32_mask): Ditto.
10711 (_mm512_cmplt_epi64_mask): Ditto.
10712 (_mm512_cmplt_epu64_mask): Ditto.
10713 (_mm512_cmpneq_epi32_mask): Ditto.
10714 (_mm512_cmpneq_epu32_mask): Ditto.
10715 (_mm512_cmpneq_epi64_mask): Ditto.
10716 (_mm512_cmpneq_epu64_mask): Ditto.
10717 (_mm512_expand_pd): Ditto.
10718 (_mm512_expand_ps): Ditto.
10719 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10720 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10721 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10722 * config/i386/i386.c (ix86_builtins): Add
10723 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10724 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10725 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10726 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10727 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10728 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10729 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10730 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10731 IX86_BUILTIN_PMOVUSQW512_MEM.
10732 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10733 __builtin_ia32_pmovsqd512mem_mask,
10734 __builtin_ia32_pmovqd512mem_mask,
10735 __builtin_ia32_pmovusqw512mem_mask,
10736 __builtin_ia32_pmovsqw512mem_mask,
10737 __builtin_ia32_pmovqw512mem_mask,
10738 __builtin_ia32_pmovusdw512mem_mask,
10739 __builtin_ia32_pmovsdw512mem_mask,
10740 __builtin_ia32_pmovdw512mem_mask,
10741 __builtin_ia32_pmovqb512mem_mask,
10742 __builtin_ia32_pmovusqb512mem_mask,
10743 __builtin_ia32_pmovsqb512mem_mask,
10744 __builtin_ia32_pmovusdb512mem_mask,
10745 __builtin_ia32_pmovsdb512mem_mask,
10746 __builtin_ia32_pmovdb512mem_mask.
10747 (bdesc_args): Add __builtin_ia32_expanddf512,
10748 __builtin_ia32_expandsf512.
10749 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10750 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10751 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10752 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10753 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10754 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10755 (avx512f_<code>v8div16qi2_mask_store): This.
10756 (avx512f_expand<mode>): New.
10758 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10760 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10762 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10763 (_mm512_prefetch_i32scatter_pd): Ditto.
10764 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10765 (_mm512_prefetch_i64scatter_pd): Ditto.
10766 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10767 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10768 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10769 (_mm512_prefetch_i32scatter_ps): Ditto.
10770 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10771 (_mm512_prefetch_i64scatter_ps): Ditto.
10772 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10773 * config/i386/i386-builtin-types.def: Define
10774 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10775 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10776 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10777 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10778 IX86_BUILTIN_SCATTERPFQPD.
10779 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10780 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10781 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10782 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10783 __builtin_ia32_scatterpfqps.
10784 (ix86_expand_builtin): Expand new built-ins.
10785 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10786 fix memory access data type.
10787 (*avx512pf_gatherpf<mode>_mask): Ditto.
10788 (*avx512pf_gatherpf<mode>): Ditto.
10789 (avx512pf_scatterpf<mode>): Ditto.
10790 (*avx512pf_scatterpf<mode>_mask): Ditto.
10791 (*avx512pf_scatterpf<mode>): Ditto.
10792 (GATHER_SCATTER_SF_MEM_MODE): New.
10793 (avx512pf_gatherpf<mode>df): Ditto.
10794 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10795 (*avx512pf_scatterpf<mode>df): Ditto.
10797 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10800 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10801 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10804 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10806 * common/config/arm/arm-common.c
10807 (arm_rewrite_mcpu): Handle multiple names.
10809 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10811 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10813 * gimple-builder.h (create_gimple_tmp): Delete.
10815 2014-01-27 Christian Bruel <christian.bruel@st.com>
10817 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10820 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10822 * config/pa/pa.md (call): Generate indirect long calls to non-local
10823 functions when outputing 32-bit code.
10824 (call_value): Likewise except for special call to buggy powf function.
10826 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10827 portable runtime and PIC indirect calls.
10828 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10829 and PIC call sequences. Use ldo instead of blr to set return register
10830 in PIC call sequence.
10832 2014-01-25 Walter Lee <walt@tilera.com>
10834 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10835 avoid clobbering a live register.
10837 2014-01-25 Walter Lee <walt@tilera.com>
10839 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10840 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10841 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10842 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10844 2014-01-25 Walter Lee <walt@tilera.com>
10846 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10847 arguments on even registers.
10848 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10850 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10851 (BIGGEST_ALIGNMENT): Ditto.
10852 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10854 2014-01-25 Walter Lee <walt@tilera.com>
10856 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10857 insns before bundling.
10858 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10860 2014-01-25 Walter Lee <walt@tilera.com>
10862 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10863 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10864 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10866 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10868 * config/mips/constraints.md (kl): Delete.
10869 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10870 define expands, using...
10871 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10872 instructions for MIPS16.
10873 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10874 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10876 2014-01-25 Walter Lee <walt@tilera.com>
10878 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10882 2014-01-25 Walter Lee <walt@tilera.com>
10884 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10885 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10887 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10889 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10892 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10894 * print-rtl.c (in_call_function_usage): New var.
10895 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10896 EXPR_LIST mode as mode and not as reg note name.
10898 PR middle-end/59561
10899 * cfgloopmanip.c (copy_loop_info): If
10900 loop->warned_aggressive_loop_optimizations, make sure
10901 the flag is set in target loop too.
10903 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10905 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10907 * builtins.def: Likewise.
10908 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10909 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10910 * ira.c (ira_setup_eliminable_regset): Likewise.
10911 * omp-low.c (gate_expand_omp): Likewise.
10912 (execute_lower_omp): Likewise.
10913 (diagnose_sb_0): Likewise.
10914 (gate_diagnose_omp_blocks): Likewise.
10915 (simd_clone_clauses_extract): Likewise.
10918 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10920 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10921 correction for little endian...
10922 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10925 2014-01-24 Jeff Law <law@redhat.com>
10927 PR tree-optimization/59919
10928 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10929 for non-returning calls.
10931 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10933 * common/config/aarch64/aarch64-common.c
10934 (aarch64_rewrite_mcpu): Handle multiple names.
10935 * config/aarch64/aarch64.h
10936 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10938 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10940 * input.c (add_file_to_cache_tab): Handle the case where fopen
10943 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10946 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10947 from push operand if code of push isn't PRE_DEC.
10949 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10952 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10953 -mquad-memory-atomic. Update -mquad-memory documentation to say
10954 it is only used for non-atomic loads/stores.
10956 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10957 -mquad-memory or -mquad-memory-atomic switches.
10959 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10960 -mquad-memory-atomic to ISA 2.07 support.
10962 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10963 to separate support of normal quad word memory operations (ldq, stq)
10964 from the atomic quad word memory operations.
10966 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10967 support to separate non-atomic quad word operations from atomic
10968 quad word operations. Disable non-atomic quad word operations in
10969 little endian mode so that we don't have to swap words after the
10970 load and before the store.
10971 (quad_load_store_p): Add comment about atomic quad word support.
10972 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10973 options printed with -mdebug=reg.
10975 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10976 -mquad-memory-atomic as the test for whether we have quad word
10977 atomic instructions.
10978 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10979 or -mp8-vector are used, allow byte/half-word atomic operations.
10981 * config/rs6000/sync.md (load_lockedti): Insure that the address
10982 is a proper indexed or indirect address for the lqarx instruction.
10983 On little endian systems, swap the hi/lo registers after the lqarx
10985 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10986 insure the address is valid for the lqarx instruction.
10987 (store_conditionalti): Insure that the address is a proper indexed
10988 or indirect address for the stqcrx. instruction. On little endian
10989 systems, swap the hi/lo registers before doing the stqcrx.
10991 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10992 insure the address is valid for the stqcrx. instruction.
10994 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10995 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10996 type of quad memory support is available.
10998 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11000 PR regression/59915
11001 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11002 there is a danger of looping.
11004 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11006 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11007 force flag_ira_loop_pressure if set via command line.
11009 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11011 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11012 (ashr_simd): New builtin handling DI mode.
11013 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11014 (aarch64_sshr_simddi): New match pattern.
11015 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11016 (vshrd_n_s64): Likewise.
11017 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11019 2014-01-23 Nick Clifton <nickc@redhat.com>
11021 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11022 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11023 favour of mcu specific scripts.
11024 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11027 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11028 Alex Velenko <Alex.Velenko@arm.com>
11030 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11031 (vaddv_s16): Likewise.
11032 (vaddv_s32): Likewise.
11033 (vaddv_u8): Likewise.
11034 (vaddv_u16): Likewise.
11035 (vaddv_u32): Likewise.
11036 (vaddvq_s8): Likewise.
11037 (vaddvq_s16): Likewise.
11038 (vaddvq_s32): Likewise.
11039 (vaddvq_s64): Likewise.
11040 (vaddvq_u8): Likewise.
11041 (vaddvq_u16): Likewise.
11042 (vaddvq_u32): Likewise.
11043 (vaddvq_u64): Likewise.
11044 (vaddv_f32): Likewise.
11045 (vaddvq_f32): Likewise.
11046 (vaddvq_f64): Likewise.
11047 (vmaxv_f32): Likewise.
11048 (vmaxv_s8): Likewise.
11049 (vmaxv_s16): Likewise.
11050 (vmaxv_s32): Likewise.
11051 (vmaxv_u8): Likewise.
11052 (vmaxv_u16): Likewise.
11053 (vmaxv_u32): Likewise.
11054 (vmaxvq_f32): Likewise.
11055 (vmaxvq_f64): Likewise.
11056 (vmaxvq_s8): Likewise.
11057 (vmaxvq_s16): Likewise.
11058 (vmaxvq_s32): Likewise.
11059 (vmaxvq_u8): Likewise.
11060 (vmaxvq_u16): Likewise.
11061 (vmaxvq_u32): Likewise.
11062 (vmaxnmv_f32): Likewise.
11063 (vmaxnmvq_f32): Likewise.
11064 (vmaxnmvq_f64): Likewise.
11065 (vminv_f32): Likewise.
11066 (vminv_s8): Likewise.
11067 (vminv_s16): Likewise.
11068 (vminv_s32): Likewise.
11069 (vminv_u8): Likewise.
11070 (vminv_u16): Likewise.
11071 (vminv_u32): Likewise.
11072 (vminvq_f32): Likewise.
11073 (vminvq_f64): Likewise.
11074 (vminvq_s8): Likewise.
11075 (vminvq_s16): Likewise.
11076 (vminvq_s32): Likewise.
11077 (vminvq_u8): Likewise.
11078 (vminvq_u16): Likewise.
11079 (vminvq_u32): Likewise.
11080 (vminnmv_f32): Likewise.
11081 (vminnmvq_f32): Likewise.
11082 (vminnmvq_f64): Likewise.
11084 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11086 * config/aarch64/aarch64-simd.md
11087 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11088 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11089 (*aarch64_mul3_elt<mode>): Likewise.
11090 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11091 (*aarch64_mul3_elt_to_64v2df): Likewise.
11092 (*aarch64_mla_elt<mode>): Likewise.
11093 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11094 (*aarch64_mls_elt<mode>): Likewise.
11095 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11096 (*aarch64_fma4_elt<mode>): Likewise.
11097 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11098 (*aarch64_fma4_elt_to_64v2df): Likewise.
11099 (*aarch64_fnma4_elt<mode>): Likewise.
11100 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11101 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11102 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11103 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11104 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11105 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11106 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11107 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11108 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11110 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11112 * config/aarch64/aarch64-simd.md
11113 (aarch64_be_checked_get_lane<mode>): New define_expand.
11114 * config/aarch64/aarch64-simd-builtins.def
11115 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11116 New builtin definition.
11117 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11118 Use new safe be builtin.
11120 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11122 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11124 (aarch64_be_st1<mode>): Likewise.
11125 (aarch_ld1<VALL:mode>): Define_expand modified.
11126 (aarch_st1<VALL:mode>): Likewise.
11127 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11128 (UNSPEC_ST1): Likewise.
11130 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11132 * config/microblaze/microblaze.md: Add trap insn and attribute
11134 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11136 PR preprocessor/58580
11137 * input.h (location_get_source_line): Take an additional line_size
11139 (void diagnostics_file_cache_fini): Declare new function.
11140 * input.c (struct fcache): New type.
11141 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11142 New static constants.
11143 (diagnostic_file_cache_init, total_lines_num)
11144 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11145 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11146 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11147 (get_next_line, read_next_line, goto_next_line, read_line_num):
11148 New static function definitions.
11149 (diagnostic_file_cache_fini): New function.
11150 (location_get_source_line): Take an additional output line_len
11151 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11153 * diagnostic.c (diagnostic_finish): Call
11154 diagnostic_file_cache_fini.
11155 (adjust_line): Take an additional input parameter for the length
11156 of the line, rather than calculating it with strlen.
11157 (diagnostic_show_locus): Adjust the use of
11158 location_get_source_line and adjust_line with respect to their new
11159 signature. While displaying a line now, do not stop at the first
11160 null byte. Rather, display the zero byte as a space and keep
11161 going until we reach the size of the line.
11162 * Makefile.in: Add vec.o to OBJS-libcommon
11164 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11165 Ilya Tocar <ilya.tocar@intel.com>
11167 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11168 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11169 (__builtin_ia32_kmov16): Ditto.
11170 * config/i386/i386.md (UNSPEC_KMOV): New.
11173 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11175 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11176 (_mm512_storeu_si512): Ditto.
11178 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11181 * rtl.h (get_referenced_operands): Declare.
11182 * recog.c (get_referenced_operands): New function.
11183 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11184 operands have been referenced when recording LO_SUM references.
11186 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11188 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11190 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11192 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11193 Enable for generic and recent AMD targets.
11195 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11197 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11198 ARG_SIZE note when adjustment was eliminated.
11200 2014-01-22 Jeff Law <law@redhat.com>
11202 PR tree-optimization/59597
11203 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11204 in file. Accept new argument REGISTERING and use it to modify
11205 dump output appropriately.
11206 (register_jump_thread): Corresponding changes.
11207 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11208 thread paths involving joiner blocks. Add code to dump cancelled
11209 jump threading paths.
11211 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11213 PR rtl-optimization/59477
11214 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11215 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11217 2014-01-22 Tom Tromey <tromey@redhat.com>
11219 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11221 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11223 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11225 PR rtl-optimization/59896
11226 * lra-constraints.c (process_alt_operands): Check unused note for
11227 matched operands of insn with no output reloads.
11229 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11231 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11232 (mips_move_from_gpr_cost): Likewise.
11234 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11236 PR rtl-optimization/59858
11237 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11238 ira_class_hard_regs_num.
11239 (process_alt_operands): Increase reject for dying matched operand.
11241 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11244 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11245 smaller than size, perform several stores or loads and stores
11246 at dst + count - size to store or copy all of size bytes, rather
11247 than just last modesize bytes.
11249 2014-01-20 DJ Delorie <dj@redhat.com>
11251 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11252 that CLOBBERs are REGs before propogating their values.
11254 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11256 PR middle-end/59789
11257 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11258 (cgraph_inline_failed_type): New function.
11259 * cgraph.h (DEFCIFCODE): Add type.
11260 (cgraph_inline_failed_type_t): New enum.
11261 (cgraph_inline_failed_type): New prototype.
11262 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11263 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11264 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11265 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11266 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11267 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11268 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11269 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11270 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11271 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11272 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11273 OPTIMIZATION_MISMATCH.
11274 * tree-inline.c (expand_call_inline): Emit errors during
11275 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11277 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11280 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11281 mode attribute in insn output.
11283 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11285 * output.h (output_constant): Delete.
11286 * varasm.c (output_constant): Make private.
11288 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11290 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11292 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11294 PR middle-end/59860
11295 * tree.h (fold_builtin_strcat): New prototype.
11296 * builtins.c (fold_builtin_strcat): No longer static. Add len
11297 argument, if non-NULL, don't call c_strlen. Optimize
11298 directly into __builtin_memcpy instead of __builtin_strcpy.
11299 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11300 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11302 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11304 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11305 for SImode_address_operand operands, having only a REG argument.
11307 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11309 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11310 loader name using mbig-endian.
11311 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11313 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11315 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11316 (-mtune): Likewise.
11319 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11321 * config/aarch64/aarch64-protos.h
11322 (aarch64_cannot_change_mode_class_ptr): Declare.
11323 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11324 aarch64_cannot_change_mode_class_ptr): New.
11325 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11326 backend hook aarch64_cannot_change_mode_class.
11328 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11330 * common/config/aarch64/aarch64-common.c
11331 (aarch64_handle_option): Don't handle any option order logic here.
11332 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11333 selected_cpu, warn on architecture version mismatch.
11334 (aarch64_override_options): Fix parsing order for option strings.
11336 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11337 Iain Sandoe <iain@codesourcery.com>
11340 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11341 warning. Amend comment to reflect current functionality.
11343 2014-01-20 Richard Biener <rguenther@suse.de>
11345 PR middle-end/59860
11346 * builtins.c (fold_builtin_strcat): Remove case better handled
11347 by tree-ssa-strlen.c.
11349 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11351 * config/aarch64/aarch64.opt
11352 (mcpu, march, mtune): Make case-insensitive.
11354 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11357 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11358 if operands[1] is a REG or ZERO_EXTEND of a REG.
11360 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11362 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11364 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11366 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11367 long non-pic millicode calls.
11369 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11371 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11373 2014-01-19 Kito Cheng <kito@0xlab.org>
11375 * builtins.c (expand_movstr): Check movstr expand done or fail.
11377 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11378 H.J. Lu <hongjiu.lu@intel.com>
11381 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11382 to DImode for zero-extended addresses.
11384 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11386 PR rtl-optimization/57763
11387 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11388 on the new indirect jump_insn and increment LABEL_NUSES (label).
11390 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11394 * config.gcc (x86_archs): New variable.
11395 (x86_64_archs): Likewise.
11396 (x86_cpus): Likewise.
11397 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11398 --with-arch/--with-cpu= options.
11399 Support --with-arch=/--with-cpu={nehalem,westmere,
11400 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11402 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11404 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11405 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11407 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11409 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11411 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11414 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11415 clear cpp_get_options (parse_in)->warn_unused_macros for
11416 ix86_target_macros_internal with cpp_define.
11418 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11420 * jump.c (delete_related_insns): Keep (use (insn))s.
11421 * reorg.c (redundant_insn): Check for barriers too.
11423 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11425 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11427 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11429 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11430 call to $$dyncall when TARGET_LONG_CALLS is true.
11432 2014-01-17 Jeff Law <law@redhat.com>
11434 * ree.c (combine_set_extension): Temporarily disable test for
11435 changing number of hard registers.
11437 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11439 PR middle-end/58125
11440 * ipa-inline-analysis.c (inline_free_summary):
11441 Do not free summary of aliases.
11443 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11445 PR middle-end/59706
11446 * gimplify.c (gimplify_expr): Use create_tmp_var
11447 instead of create_tmp_var_raw. If cond doesn't have
11448 integral type, don't add the IFN_ANNOTATE builtin at all.
11450 2014-01-17 Martin Jambor <mjambor@suse.cz>
11453 * ipa-cp.c (prev_edge_clone): New variable.
11454 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11455 Also resize prev_edge_clone vector.
11456 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11457 (ipcp_edge_removal_hook): New function.
11458 (ipcp_driver): Register ipcp_edge_removal_hook.
11460 2014-01-17 Andrew Pinski <apinski@cavium.com>
11461 Steve Ellcey <sellcey@mips.com>
11464 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11467 2014-01-17 Jeff Law <law@redhat.com>
11469 PR middle-end/57904
11470 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11471 so that pass_ccp runs first.
11473 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11475 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11476 (ix86_adjust_cost): Use !TARGET_XXX.
11477 (do_reorder_for_imul): Likewise.
11478 (swap_top_of_ready_list): Likewise.
11479 (ix86_sched_reorder): Likewise.
11481 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11483 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11484 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11485 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11486 (intel_memset): New. Duplicate slm_memset.
11487 (intel_cost): New. Duplicate slm_cost.
11488 (m_INTEL): New macro.
11489 (processor_target_table): Add "intel".
11490 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11491 with PROCESSOR_INTEL for "intel".
11492 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11493 PROCESSOR_SILVERMONT.
11494 (ix86_issue_rate): Likewise.
11495 (ix86_adjust_cost): Likewise.
11496 (ia32_multipass_dfa_lookahead): Likewise.
11497 (swap_top_of_ready_list): Likewise.
11498 (ix86_sched_reorder): Likewise.
11499 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11500 instead of TARGET_OPT_AGU.
11501 * config/i386/i386.h (TARGET_INTEL): New.
11502 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11503 (processor_type): Add PROCESSOR_INTEL.
11504 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11505 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11507 2014-01-17 Marek Polacek <polacek@redhat.com>
11510 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11513 2014-01-17 Richard Biener <rguenther@suse.de>
11515 PR tree-optimization/46590
11516 * opts.c (default_options_table): Add entries for
11517 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11518 all enabled at -O1 but not for -Og.
11519 * common.opt (fbranch-count-reg): Remove Init(1).
11520 (fmove-loop-invariants): Likewise.
11521 (ftree-pta): Likewise.
11523 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11525 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11526 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11527 decls to at least the GCC 4.8 used alignments.
11530 * tree-nested.c (convert_nonlocal_reference_stmt,
11531 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11532 of GIMPLE_BIND stmts, adjust associated decls.
11534 2014-01-17 Richard Biener <rguenther@suse.de>
11536 PR tree-optimization/46590
11537 * vec.h (vec<>::bseach): New member function implementing
11538 binary search according to C89 bsearch.
11539 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11540 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11541 bitmap pointer again. Make accesses_in_loop a flat array.
11542 (mem_ref_obstack): New global.
11543 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11544 (mark_ref_stored): Likewise.
11545 (ref_indep_loop_p_2): Likewise.
11546 (set_ref_stored_in_loop): New helper function.
11547 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11548 (memref_free): Adjust.
11549 (record_mem_ref_loc): Simplify.
11550 (gather_mem_refs_stmt): Adjust.
11551 (sort_locs_in_loop_postorder_cmp): New function.
11552 (analyze_memory_references): Sort accesses_in_loop after
11553 loop postorder number.
11554 (find_ref_loc_in_loop_cmp): New function.
11555 (for_all_locs_in_loop): Find relevant cluster of locs in
11556 accesses_in_loop and iterate without recursion.
11557 (execute_sm): Avoid uninit warning.
11558 (struct ref_always_accessed): Simplify.
11559 (ref_always_accessed::operator ()): Likewise.
11560 (ref_always_accessed_p): Likewise.
11561 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11562 loop postorder numbers here.
11563 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11566 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11569 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11570 on decls for which assemble_alias has been called.
11572 2014-01-17 Nick Clifton <nickc@redhat.com>
11574 * config/msp430/msp430.opt: (mcpu): New option.
11575 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11576 (msp430_option_override): Parse target_cpu. If the MCU name
11577 matches a generic string, clear target_mcu.
11578 (msp430_attr): Allow numeric interrupt values up to 63.
11579 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11580 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11582 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11584 * config/msp430/msp430.md (popm): Use %J rather than %I.
11585 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11586 (addhi_cy_i): Use immediate_operand for operand 2.
11587 * doc/invoke.texi: Document -mcpu option.
11589 2014-01-17 Richard Biener <rguenther@suse.de>
11591 PR rtl-optimization/38518
11592 * df.h (df_analyze_loop): Declare.
11593 * df-core.c: Include cfgloop.h.
11594 (df_analyze_1): Split out main part of df_analyze.
11595 (df_analyze): Adjust.
11596 (loop_inverted_post_order_compute): New function.
11597 (loop_post_order_compute): Likewise.
11598 (df_analyze_loop): New function avoiding whole-function
11599 postorder computes.
11600 * loop-invariant.c (find_defs): Use df_analyze_loop.
11601 (find_invariants): Adjust.
11602 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11604 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11606 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11607 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11609 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11611 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11612 traversal when removing references.
11614 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11617 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11619 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11621 PR middle-end/56791
11622 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11623 pushing a reload for an autoinc when we had previously reloaded an
11624 inner part of the address.
11626 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11628 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11630 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11631 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11632 when not giving up or versioning for alias only because of
11634 (vect_analyze_data_ref_dependences): Set to true.
11635 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11637 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11638 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11641 PR middle-end/58344
11642 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11645 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11646 operand 0 predicate for gathers, use a new pseudo as subtarget.
11648 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11650 PR middle-end/59609
11651 * lra-constraints.c (process_alt_operands): Add printing debug info.
11652 Check absence of input/output reloads for matched operands too.
11654 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11656 PR rtl-optimization/59835
11657 * ira.c (ira_init_register_move_cost): Increase cost for
11660 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11662 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11664 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11667 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11668 non-register objects. Use gen_(high/low)part more consistently.
11671 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11674 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11675 endian support, remove tests for WORDS_BIG_ENDIAN.
11676 (p8_mfvsrd_3_<mode>): Likewise.
11677 (reload_gpr_from_vsx<mode>): Likewise.
11678 (reload_gpr_from_vsxsf): Likewise.
11679 (p8_mfvsrd_4_disf): Likewise.
11681 2014-01-16 Richard Biener <rguenther@suse.de>
11683 PR rtl-optimization/46590
11684 * lcm.c (compute_antinout_edge): Use postorder iteration.
11685 (compute_laterin): Use inverted postorder iteration.
11687 2014-01-16 Nick Clifton <nickc@redhat.com>
11689 PR middle-end/28865
11690 * varasm.c (output_constant): Return the number of bytes actually
11692 (output_constructor_array_range): Update the field size with the
11693 number of bytes emitted by output_constant.
11694 (output_constructor_regular_field): Likewise. Also do not
11695 complain if the total number of bytes emitted is now greater
11696 than the expected fieldpos.
11697 * output.h (output_constant): Update prototype and descriptive comment.
11699 2014-01-16 Marek Polacek <polacek@redhat.com>
11701 PR middle-end/59827
11702 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11703 it is error_mark_node.
11705 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11707 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11708 VALID_AVX256_REG_OR_OI_MODE.
11710 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11712 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11713 current procedure should be profiled.
11715 2014-01-15 Andrew Pinski <apinski@cavium.com>
11717 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11718 of moving from/to the STACK_REG register class.
11720 2014-01-15 Richard Henderson <rth@redhat.com>
11723 * reginfo.c (global_regs_decl): Globalize.
11724 * rtl.h (global_regs_decl): Declare.
11725 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11726 reserved via global_regs.
11728 2014-01-15 Teresa Johnson <tejohnson@google.com>
11730 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11732 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11734 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11735 and vmulosh rather than call gen_vec_widen_smult_*.
11736 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11737 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11738 (vec_widen_smult_even_v16qi): Likewise.
11739 (vec_widen_umult_even_v8hi): Likewise.
11740 (vec_widen_smult_even_v8hi): Likewise.
11741 (vec_widen_umult_odd_v16qi): Likewise.
11742 (vec_widen_smult_odd_v16qi): Likewise.
11743 (vec_widen_umult_odd_v8hi): Likewise.
11744 (vec_widen_smult_odd_v8hi): Likewise.
11745 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11746 vmuloub rather than call gen_vec_widen_umult_*.
11747 (vec_widen_umult_lo_v16qi): Likewise.
11748 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11749 vmulosb rather than call gen_vec_widen_smult_*.
11750 (vec_widen_smult_lo_v16qi): Likewise.
11751 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11752 rather than call gen_vec_widen_umult_*.
11753 (vec_widen_umult_lo_v8hi): Likewise.
11754 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11755 rather than call gen_vec_widen_smult_*.
11756 (vec_widen_smult_lo_v8hi): Likewise.
11758 2014-01-15 Jeff Law <law@redhat.com>
11760 PR tree-optimization/59747
11761 * ree.c (find_and_remove_re): Properly handle case where a second
11762 eliminated extension requires widening a copy created for elimination
11763 of a prior extension.
11764 (combine_set_extension): Ensure that the number of hard regs needed
11765 for a destination register does not change when we widen it.
11767 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11769 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11770 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11771 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11772 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11773 (avr-*-rtems*): Likewise.
11774 (bfin*-rtems*): Likewise.
11775 (moxie-*-rtems*): Likewise.
11776 (h8300-*-rtems*): Likewise.
11777 (i[34567]86-*-rtems*): Likewise.
11778 (lm32-*-rtems*): Likewise.
11779 (m32r-*-rtems*): Likewise.
11780 (m68k-*-rtems*): Likewise.
11781 (microblaze*-*-rtems*): Likewise.
11782 (mips*-*-rtems*): Likewise.
11783 (powerpc-*-rtems*): Likewise.
11784 (sh-*-rtems*): Likewise.
11785 (sparc-*-rtems*): Likewise.
11786 (sparc64-*-rtems*): Likewise.
11787 (v850-*-rtems*): Likewise.
11788 (m32c-*-rtems*): Likewise.
11790 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11792 PR rtl-optimization/59511
11793 * ira.c (ira_init_register_move_cost): Use memory costs for some
11794 cases of register move cost calculations.
11795 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11796 instead of BB frequency.
11797 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11798 * lra-assigns.c (find_hard_regno_for): Ditto.
11800 2014-01-15 Richard Biener <rguenther@suse.de>
11802 PR tree-optimization/59822
11803 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11804 (vectorizable_load): Use it to hoist defs of uses of invariant
11805 loads out of the loop.
11807 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11808 Kugan Vivekanandarajah <kuganv@linaro.org>
11811 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11814 2014-01-15 Richard Biener <rguenther@suse.de>
11816 PR rtl-optimization/59802
11817 * lcm.c (compute_available): Use inverted postorder to seed
11818 the initial worklist.
11820 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11823 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11824 ADDR_REGS for invalid symrefs in non-PIC code.
11826 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11829 * builtins.c (determine_block_size): Initialize *probable_max_size
11830 even if len_rtx is CONST_INT.
11832 2014-01-14 Andrew Pinski <apinski@cavium.com>
11834 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11835 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11836 (cortexa53_tunings): Likewise.
11837 (aarch64_sched_issue_rate): New function.
11838 (TARGET_SCHED_ISSUE_RATE): Define.
11840 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11842 * ira-costs.c (find_costs_and_classes): Add missed
11843 ira_init_register_move_cost_if_necessary.
11845 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11848 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11850 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11853 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11854 to indicate if type is used for function return value. Warn ABI
11855 change if the vector mode isn't available for function return value.
11856 (ix86_function_arg_advance): Pass false to type_natural_mode.
11857 (ix86_function_arg): Likewise.
11858 (ix86_gimplify_va_arg): Likewise.
11859 (function_arg_32): Don't warn ABI change.
11860 (ix86_function_value): Pass true to type_natural_mode.
11861 (ix86_return_in_memory): Likewise.
11862 (ix86_struct_value_rtx): Removed.
11863 (TARGET_STRUCT_VALUE_RTX): Likewise.
11865 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11867 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11868 converting a conditional jump into a conditional return.
11870 2014-01-14 Richard Biener <rguenther@suse.de>
11872 PR tree-optimization/58921
11873 PR tree-optimization/59006
11874 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11875 hoisting invariant stmts.
11876 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11877 invariant loads on the preheader edge if possible.
11879 2014-01-14 Joey Ye <joey.ye@arm.com>
11881 * doc/plugin.texi (Building GCC plugins): Update to C++.
11883 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11885 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11886 (_mm_rcp28_round_ss): Ditto.
11887 (_mm_rsqrt28_round_sd): Ditto.
11888 (_mm_rsqrt28_round_ss): Ditto.
11889 (_mm_rcp28_sd): Ditto.
11890 (_mm_rcp28_ss): Ditto.
11891 (_mm_rsqrt28_sd): Ditto.
11892 (_mm_rsqrt28_ss): Ditto.
11893 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11894 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11895 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11896 (IX86_BUILTIN_RCP28SD): Ditto.
11897 (IX86_BUILTIN_RCP28SS): Ditto.
11898 (IX86_BUILTIN_RSQRT28SD): Ditto.
11899 (IX86_BUILTIN_RSQRT28SS): Ditto.
11900 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11901 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11902 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11903 (ix86_expand_special_args_builtin): Expand new FTYPE.
11904 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11905 (srcp14<mode>): Make insn unary.
11906 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11907 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11908 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11909 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11910 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11911 Fix rounding: make it SAE only.
11912 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11914 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11916 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11917 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11918 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11919 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11920 (round_saeonly_mask_scalar_operand4): Ditto.
11921 (round_saeonly_mask_scalar_op3): Ditto.
11922 (round_saeonly_mask_scalar_op4): Ditto.
11924 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11926 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11927 Implement -maltivec=be for vec_insert and vec_extract.
11929 2014-01-10 DJ Delorie <dj@redhat.com>
11931 * config/msp430/msp430.md (call_internal): Don't allow memory
11932 references with SP as the base register.
11933 (call_value_internal): Likewise.
11934 * config/msp430/constraints.md (Yc): New. For memory references
11935 that don't use SP as a base register.
11937 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11938 "an integer without a # prefix"
11939 * config/msp430/msp430.md (epilogue_helper): Use it.
11941 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11944 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11945 AVX512F gather builtins.
11946 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11947 on gather decls with INTEGER_TYPE masktype.
11948 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11949 directly into the builtin rather than hoisting it before loop.
11951 PR tree-optimization/59387
11952 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11953 (scev_const_prop): If folded_casts and type has undefined overflow,
11954 use force_gimple_operand instead of force_gimple_operand_gsi and
11955 for each added stmt if it is assign with
11956 arith_code_with_undefined_signed_overflow, call
11957 rewrite_to_defined_overflow.
11958 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11959 gimple-fold.h instead.
11960 (arith_code_with_undefined_signed_overflow,
11961 rewrite_to_defined_overflow): Moved to ...
11962 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11963 rewrite_to_defined_overflow): ... here. No longer static.
11964 Include gimplify-me.h.
11965 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11966 rewrite_to_defined_overflow): New prototypes.
11968 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11970 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11972 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11974 * builtins.c (get_object_alignment_2): Minor tweak.
11975 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11977 2014-01-13 Christian Bruel <christian.bruel@st.com>
11979 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11980 optimized non constant lengths.
11982 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11985 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11986 load as __atomic_load_N if possible.
11988 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11990 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11992 (rs6000_expand_builtin): Adjust call.
11994 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11997 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11998 * config/rs6000/rs6000.c: Include target-globals.h.
11999 (rs6000_set_current_function): Instead of doing target_reinit
12000 unconditionally, use save_target_globals_default_opts and
12001 restore_target_globals.
12003 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12005 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12006 (rs6000_expand_builtin): Handle mffs and mtfsf.
12007 (rs6000_init_builtins): Define mffs and mtfsf.
12008 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12009 (rs6000_mffs): New pattern.
12010 (rs6000_mtfsf): New pattern.
12012 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12014 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12015 Start narrowing with START. Apply candidate-use pair
12016 and check overall cost in narrowing.
12017 (iv_ca_prune): Pass new argument.
12019 2014-01-10 Jeff Law <law@redhat.com>
12021 PR middle-end/59743
12022 * ree.c (combine_reaching_defs): Ensure the defining statement
12023 occurs before the extension when optimizing extensions with
12024 different source and destination hard registers.
12026 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12029 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12030 vtables into the type inheritance graph.
12032 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12034 PR rtl-optimization/59754
12035 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12036 modes in the REGNO != REGNO case.
12038 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12040 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12042 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12044 PR tree-optimization/59745
12045 * tree-predcom.c (tree_predictive_commoning_loop): Call
12046 free_affine_expand_cache if giving up because components is NULL.
12048 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12049 GC in payload of target_globals struct instead of allocating them on
12050 the heap and the larger structs separately using GC.
12051 * target-globals.h (struct target_globals): Make regs, hard_regs,
12052 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12053 of GTY((skip)) and change type to void *.
12054 (reset_target_globals): Cast loads from those fields to corresponding
12057 2014-01-10 Steve Ellcey <sellcey@mips.com>
12060 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12061 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12062 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12064 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12067 * aarch64-modes.def (CC_Zmode): New flags mode.
12068 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12069 represents an equality.
12070 (aarch64_get_condition_code): Handle CC_Zmode.
12071 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12073 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12075 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12076 extraction in good case.
12078 2014-01-10 Richard Biener <rguenther@suse.de>
12080 PR tree-optimization/59374
12081 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12082 checking after SLP discovery. Mark stmts not participating
12083 in any SLP instance properly.
12085 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12087 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12088 when handling a SET rtx.
12090 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12092 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12093 (cortex-a57): Likewise.
12094 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12096 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12098 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12099 non-iwmmxt builtins.
12101 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12105 * ipa-devirt.c record_target_from_binfo): Take as argument
12106 stack of binfos and lookup matching one for virtual inheritance.
12107 (possible_polymorphic_call_targets_1): Update.
12109 2014-01-10 Huacai Chen <chenhc@lemote.com>
12111 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12112 kernel strings for Loongson-2E/2F/3A.
12114 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12116 PR middle-end/59670
12117 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12118 is_gimple_call before calling gimple_call_internal_p.
12120 2014-01-09 Steve Ellcey <sellcey@mips.com>
12122 * Makefile.in (TREE_FLOW_H): Remove.
12123 (TREE_SSA_H): Add file names from tree-flow.h.
12124 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12125 * tree.h: Remove tree-flow.h reference.
12126 * hash-table.h: Remove tree-flow.h reference.
12127 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12128 reference with tree-ssa-loop.h.
12130 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12132 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12133 default element-order behavior for -maltivec.
12134 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12135 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12136 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12137 when targeting big endian, at least for now.
12138 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12140 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12142 PR middle-end/47735
12143 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12144 var satisfies use_register_for_decl, just take into account type
12145 alignment, rather than decl alignment.
12147 PR tree-optimization/59622
12148 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12149 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12150 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12151 Don't devirtualize for inplace at all. For targets.length () == 1,
12152 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12154 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12156 * config/i386/i386.md (cpu): Remove the unused btver1.
12158 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12160 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12162 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12165 * tree-core.h (struct target_globals): New forward declaration.
12166 (struct tree_target_option): Add globals field.
12167 * tree.h (TREE_TARGET_GLOBALS): Define.
12168 (prepare_target_option_nodes_for_pch): New prototype.
12169 * target-globals.h (struct target_globals): Define even if
12170 !SWITCHABLE_TARGET.
12171 * tree.c (prepare_target_option_node_for_pch,
12172 prepare_target_option_nodes_for_pch): New functions.
12173 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12174 * config/i386/i386.c: Include target-globals.h.
12175 (ix86_set_current_function): Instead of doing target_reinit
12176 unconditionally, use save_target_globals_default_opts and
12177 restore_target_globals.
12179 2014-01-09 Richard Biener <rguenther@suse.de>
12181 PR tree-optimization/59715
12182 * tree-cfg.h (split_critical_edges): Declare.
12183 * tree-cfg.c (split_critical_edges): Export.
12184 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12186 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12188 * cfgexpand.c (expand_stack_vars): Optionally disable
12189 asan stack protection.
12190 (expand_used_vars): Likewise.
12191 (partition_stack_vars): Likewise.
12192 * asan.c (asan_emit_stack_protection): Optionally disable
12193 after return stack usage.
12194 (instrument_derefs): Optionally disable memory access instrumentation.
12195 (instrument_builtin_call): Likewise.
12196 (instrument_strlen_call): Likewise.
12197 (asan_protect_global): Optionally disable global variables protection.
12198 * doc/invoke.texi: Added doc for new options.
12199 * params.def: Added new options.
12200 * params.h: Likewise.
12202 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12204 PR rtl-optimization/59724
12205 * ifcvt.c (cond_exec_process_if_block): Don't call
12206 flow_find_head_matching_sequence with 0 longest_match.
12207 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12208 non-active insns if !stop_after.
12209 (try_head_merge_bb): Revert 2014-01-07 changes.
12211 2014-01-08 Jeff Law <law@redhat.com>
12213 * ree.c (get_sub_rtx): New function, extracted from...
12214 (merge_def_and_ext): Here.
12215 (combine_reaching_defs): Use get_sub_rtx.
12217 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12219 * cgraph.h (varpool_variable_node): Do not choke on null node.
12221 2014-01-08 Catherine Moore <clm@codesourcery.com>
12223 * config/mips/mips.md (simple_return): Attempt to use JRC
12225 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12227 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12229 PR rtl-optimization/59137
12230 * reorg.c (steal_delay_list_from_target): Call update_block for
12232 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12234 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12236 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12237 two duplicate entries.
12239 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12242 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12244 * config/mips/mips.c (mips_truncated_op_cost): New function.
12245 (mips_rtx_costs): Adjust test for BADDU.
12246 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12248 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12250 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12251 (*baddu_si): ...this new pattern.
12253 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12256 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12258 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12260 PR middle-end/57748
12261 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12263 (expand_expr, expand_normal): Adjust.
12264 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12265 inner_reference_p. Use inner_reference_p to expand inner references.
12266 (store_expr): Adjust.
12267 * cfgexpand.c (expand_call_stmt): Adjust.
12269 2014-01-08 Rong Xu <xur@google.com>
12271 * gcov-io.c (gcov_var): Move from gcov-io.h.
12272 (gcov_position): Ditto.
12273 (gcov_is_error): Ditto.
12274 (gcov_rewrite): Ditto.
12275 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12276 only part to libgcc/libgcov.h.
12278 2014-01-08 Marek Polacek <polacek@redhat.com>
12280 PR middle-end/59669
12281 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12283 2014-01-08 Marek Polacek <polacek@redhat.com>
12286 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12288 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12290 PR rtl-optimization/59649
12291 * stor-layout.c (get_mode_bounds): For BImode return
12292 0 and STORE_FLAG_VALUE.
12294 2014-01-08 Richard Biener <rguenther@suse.de>
12296 PR middle-end/59630
12297 * gimple.h (is_gimple_builtin_call): Remove.
12298 (gimple_builtin_call_types_compatible_p): New.
12299 (gimple_call_builtin_p): New overload.
12300 * gimple.c (is_gimple_builtin_call): Remove.
12301 (validate_call): Rename to ...
12302 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12303 check return types.
12304 (validate_type): New static function.
12305 (gimple_call_builtin_p): New overload and adjust.
12306 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12307 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12308 (gimple_fold_stmt_to_constant_1): Likewise.
12309 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12311 2014-01-08 Richard Biener <rguenther@suse.de>
12313 PR middle-end/59471
12314 * gimplify.c (gimplify_expr): Gimplify register-register type
12315 VIEW_CONVERT_EXPRs to separate stmts.
12317 2014-01-07 Jeff Law <law@redhat.com>
12319 PR middle-end/53623
12320 * ree.c (combine_set_extension): Handle case where source
12321 and destination registers in an extension insn are different.
12322 (combine_reaching_defs): Allow source and destination registers
12323 in extension to be different under limited circumstances.
12324 (add_removable_extension): Remove restriction that the
12325 source and destination registers in the extension are the same.
12326 (find_and_remove_re): Emit a copy from the extension's
12327 destination to its source after the defining insn if
12328 the source and destination registers are different.
12330 PR middle-end/59285
12331 * ifcvt.c (merge_if_block): If we are merging a block with more than
12332 one successor with a block with no successors, remove any BARRIER
12333 after the second block.
12335 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12337 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12339 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12342 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12343 for 14-bit register offsets when INT14_OK_STRICT is false.
12345 2014-01-07 Roland Stigge <stigge@antcom.de>
12346 Michael Meissner <meissner@linux.vnet.ibm.com>
12349 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12350 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12352 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12354 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12357 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12359 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12360 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12361 rtx is const0_rtx or not.
12363 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12366 * target-globals.c (save_target_globals): Remove this_fn_optab
12368 * toplev.c: Include optabs.h.
12369 (target_reinit): Temporarily restore the global options if another
12370 set of options are in force.
12372 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12374 PR rtl-optimization/58668
12375 * cfgcleanup.c (flow_find_cross_jump): Don't count
12376 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12377 to determine what is counted.
12378 (flow_find_head_matching_sequence): Use active_insn_p to determine
12380 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12382 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12383 determine what is counted.
12385 PR tree-optimization/59643
12386 * tree-predcom.c (split_data_refs_to_components): If one dr is
12387 read and one write, determine_offset fails and the write isn't
12388 in the bad component, just put the read into the bad component.
12390 2014-01-07 Mike Stump <mikestump@comcast.net>
12391 Jakub Jelinek <jakub@redhat.com>
12394 * tree-core.h (struct tree_optimization_option): Change optabs
12395 type from unsigned char * to void *.
12396 * optabs.c (init_tree_optimization_optabs): Adjust
12397 TREE_OPTIMIZATION_OPTABS initialization.
12399 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12402 * config/i386/i386.h (struct machine_function): Add
12403 no_drap_save_restore field.
12404 * config/i386/i386.c (ix86_save_reg): Use
12405 !cfun->machine->no_drap_save_restore instead of
12406 crtl->stack_realign_needed.
12407 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12408 this function clears frame_pointer_needed. Set
12409 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12410 and DRAP reg is needed.
12412 2014-01-06 Marek Polacek <polacek@redhat.com>
12415 * doc/implement-c.texi: Mention that other integer types are
12416 permitted as bit-field types in strictly conforming mode.
12418 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12420 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12421 is newly allocated.
12423 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12425 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12427 2014-01-06 Martin Jambor <mjambor@suse.cz>
12430 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12432 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12434 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12438 * var-tracking.c (add_stores): Preserve the value of the source even if
12439 we don't record the store.
12441 2014-01-06 Terry Guo <terry.guo@arm.com>
12443 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12445 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12448 * config/darwin.c (darwin_function_section): Adjust return values to
12449 correspond to optimisation changes made in r206070.
12451 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12453 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12454 from prefetch_block tune setting.
12455 (nocona_cost): Correct size of prefetch block to 64.
12457 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12459 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12460 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12461 used to save the static chain register in the computation of the offset
12462 from which the FP registers need to be restored.
12464 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12466 PR tree-optimization/59519
12467 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12468 ICE if get_current_def (current_new_name) is already non-NULL, as long
12469 as it is a phi result of some other phi in *new_exit_bb that has
12472 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12473 or vmovdqu* for misaligned_operand.
12474 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12475 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12476 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12477 aligned_mem for AVX512F masked aligned load and store builtins and for
12478 non-temporal moves.
12480 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12482 PR tree-optimization/59651
12483 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12484 Address range for negative step should be added by TYPE_SIZE_UNIT.
12486 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12488 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12489 overlapping registers also for registers other than the stack pointer.
12491 2014-01-03 Marek Polacek <polacek@redhat.com>
12494 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12497 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12500 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12503 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12504 (push splitter): Use <P:MODE_SIZE> instead of
12505 GET_MODE_SIZE (<P:MODE>mode).
12506 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12507 (mov -1, reg peephole2): Likewise.
12508 * config/i386/sse.md (*mov<mode>_internal,
12509 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12510 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12511 *<code><mode>3, *andnot<mode>3<mask_name>,
12512 <mask_codefor><code><mode>3<mask_name>): Likewise.
12513 * config/i386/subst.md (mask_mode512bit_condition,
12514 sd_mask_mode512bit_condition): Likewise.
12516 2014-01-02 Xinliang David Li <davidxl@google.com>
12518 PR tree-optimization/59303
12519 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12520 (dump_predicates): Better output format.
12521 (pred_equal_p): New function.
12522 (is_neq_relop_p): Ditto.
12523 (is_neq_zero_form_p): Ditto.
12524 (pred_expr_equal_p): Ditto.
12525 (pred_neg_p): Ditto.
12526 (simplify_pred): Ditto.
12527 (simplify_preds_2): Ditto.
12528 (simplify_preds_3): Ditto.
12529 (simplify_preds_4): Ditto.
12530 (simplify_preds): Ditto.
12531 (push_pred): Ditto.
12532 (push_to_worklist): Ditto.
12533 (get_pred_info_from_cmp): Ditto.
12534 (is_degenerated_phi): Ditto.
12535 (normalize_one_pred_1): Ditto.
12536 (normalize_one_pred): Ditto.
12537 (normalize_one_pred_chain): Ditto.
12538 (normalize_preds): Ditto.
12539 (normalize_cond_1): Remove function.
12540 (normalize_cond): Ditto.
12541 (is_gcond_subset_of): Ditto.
12542 (is_subset_of_any): Ditto.
12543 (is_or_set_subset_of): Ditto.
12544 (is_and_set_subset_of): Ditto.
12545 (is_norm_cond_subset_of): Ditto.
12546 (pred_chain_length_cmp): Ditto.
12547 (convert_control_dep_chain_into_preds): Type change.
12548 (find_predicates): Ditto.
12549 (find_def_preds): Ditto.
12550 (destroy_predicates_vecs): Ditto.
12551 (find_matching_predicates_in_rest_chains): Ditto.
12552 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12553 (is_pred_expr_subset): Ditto.
12554 (is_pred_chain_subset_of): Ditto.
12555 (is_included_in): Ditto.
12556 (is_superset_of): Ditto.
12558 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12560 Update copyright years.
12562 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12564 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12565 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12566 config/arc/arc.md, config/arc/arc.opt,
12567 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12568 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12569 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12570 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12571 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12572 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12573 vtable-verify.c, vtable-verify.h: Use the standard form for the
12576 2014-01-02 Tobias Burnus <burnus@net-b.de>
12578 * gcc.c (process_command): Update copyright notice dates.
12579 * gcov-dump.c: Ditto.
12581 * doc/cpp.texi: Bump @copying's copyright year.
12582 * doc/cppinternals.texi: Ditto.
12583 * doc/gcc.texi: Ditto.
12584 * doc/gccint.texi: Ditto.
12585 * doc/gcov.texi: Ditto.
12586 * doc/install.texi: Ditto.
12587 * doc/invoke.texi: Ditto.
12589 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12591 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12593 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12595 * config/i386/sse.md (*mov<mode>_internal): Guard
12596 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12598 PR rtl-optimization/59647
12599 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12600 new_rtx into UNSIGNED_FLOAT rtxes.
12602 Copyright (C) 2014 Free Software Foundation, Inc.
12604 Copying and distribution of this file, with or without modification,
12605 are permitted in any medium without royalty provided the copyright
12606 notice and this notice are preserved.