1 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
4 * config/i386/i386.c (ix86_rtx_costs)
5 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
8 2014-05-26 Martin Jambor <mjambor@suse.cz>
10 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
13 2014-05-26 Richard Biener <rguenther@suse.de>
15 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
16 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
17 Provide specializations.
18 (wi::int_traits <HOST_WIDE_INT>,
19 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
21 2014-05-26 Alan Modra <amodra@gmail.com>
24 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
25 params and return a bool. Remove dead code. Update comment.
26 Assert we have a const_int source. Remove bogus code from
27 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
28 handling of constants > 2G and reg_equal note, from..
29 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
30 return value. Update comment. If we can, use a new pseudo
31 for intermediate calculations.
32 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
34 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
35 call to rs6000_emit_set_const in splitter.
36 (movdi_internal64+2, +3): Likewise.
38 2014-05-26 Richard Biener <rguenther@suse.de>
40 * system.h: Define __STDC_FORMAT_MACROS before
42 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
43 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
44 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
45 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
46 HOST_WIDEST_INT_C): Remove.
47 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
48 if C99 inttypes.h is not available.
49 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
50 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
51 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
52 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
53 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
54 (struct output_info): Likewise.
55 (print_statistics): Adjust.
56 (dump_bitmap_statistics): Likewise.
57 * bt-load.c (migrate_btr_defs): Print with PRId64.
58 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
59 (MAX_SAFE_MULTIPLIER): Adjust.
60 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
61 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
62 dump_cgraph_node): Likewise.
63 * final.c (dump_basic_block_info): Likewise.
64 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
65 * gcov.c (format_gcov): Likewise.
66 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
68 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
69 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
70 (inline_small_functions, dump_overall_stats, dump_inline_stats):
71 Use PRId64 for dumping.
72 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
73 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
74 (add_allocno_hard_regs): Adjust.
75 * loop-doloop.c (doloop_modify): Print using PRId64.
76 * loop-iv.c (inverse): Compute in uint64_t.
77 (determine_max_iter, iv_number_of_iterations): Likewise.
78 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
80 * lto-streamer-out.c (write_symbol): Use uint64_t.
81 * mcf.c (CAP_INFINITY): Use int64_t maximum.
82 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
83 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
84 * modulo-sched.c (const_iteration_count): Use int64_t.
85 (sms_schedule): Dump using PRId64.
86 * predict.c (dump_prediction): Likewise.
87 * pretty-print.h (pp_widest_integer): Remove.
88 * profile.c (get_working_sets, is_edge_inconsistent,
89 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
90 * tree-pretty-print.c (pp_double_int): Remove case handling
91 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
92 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
94 (pass_optimize_bswap::execute): Remove restriction on hosts.
95 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
96 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
97 * tree.c (widest_int_cst_value): Remove.
98 * tree.h (widest_int_cst_value): Likewise.
99 * value-prof.c (dump_histogram_value): Print using PRId64.
100 * gengtype.c (main): Also inject int64_t.
101 * ggc-page.c (struct max_alignment): Use int64_t.
102 * alloc-pool.c (struct allocation_object_def): Likewise.
103 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
105 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
106 * doc/tm.texi: Regenerated.
107 * gengtype-lex.l (IWORD): Handle [u]int64_t.
108 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
109 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
110 mmix_output_register_setting): Use [u]int64_t in prototypes.
111 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
112 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
113 mmix_output_octa, mmix_output_shifted_value): Adjust.
114 (mmix_intval): Adjust. Remove unreachable case.
115 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
117 2014-05-26 Richard Biener <rguenther@suse.de>
119 * configure.ac: Drop __int64 type check. Insist that we
120 found uint64_t and int64_t.
121 * hwint.h (HOST_BITS_PER___INT64): Remove.
122 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
123 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
124 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
125 (HOST_WIDEST_FAST_INT): Remove __int64 case.
126 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
127 for dst_q_src_df_rms_cdt.
128 * configure: Regenerate.
129 * config.in: Likewise.
131 2014-05-26 Michael Tautschnig <mt@debian.org>
134 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
135 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
137 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
139 PR rtl-optimization/61278
140 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
142 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
144 PR rtl-optimization/61220
145 Part of PR rtl-optimization/61225
146 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
147 insn; skip split_edge for a block with only one successor.
149 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
151 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
154 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
156 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
157 (update_vtable_references): New function.
158 (function_and_variable_visibility): Rewrite also vtable initializers.
159 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
161 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
163 * ggc.h (ggc_grow): New function.
164 * ggc-none.c (ggc_grow): New function.
165 * ggc-page.c (ggc_grow): Likewise.
167 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
169 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
170 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
171 comdat_can_be_unshared_p, cgraph_externally_visible_p,
172 varpool_externally_visible_p, can_replace_by_local_alias,
173 update_visibility_by_resolution_info, function_and_variable_visibility,
174 pass_data_ipa_function_and_variable_visibility,
175 make_pass_ipa_function_and_variable_visibility,
176 whole_program_function_and_variable_visibility,
177 pass_data_ipa_whole_program_visibility,
178 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
179 * cgraph.h (cgraph_local_node_p): Declare.
180 * ipa-visibility.c: New file.
181 * Makefile.in (OBJS): Add ipa-visiblity.o
183 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
185 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
186 that var decl is available.
188 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
190 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
192 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
193 (find_decls_types_r): Do not walk COMDAT_GROUP.
194 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
195 * varasm.c (make_decl_one_only): Use set_comdat_group;
196 create node if needed.
197 * ipa-inline-transform.c (save_inline_function_body): Update
198 way we decl->symtab mapping.
199 * symtab.c (symtab_hash, hash_node, eq_node
200 symtab_insert_node_to_hashtable): Remove.
201 (symtab_register_node): Update.
202 (symtab_unregister_node): Update.
203 (symtab_get_node): Reimplement as inline function.
204 (symtab_add_to_same_comdat_group): Update.
205 (symtab_dissolve_same_comdat_group_list): Update.
206 (dump_symtab_base): Update.
207 (verify_symtab_base): Update.
208 (symtab_make_decl_local): Update.
209 (fixup_same_cpp_alias_visibility): Update.
210 (symtab_nonoverwritable_alias): Update.
211 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
212 * ipa.c (update_visibility_by_resolution_info): UPdate.
213 * bb-reorder.c: Include cgraph.h
214 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
216 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
217 * cgraph.c (cgraph_get_create_node): Update.
218 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
220 (symtab_get_node): Make inline.
221 (symtab_insert_node_to_hashtable): Remove.
222 (symtab_can_be_discarded): Update.
223 (decl_comdat_group): New function.
224 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
226 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
228 (read_comdat_group): New function.
229 (input_node, input_varpool_node): Use it.
230 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
232 * mips.c (mips_start_unique_function): Likewise.
233 (ix86_code_end): Likewise.
234 (rs6000_code_end): Likweise.
235 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
237 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
239 * gengtype-state.c (fatal_reading_state): Bring offline.
240 * optabs.c (widening_optab_handler): Bring offline.
241 * optabs.h (widening_optab_handler): Likewise.
242 * final.c (get_attr_length_1): Likewise.
244 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
246 * sched-int.h (sd_iterator_cond): Manually tail recurse.
248 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
250 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
251 (ppc440-compare): Include shift with dot.
252 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
253 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
255 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
257 (e6500_sfx2): Include it.
258 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
259 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
260 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
261 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
262 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
263 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
264 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
265 *lshiftrt_internal1le, *lshiftrt_internal1be,
266 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
267 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
268 *rotldi3_internal10le, *rotldi3_internal10be,
269 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
270 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
271 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
272 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
273 define_insns): Use type "shift" in the appropriate alternatives.
275 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
277 * config/rs6000/rs6000.md (type): Add "logical". Delete
279 (dot): Adjust comment.
280 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
281 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
282 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
283 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
284 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
285 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
286 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
287 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
289 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
290 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
291 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
292 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
293 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
294 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
295 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
296 * config/rs6000/8540.md (ppc8540_su): Adjust.
297 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
298 cell-cmp-microcoded): Adjust.
299 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
300 * config/rs6000/e500mc.md (e500mc_su): Adjust.
301 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
302 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
303 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
304 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
305 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
306 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
307 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
309 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
310 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
311 Adjust. Adjust comment.
312 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
313 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
315 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
317 * config/rs6000/rs6000.md (type): Add "add".
318 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
319 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
320 define_insns): Use it.
321 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
323 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
324 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
325 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
326 * config/rs6000/601.md (ppc601-integer): Adjust.
327 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
328 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
329 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
330 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
331 * config/rs6000/8540.md (ppc8540_su): Adjust.
332 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
333 cell-cmp-microcoded): Adjust.
334 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
335 * config/rs6000/e500mc.md (e500mc_su): Adjust.
336 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
337 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
338 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
339 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
340 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
341 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
342 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
344 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
345 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
346 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
347 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
349 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
351 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
352 "delayed_compare", "var_delayed_compare".
353 (var_shift): New attribute.
354 (cell_micro): Adjust.
355 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
356 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
357 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
358 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
359 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
360 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
361 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
362 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
363 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
364 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
365 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
366 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
367 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
368 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
369 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
370 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
371 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
372 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
373 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
374 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
375 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
376 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
377 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
378 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
379 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
381 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
382 * config/rs6000/440.md (ppc440-integer): Adjust.
383 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
384 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
385 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
386 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
387 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
388 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
389 * config/rs6000/8540.md (ppc8540_su): Adjust.
390 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
391 cell-cmp-microcoded): Adjust.
392 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
393 * config/rs6000/e500mc.md (e500mc_su): Adjust.
394 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
395 e500mc64_delayed): Adjust.
396 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
397 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
398 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
399 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
400 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
401 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
402 power6-delayed-compare, power6-var-delayed-compare): Adjust.
403 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
404 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
406 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
407 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
409 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
411 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
412 (bits): New mode_attr.
413 (idiv_ldiv): Delete mode_attr.
414 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
415 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
416 rs6000_adjust_priority, is_nonpipeline_insn,
417 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
419 * config/rs6000/40x.md (ppc403-idiv): Adjust.
420 * config/rs6000/440.md (ppc440-idiv): Adjust.
421 * config/rs6000/476.md (ppc476-idiv): Adjust.
422 * config/rs6000/601.md (ppc601-idiv): Adjust.
423 * config/rs6000/603.md (ppc603-idiv): Adjust.
424 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
425 ppc620-ldiv): Adjust.
426 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
427 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
428 * config/rs6000/8540.md (ppc8540_divide): Adjust.
429 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
430 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
431 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
432 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
433 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
434 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
435 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
436 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
437 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
438 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
439 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
440 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
441 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
442 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
443 * config/rs6000/titan.md (titan_fxu_div): Adjust.
445 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
447 * config/rs6000/rs6000.md (type): Delete "insert_word",
448 "insert_dword". Add "insert".
449 (size): Update comment.
450 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
451 insn_must_be_first_in_group): Adjust.
452 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
453 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
454 *insvsi_internal6, insvdi_internal): Adjust.
456 * config/rs6000/40x.md (ppc403-integer): Adjust.
457 * config/rs6000/440.md (ppc440-integer): Adjust.
458 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
459 * config/rs6000/601.md (ppc601-integer): Adjust.
460 * config/rs6000/603.md (ppc603-integer): Adjust.
461 * config/rs6000/6xx.md (ppc604-integer): Adjust.
462 * config/rs6000/7450.md (ppc7450-integer): Adjust.
463 * config/rs6000/7xx.md (ppc750-integer): Adjust.
464 * config/rs6000/8540.md (ppc8540_su): Adjust.
465 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
466 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
467 * config/rs6000/e500mc.md (e500mc_su): Adjust.
468 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
469 * config/rs6000/e5500.md (e5500_sfx): Adjust.
470 * config/rs6000/e6500.md (e6500_sfx): Adjust.
471 * config/rs6000/mpc.md (mpccore-integer): Adjust.
472 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
473 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
474 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
475 * config/rs6000/power7.md (power7-integer): Adjust.
476 * config/rs6000/power8.md (power8-1cyc): Adjust.
477 * config/rs6000/rs64.md (rs64a-integer): Adjust.
478 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
480 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
482 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
483 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
484 (size): New attribute.
485 (dot): New attribute.
486 (cell_micro): Adjust.
487 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
488 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
489 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
490 umuldi3_highpart): Adjust.
491 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
492 rs6000_adjust_priority, is_nonpipeline_insn,
493 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
495 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
496 ppc405-imul3): Adjust.
497 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
498 * config/rs6000/476.md (ppc476-imul): Adjust.
499 * config/rs6000/601.md (ppc601-imul): Adjust.
500 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
501 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
502 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
503 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
504 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
506 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
507 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
508 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
510 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
511 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
512 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
513 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
514 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
515 * config/rs6000/mpc.md (mpccore-imul): Adjust.
516 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
517 power4-lmul, power4-imul, power4-imul3): Adjust.
518 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
519 power5-lmul, power5-imul, power5-imul3): Adjust.
520 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
521 power6-lmul, power6-imul, power6-imul3): Adjust.
522 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
523 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
525 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
527 * config/rs6000/titan.md (titan_imul): Adjust.
529 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
531 * config/rs6000/rs6000.md (type): Add new value "halfmul".
532 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
533 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
534 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
535 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
536 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
537 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
538 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
539 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
540 * config/rs6000/titan.md: Delete nonsensical comment.
541 (titan_imul): Add type imul3.
542 (titan_mulhw): Remove type imul3; add type halfmul.
544 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
546 * config/rs6000/rs6000.md (type): Reorder, reformat.
548 2014-05-23 Martin Jambor <mjambor@suse.cz>
550 PR tree-optimization/53787
551 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
552 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
553 analysis_done, update all uses.
554 * ipa-prop.c: Include domwalk.h
555 (param_analysis_info): Removed.
556 (param_aa_status): New type.
557 (ipa_bb_info): Likewise.
558 (func_body_info): Likewise.
559 (ipa_get_bb_info): New function.
560 (aa_overwalked): Likewise.
561 (find_dominating_aa_status): Likewise.
562 (parm_bb_aa_status_for_bb): Likewise.
563 (parm_preserved_before_stmt_p): Changed to use new param AA info.
564 (load_from_unmodified_param): Accept func_body_info as a parameter
565 instead of parms_ainfo.
566 (parm_ref_data_preserved_p): Changed to use new param AA info.
567 (parm_ref_data_pass_through_p): Likewise.
568 (ipa_load_from_parm_agg_1): Likewise. Update callers.
569 (compute_complex_assign_jump_func): Changed to use new param AA info.
570 (compute_complex_ancestor_jump_func): Likewise.
571 (ipa_compute_jump_functions_for_edge): Likewise.
572 (ipa_compute_jump_functions): Removed.
573 (ipa_compute_jump_functions_for_bb): New function.
574 (ipa_analyze_indirect_call_uses): Likewise, moved variable
576 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
577 and info, moved variable declarations down.
578 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
580 (ipa_analyze_stmt_uses): Likewise.
581 (ipa_analyze_params_uses): Removed.
582 (ipa_analyze_params_uses_in_bb): New function.
583 (ipa_analyze_controlled_uses): Likewise.
584 (free_ipa_bb_info): Likewise.
585 (analysis_dom_walker): New class.
586 (ipa_analyze_node): Handle node-specific forbidden analysis,
587 initialize and free func_body_info, use dominator walker.
588 (ipcp_modif_dom_walker): New class.
589 (ipcp_transform_function): Create and free func_body_info, use
590 ipcp_modif_dom_walker, moved a lot of functionality there.
592 2014-05-23 Marek Polacek <polacek@redhat.com>
593 Jakub Jelinek <jakub@redhat.com>
595 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
596 * gcc.c (sanitize_spec_function): Likewise.
597 * convert.c (convert_to_integer): Include "ubsan.h". Add
598 floating-point to integer instrumentation.
599 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
600 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
602 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
603 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
604 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
605 * ubsan.c: Include "realmpfr.h" and "dfp.h".
606 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
607 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
608 float/double/long double.
609 (ubsan_instrument_float_cast): New function.
610 * ubsan.h (ubsan_instrument_float_cast): Declare.
612 2014-05-23 Jiong Wang <jiong.wang@arm.com>
614 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
616 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
617 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
618 Adjust for tailcalling through registers.
619 * config/aarch64/aarch64.h (enum reg_class): New caller save
621 (REG_CLASS_NAMES): Likewise.
622 (REG_CLASS_CONTENTS): Likewise.
623 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
624 Allow tailcalling without decls.
626 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
628 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
629 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
631 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
632 gsi, and variables v_* to v*.
634 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
636 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
638 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
640 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
641 * omp-low.c: Update accordingly.
643 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
644 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
645 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
646 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
647 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
648 GF_OMP_TARGET_KIND_UPDATE.
650 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
651 Explicitly enumerate the expected region types.
653 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
656 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
657 documentation; the old documentation didn't clearly state the
658 constraints on the contents of the pointed-to storage.
660 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
662 Fix bootstrap error on ia64
663 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
664 Return default value.
666 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
668 PR tree-optimization/54733
669 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
671 (find_bswap_or_nop_load): New.
672 (find_bswap_1): Renamed to ...
673 (find_bswap_or_nop_1): This. Also add support for memory source.
674 (find_bswap): Renamed to ...
675 (find_bswap_or_nop): This. Also add support for memory source and
676 detection of bitwise operations equivalent to load in host endianness.
677 (execute_optimize_bswap): Likewise. Also move its leading comment back
678 in place and split statement transformation into ...
679 (bswap_replace): This.
681 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
683 PR rtl-optimization/61215
684 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
685 simplify_gen_subreg until final substitution.
687 2014-05-23 Alan Modra <amodra@gmail.com>
690 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
691 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
692 Use "Y" constraint rather than "m".
694 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
696 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
698 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
699 New function declaration.
700 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
701 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
702 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
703 (aarch64_init_builtins) : Initialize builtins
704 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
705 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
706 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
707 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
708 and __builtins_aarch64_set_fpsr.
709 (aarch64_atomic_assign_expand_fenv): New function.
710 * config/aarch64/aarch64.md (set_fpcr): New pattern.
711 (get_fpcr) : Likewise.
712 (set_fpsr) : Likewise.
713 (get_fpsr) : Likewise.
714 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
715 and UNSPECV_SET_FPSR.
716 * doc/extend.texi (AARCH64 Built-in Functions) : Document
717 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
718 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
720 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
722 PR rtl-optimization/60969
723 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
724 constraints. Set up mem cost for NO_REGS case.
726 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
728 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
730 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
732 * config/darwin.c: Include "lto-section-names.h".
733 (LTO_SEGMENT_NAME): Don't define.
734 * config/i386/winnt.c: Include "lto-section-names.h".
735 * lto-streamer.c: Include "lto-section-names.h".
736 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
737 * lto-wrapper.c: Include "lto-section-names.h".
738 (LTO_SECTION_NAME_PREFIX): Don't define.
739 * lto-section-names.h: New file.
740 * cgraphunit.c: Include "lto-section-names.h".
742 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
744 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
746 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
749 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
751 2014-05-22 Nick Clifton <nickc@redhat.com>
753 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
755 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
757 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
758 -> (T)A transformation to integer types.
760 2014-05-22 Teresa Johnson <tejohnson@google.com>
762 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
763 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
764 (gcov_rewrite): Use gcov_nonruntime_assert.
766 (gcov_write_words): Ditto.
767 (gcov_write_length): Ditto.
768 (gcov_read_words): Use gcov_nonruntime_assert, and remove
769 gcc_assert from IN_LIBGCOV code.
770 (gcov_read_summary): Use gcov_error to flag profile corruption.
771 (gcov_sync): Use gcov_nonruntime_assert.
772 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
773 (gcov_histo_index): Use gcov_nonruntime_assert.
774 (static void gcov_histogram_merge): Ditto.
775 (compute_working_sets): Ditto.
776 * gcov-io.h (gcov_nonruntime_assert): Define.
777 (gcov_error): Define for !IN_LIBGCOV
779 2014-05-22 Richard Biener <rguenther@suse.de>
781 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
782 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
783 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
784 and deallocation site.
785 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
786 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
787 passing through the incoming points-to set.
788 (handle_lhs_call): Use flags argument instead of recomputing it.
789 (find_func_aliases_for_call): Call handle_lhs_call with proper
792 2014-05-22 Jakub Jelinek <jakub@redhat.com>
794 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
795 all padding bits in REAL_VALUE_TYPE are cleared.
797 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
799 Cleanup and improve multipass_dfa_lookahead_guard
800 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
801 (core2i7_first_cycle_multipass_begin,)
802 (core2i7_first_cycle_multipass_issue,)
803 (core2i7_first_cycle_multipass_backtrack): Update signature.
805 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
806 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
807 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
809 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
810 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
812 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
814 * doc/tm.texi: Regenerate.
816 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
817 * haifa-sched.c (ready_try): Make signed to allow negative values.
818 (rebug_ready_list_1): Update.
819 (choose_ready): Simplify.
820 (sched_extend_ready_list): Update.
822 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
824 Remove IA64 speculation tweaking flags
825 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
826 speculation tuning flags.
827 (msched-prefer-non-data-spec-insns,)
828 (msched-prefer-non-control-spec-insns): Obsolete options.
829 * haifa-sched.c (choose_ready): Remove handling of
830 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
831 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
832 and PREFER_NON_DATA_SPEC.
833 * sel-sched.c (process_spec_exprs): Remove handling of
834 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
836 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
838 Improve scheduling debug output
839 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
840 (advance_one_cycle): Update.
841 (schedule_insn, queue_to_ready): Add debug printouts.
842 (debug_ready_list_1): New static function.
843 (debug_ready_list): Update.
844 (max_issue): Add debug printouts.
845 (dump_insn_stream): New static function.
846 (schedule_block): Use it. Also better indent printouts.
848 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
850 Fix sched_insn debug counter
851 * haifa-sched.c (schedule_insn): Update.
852 (struct haifa_saved_data): Add nonscheduled_insns_begin.
853 (save_backtrack_point, restore_backtrack_point): Update.
854 (first_nonscheduled_insn): New static function.
855 (queue_to_ready, choose_ready): Use it.
856 (schedule_block): Init nonscheduled_insns_begin.
857 (sched_emit_insn): Update.
860 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
862 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
864 (aarch64_secondary_reload) : LikeWise.
865 (aarch64_class_max_nregs) : Remove CORE_REGS.
866 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
867 (REG_CLASS_NAMES) : Likewise.
868 (REG_CLASS_CONTENTS) : LikeWise.
869 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
871 2014-05-21 Guozhi Wei <carrot@google.com>
874 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
876 (vqdmulhq_n_s16): Likewise.
878 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
880 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
882 2014-05-21 Marek Polacek <polacek@redhat.com>
885 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
887 2014-05-21 Martin Jambor <mjambor@suse.cz>
889 * doc/invoke.texi (Optimize Options): Document parameters
890 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
891 ipa-cp-array-index-hint-bonus.
893 2014-05-21 Mark Wielaard <mjw@redhat.com>
896 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
897 version >= 3 or not strict DWARF.
898 * langhooks.h (struct lang_hooks_for_types): Add
899 enum_underlying_base_type.
900 * langhooks.c (lhd_enum_underlying_base_type): New function.
901 * gcc/langhooks.h (struct lang_hooks_for_types): Add
902 enum_underlying_base_type.
903 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
904 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
905 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
907 2014-05-21 Richard Biener <rguenther@suse.de>
909 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
911 2014-05-21 John Marino <gnugcc@marino.st>
913 * config.gcc (*-*-dragonfly*): New target.
914 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
915 * configure: Regenerate.
916 * config/dragonfly-stdint.h: New.
917 * config/dragonfly.h: New.
918 * config/dragonfly.opt: New.
919 * config/i386/dragonfly.h: New.
920 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
922 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
924 * tree.def (VOID_CST): New.
925 * tree-core.h (TI_VOID): New.
926 * tree.h (void_node): New.
927 * tree.c (tree_node_structure_for_code, tree_code_size)
928 (iterative_hash_expr): Handle VOID_CST.
929 (build_common_tree_nodes): Initialize void_node.
931 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
933 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
935 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
937 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
940 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
941 flag_reorder_blocks_and_partition.
942 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
944 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
947 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
949 (*addc_r_t): Add new insn_and_split.
951 2014-05-21 Jakub Jelinek <jakub@redhat.com>
954 * omp-low.c (handle_simd_reference): New function.
955 (lower_rec_input_clauses): Use it. Defer adding reference
956 initialization even for reduction without placeholder if in simd,
957 handle it properly later on.
959 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
961 PR tree-optimization/60899
962 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
963 assume all static symbols will have definition wile parsing and
964 check the do have definition later in compilation; check that
965 variable referring symbol will be output before concluding that
966 reference is safe; be conservative for referring local statics;
967 be more precise about when comdat is output in other partition.
969 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
972 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
974 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
975 (ipa_inline): Loop inline_to_all_callers until no more aliases
978 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
980 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
981 set writeonly flag only for vars actually written to.
983 2014-05-20 Dehao Chen <dehao@google.com>
985 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
986 and callee count to get clone count.
987 * tree-inline.c (expand_call_inline): Use callee count instead of bb
990 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
992 PR rtl-optimization/61243
993 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
995 2014-05-20 Xinliang David Li <davidxl@google.com>
997 * cgraphunit.c (walk_polymorphic_call_targets): Add
998 dbgcnt and fopt-info support.
999 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1000 * ipa-devirt.c (ipa_devirt): Ditto.
1001 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1002 * ipa.c (walk_polymorphic_call_targets): Ditto.
1003 * gimple-fold.c (fold_gimple_assign): Ditto.
1004 (gimple_fold_call): Ditto.
1005 * dbgcnt.def: New counter.
1007 2014-05-20 DJ Delorie <dj@redhat.com>
1009 * config/msp430/msp430.md (split): Don't allow subregs when
1010 splitting SImode adds.
1011 (andneghi): Fix subtraction logic.
1012 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1014 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1016 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1018 * except.c (switch_to_exception_section, resolve_unique_section,
1019 get_named_text_section, default_function_rodata_section,
1020 align_variable, get_block_for_decl, default_section_type_flags):
1021 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1022 * symtab.c (symtab_add_to_same_comdat_group,
1023 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1024 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1026 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1027 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1028 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1029 (c6x_function_in_section_p): Likewise.
1030 * config/darwin.c (machopic_select_section): Likewise.
1031 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1032 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1033 * config/mep/mep.c (mep_select_section): LIkewise.
1034 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1036 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1038 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1039 EH region of calls to pure functions that can throw an exception.
1040 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1041 (copy_reference_ops_from_call): Also copy the EH region of the call if
1042 it can throw an exception.
1044 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1046 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1047 nested VEC_SELECTs that are inverses of each other.
1049 2014-05-20 Richard Biener <rguenther@suse.de>
1051 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1052 (extract_and_process_scc_for_name): not here.
1053 (cond_dom_walker::before_dom_children): Only process
1054 stmts that end the BB in interesting ways.
1055 (run_scc_vn): Mark param uses as visited.
1057 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1059 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1062 2014-05-20 Nick Clifton <nickc@redhat.com>
1064 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1065 (msp430_gimplify_va_arg_expr): New function.
1066 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1068 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1069 operand 0 in order to prevent confusion about the number of
1072 2014-05-20 Richard Biener <rguenther@suse.de>
1074 PR tree-optimization/61221
1075 * tree-ssa-pre.c (el_to_update): Remove.
1076 (eliminate_dom_walker::before_dom_children): Handle released
1077 VDEFs by value-numbering them to the associated VUSE. Update
1078 stmt immediately for substituted call address.
1079 (eliminate): Remove delayed stmt updating code.
1080 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1081 possibly late re-numbered vuses.
1082 (vn_reference_lookup_2): Adjust.
1083 (vn_reference_lookup_pieces): Likewise.
1084 (vn_reference_lookup): Likewise.
1086 2014-05-20 Richard Biener <rguenther@suse.de>
1088 * config.gcc: Remove need_64bit_hwint.
1089 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1090 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1092 * config.in: Regenerate.
1093 * configure: Likewise.
1095 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1097 * doc/extend.texi: Create Label Attributes section,
1098 move all label attributes into it and reference it.
1100 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1102 * arm.c (thumb1_reorg): When scanning backwards skip anything
1103 that's not a proper insn.
1105 2014-05-19 Richard Biener <rguenther@suse.de>
1107 PR tree-optimization/61221
1108 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1109 Do nothing for unreachable blocks.
1110 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1111 Improve unreachability detection.
1113 2014-05-19 Richard Biener <rguenther@suse.de>
1115 PR tree-optimization/61209
1116 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1118 2014-05-19 Nick Clifton <nickc@redhat.com>
1120 * except.c (init_eh): Fix computation of builtin setjmp buffer
1121 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1123 2014-05-19 Richard Biener <rguenther@suse.de>
1125 PR tree-optimization/61184
1126 * tree-vrp.c (is_negative_overflow_infinity): Use
1127 TREE_OVERFLOW_P and do that check first.
1128 (is_positive_overflow_infinity): Likewise.
1129 (is_overflow_infinity): Likewise.
1130 (vrp_operand_equal_p): Properly treat operands with
1131 differing overflow as not equal.
1133 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1135 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1136 shift simplification where it was intended.
1138 2014-05-19 Christian Bruel <christian.bruel@st.com>
1141 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1143 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1146 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1149 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1151 * reg-notes.def (CROSSING_JUMP): Likewise.
1152 * rtl.h (rtx_def): Update comment for jump flag.
1153 (CROSSING_JUMP_P): Define.
1154 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1155 of a REG_CROSSING_JUMP note.
1156 * cfghooks.c (tidy_fallthru_edges): Likewise.
1157 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1158 * emit-rtl.c (try_split): Likewise.
1159 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1160 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1161 * jump.c (redirect_jump_2): Likewise.
1162 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1163 (relax_delay_slots): Likewise.
1164 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1165 (bbit_di): Likewise.
1166 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1167 * config/sh/sh.md (jump_compact): Likewise.
1168 * bb-reorder.c (rotate_loop): Likewise.
1169 (pass_duplicate_computed_gotos::execute): Likewise.
1170 (add_reg_crossing_jump_notes): Rename to...
1171 (update_crossing_jump_flags): ...this.
1172 (pass_partition_blocks::execute): Update accordingly.
1174 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1176 * tree.h: Remove extraneous template <>.
1178 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1180 * ipa.c (symtab_remove_unreachable_nodes): Remove
1181 symbol from comdat group if its body was eliminated.
1182 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1183 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1184 (symtab_unregister_node): ... this one.
1185 (verify_symtab_base): More strict checking of comdats.
1186 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1188 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1190 * tree-pass.h (make_pass_ipa_comdats): New pass.
1191 * timevar.def (TV_IPA_COMDATS): New timevar.
1192 * passes.def (pass_ipa_comdats): Add.
1193 * Makefile.in (OBJS): Add ipa-comdats.o
1194 * ipa-comdats.c: New file.
1196 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1198 * ipa.c (update_visibility_by_resolution_info): New function.
1199 (function_and_variable_visibility): Use it.
1201 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1203 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1205 (FOR_EACH_DEFINED_SYMBOL): New macro.
1206 (varpool_first_static_initializer, varpool_next_static_initializer,
1207 varpool_first_defined_variable, varpool_next_defined_variable):
1209 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1211 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1213 * ggc-page.c (ggc_handle_finalizers): Add comment.
1215 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1217 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1218 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1219 (ggc_internal_cleared_alloc): Likewise.
1220 * ggc-page.c (finalizer): New class.
1221 (vec_finalizer): Likewise.
1222 (globals::finalizers): New member.
1223 (globals::vec_finalizers): Likewise.
1224 (ggc_internal_alloc): Record the finalizer if any for the block being
1226 (ggc_handle_finalizers): New function.
1227 (ggc_collect): Call ggc_handle_finalizers.
1228 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1230 (ggc_internal_cleared_alloc): Likewise.
1231 (finalize): New function.
1232 (need_finalization_p): Likewise.
1233 (ggc_alloc): Install the type's destructor as the finalizer if it
1235 (ggc_cleared_alloc): Likewise.
1236 (ggc_vec_alloc): Likewise.
1237 (ggc_cleared_vec_alloc): Likewise.
1239 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1241 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1243 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1245 * alias.c (record_alias_subset): Adjust.
1246 * bitmap.c (bitmap_element_allocate): Likewise.
1247 (bitmap_gc_alloc_stat): Likewise.
1248 * cfg.c (init_flow): Likewise.
1249 (alloc_block): Likewise.
1250 (unchecked_make_edge): Likewise.
1251 * cfgloop.c (alloc_loop): Likewise.
1252 (flow_loops_find): Likewise.
1253 (rescan_loop_exit): Likewise.
1254 * cfgrtl.c (init_rtl_bb_info): Likewise.
1255 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1256 (cgraph_allocate_node): Likewise.
1257 (cgraph_create_edge_1): Likewise.
1258 (cgraph_allocate_init_indirect_info): Likewise.
1259 * cgraphclones.c (cgraph_clone_edge): Likewise.
1260 * cgraphunit.c (add_asm_node): Likewise.
1261 (init_lowered_empty_function): Likewise.
1262 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1263 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1264 (alpha_use_linkage): Likewise.
1265 * config/arc/arc.c (arc_init_machine_status): Likewise.
1266 * config/arm/arm.c (arm_init_machine_status): Likewise.
1267 * config/avr/avr.c (avr_init_machine_status): Likewise.
1268 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1269 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1270 * config/cris/cris.c (cris_init_machine_status): Likewise.
1271 * config/darwin.c (machopic_indirection_name): Likewise.
1272 (darwin_build_constant_cfstring): Likewise.
1273 (darwin_enter_string_into_cfstring_table): Likewise.
1274 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1275 * config/frv/frv.c (frv_init_machine_status): Likewise.
1276 * config/i386/i386.c (get_dllimport_decl): Likewise.
1277 (ix86_init_machine_status): Likewise.
1278 (assign_386_stack_local): Likewise.
1279 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1280 (i386_pe_maybe_record_exported_symbol): Likewise.
1281 (i386_pe_record_stub): Likewise.
1282 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1283 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1284 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1285 (m32c_note_pragma_address): Likewise.
1286 * config/mep/mep.c (mep_init_machine_status): Likewise.
1287 (mep_note_pragma_flag): Likewise.
1288 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1289 (mips16_local_alias): Likewise.
1290 (mips_init_machine_status): Likewise.
1291 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1292 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1293 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1294 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1295 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1296 * config/pa/pa.c (pa_init_machine_status): Likewise.
1297 (pa_get_deferred_plabel): Likewise.
1298 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1299 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1300 (rs6000_init_machine_status): Likewise.
1301 (output_toc): Likewise.
1302 * config/s390/s390.c (s390_init_machine_status): Likewise.
1303 * config/score/score.c (score_output_external): Likewise.
1304 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1305 * config/spu/spu.c (spu_init_machine_status): Likewise.
1306 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1307 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1308 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1309 * coverage.c (coverage_end_function): Likewise.
1310 * dbxout.c (dbxout_init): Likewise.
1311 * doc/gty.texi: Don't mention variable_size attribute.
1312 * dwarf2cfi.c (new_cfi): Adjust.
1313 (new_cfi_row): Likewise.
1314 (copy_cfi_row): Likewise.
1315 (create_cie_data): Likewise.
1316 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1317 (new_loc_descr): Likewise.
1318 (find_AT_string_in_table): Likewise.
1319 (add_addr_table_entry): Likewise.
1320 (new_die): Likewise.
1321 (add_var_loc_to_decl): Likewise.
1322 (clone_die): Likewise.
1323 (clone_as_declaration): Likewise.
1324 (break_out_comdat_types): Likewise.
1325 (new_loc_list): Likewise.
1326 (add_loc_descr_to_each): Likewise.
1327 (add_location_or_const_value_attribute): Likewise.
1328 (add_linkage_name): Likewise.
1329 (lookup_filename): Likewise.
1330 (dwarf2out_var_location): Likewise.
1331 (new_line_info_table): Likewise.
1332 (dwarf2out_init): Likewise.
1333 (mem_loc_descriptor): Likewise.
1334 (loc_descriptor): Likewise.
1335 (add_const_value_attribute): Likewise.
1336 (tree_add_const_value_attribute): Likewise.
1337 (comp_dir_string): Likewise.
1338 (dwarf2out_vms_debug_main_pointer): Likewise.
1339 (string_cst_pool_decl): Likewise.
1340 * emit-rtl.c (set_mem_attrs): Likewise.
1341 (get_reg_attrs): Likewise.
1342 (start_sequence): Likewise.
1343 (init_emit): Likewise.
1344 (init_emit_regs): Likewise.
1345 * except.c (init_eh_for_function): Likewise.
1346 (gen_eh_region): Likewise.
1347 (gen_eh_region_catch): Likewise.
1348 (gen_eh_landing_pad): Likewise.
1349 (add_call_site): Likewise.
1350 * function.c (add_frame_space): Likewise.
1351 (insert_temp_slot_address): Likewise.
1352 (assign_stack_temp_for_type): Likewise.
1353 (get_hard_reg_initial_val): Likewise.
1354 (allocate_struct_function): Likewise.
1355 (prepare_function_start): Likewise.
1356 (types_used_by_var_decl_insert): Likewise.
1357 * gengtype.c (variable_size_p): Remove function.
1358 (enum alloc_quantity): Remove enum.
1359 (write_typed_alloc_def): Remove function.
1360 (write_typed_struct_alloc_def): Likewise.
1361 (write_typed_typedef_alloc_def): Likewise.
1362 (write_typed_alloc_defns): Likewise.
1364 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1365 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1366 * ggc.h (ggc_alloc): new function.
1367 (ggc_cleared_alloc): Likewise.
1368 (ggc_vec_alloc): Template on type of vector element, and remove
1369 element size argument.
1370 (ggc_cleared_vec_alloc): Likewise.
1371 * gimple.c (gimple_build_omp_for): Adjust.
1372 (gimple_copy): Likewise.
1373 * ipa-cp.c (get_replacement_map): Likewise.
1374 (find_aggregate_values_for_callers_subset): Likewise.
1375 (known_aggs_to_agg_replacement_list): Likewise.
1376 * ipa-devirt.c (get_odr_type): Likewise.
1377 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1378 (read_agg_replacement_chain): Likewise.
1379 * loop-iv.c (get_simple_loop_desc): Likewise.
1380 * lto-cgraph.c (input_node_opt_summary): Likewise.
1381 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1382 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1383 (input_eh_region): Likewise.
1384 (input_eh_lp): Likewise.
1385 (input_cfg): Likewise.
1386 * optabs.c (set_optab_libfunc): Likewise.
1387 (init_tree_optimization_optabs): Likewise.
1388 (set_conv_libfunc): Likewise.
1389 * passes.c (do_per_function_toporder): Likewise.
1390 * rtl.h: Don't use variable_size gty attribute.
1391 * sese.c (if_region_set_false_region): Adjust.
1392 * stringpool.c (gt_pch_save_stringpool): Likewise.
1393 * target-globals.c (save_target_globals): Likewise.
1394 * toplev.c (general_init): Likewise.
1395 * trans-mem.c (record_tm_replacement): Likewise.
1396 (split_bb_make_tm_edge): Likewise.
1397 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1398 * tree-data-ref.h (lambda_vector_new): Likewise.
1399 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1400 * tree-iterator.c (tsi_link_before): Likewise.
1401 (tsi_link_after): Likewise.
1402 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1403 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1404 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1405 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1406 * tree-ssa.c (init_tree_ssa): Adjust.
1407 * tree-ssanames.c (set_range_info): Likewise.
1408 (get_ptr_info): Likewise.
1409 (duplicate_ssa_name_ptr_info): Likewise.
1410 (duplicate_ssa_name_range_info): Likewise.
1411 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1412 (unpack_ts_fixed_cst_value_fields): Likewise.
1413 * tree.c (build_fixed): Likewise.
1414 (build_real): Likewise.
1415 (build_string): Likewise.
1416 (decl_priority_info): Likewise.
1417 (decl_debug_expr_insert): Likewise.
1418 (decl_value_expr_insert): Likewise.
1419 (decl_debug_args_insert): Likewise.
1420 (type_hash_add): Likewise.
1421 (build_omp_clause): Likewise.
1422 * ubsan.c (decl_for_type_insert): Likewise.
1423 * varasm.c (get_unnamed_section): Likewise.
1424 (get_noswitch_section): Likewise.
1425 (get_section): Likewise.
1426 (get_block_for_section): Likewise.
1427 (create_block_symbol): Likewise.
1428 (build_constant_desc): Likewise.
1429 (create_constant_pool): Likewise.
1430 (force_const_mem): Likewise.
1431 (record_tm_clone_pair): Likewise.
1432 * varpool.c (varpool_create_empty_node): Likewise.
1434 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1436 * dwarf2out.c (tree_add_const_value_attribute): Call
1437 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1438 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1439 instead of ggc_internal_<x>alloc_stat.
1440 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1441 (ggc_realloc): Likewise.
1442 * ggc-none.c (ggc_internal_alloc): Likewise.
1443 (ggc_internal_cleared_alloc): Likewise.
1444 * ggc-page.c: Likewise.
1445 * ggc.h (ggc_internal_alloc_stat): Likewise.
1446 (ggc_internal_alloc): Remove macro.
1447 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1448 (ggc_internal_cleared_alloc): Remove macro.
1449 (GGC_RESIZEVEC): Adjust.
1450 (ggc_resizevar): Remove macro.
1451 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1452 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1453 (ggc_internal_vec_cleared_alloc): Remove macro.
1454 (ggc_alloc_atomic_stat): Drop _stat suffix.
1455 (ggc_alloc_atomic): Remove macro.
1456 (ggc_alloc_cleared_atomic): Remove macro.
1457 (ggc_alloc_string_stat): Drop _stat suffix.
1458 (ggc_alloc_string): Remove macro.
1459 (ggc_alloc_rtx_def_stat): Adjust.
1460 (ggc_alloc_tree_node_stat): Likewise.
1461 (ggc_alloc_cleared_tree_node_stat): Likewise.
1462 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1463 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1464 * gimple.c (gimple_build_omp_for): Likewise.
1465 (gimple_copy): Likewise.
1466 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1467 * toplev.c (realloc_for_line_map): Adjust.
1468 * tree-data-ref.h (lambda_vector_new): Likewise.
1469 * tree-phinodes.c (allocate_phi_node): Likewise.
1470 * tree.c (grow_tree_vec_stat): Likewise.
1471 * vec.h (va_gc::reserve): Adjust.
1473 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1475 * config/microblaze/microblaze.c (break_handler): New Declaration.
1476 (microblaze_break_function_p,microblaze_is_break_handler): New.
1477 (compute_frame_size): Use microblaze_break_function_p.
1478 Add the test of break_handler.
1479 (microblaze_function_prologue) : Add the test of variable
1480 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1481 (microblaze_function_epilogue) : Add the test of break_handler.
1482 (microblaze_globalize_label) : Add the test of break_handler.
1483 Check the name by BREAK_HANDLER_NAME.
1485 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1487 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1488 microblaze_is_break_handler test.
1489 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1490 Use SYMBOL_REF_DECL.
1492 * config/microblaze/microblaze-protos.h
1493 (microblaze_break_function_p,microblaze_is_break_handler):
1496 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1497 new MicroBlaze break_handler functions.
1499 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1501 * doc/extend.texi (Size of an asm): Move node text according
1502 to its @menu entry position.
1504 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1506 PR tree-optimization/61140
1507 PR tree-optimization/61150
1508 PR tree-optimization/61197
1509 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1511 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1513 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1515 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1517 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1518 __SIZEOF_INT128__ is defined.
1520 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1522 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1523 (rs6000_delegitimize_address): Use it.
1525 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1527 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1528 inplace argument. Store the new address in the original MEM when true.
1529 * emit-rtl.c (change_address_1): Likewise.
1530 (adjust_address_1, adjust_automodify_address_1, offset_address):
1532 * rtl.h (plus_constant): Add an inplace argument.
1533 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1534 when true. Avoid generating (plus X (const_int 0)).
1535 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1536 in-place. Pass true to plus_constant.
1537 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1539 2014-05-16 Dehao Chen <dehao@google.com>
1541 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1543 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1546 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1548 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1550 2014-05-16 Dehao Chen <dehao@google.com>
1552 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1553 optimize_function_for_size_p.
1554 * regs.h (REG_FREQ_FROM_BB): Likewise.
1556 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1559 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1560 negt_reg_operand cases.
1561 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1563 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1565 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1567 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1570 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1573 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1575 * tree-cfg.c (dump_function_to_file): Dump the return type of
1576 functions, in a line to itself before the function body, mimicking
1577 the layout of a C function.
1579 2014-05-16 Dehao Chen <dehao@google.com>
1581 * cfghooks.c (make_forwarder_block): Use direct computation to
1582 get fall-through edge's count and frequency.
1584 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1586 * config/arc/arc.c (arc_init): Fix typo in error message.
1587 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1588 (split_stack_prologue_scratch_regno): Likewise.
1589 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1590 word from error message.
1592 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1594 * ira-costs.c: Fix typo in comment.
1596 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1598 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1600 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1602 * varpool.c (dump_varpool_node): Dump write-only flag.
1603 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1605 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1606 write-only variables.
1607 * ipa.c (process_references): New function.
1608 (set_readonly_bit): New function.
1609 (set_writeonly_bit): New function.
1610 (clear_addressable_bit): New function.
1611 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1612 fix handling of aliases.
1613 * cgraph.h (struct varpool_node): Add writeonly flag.
1615 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1617 PR rtl-optimization/60969
1618 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1619 Calculate costs for this case.
1621 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1623 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1624 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1626 2014-05-16 Richard Biener <rguenther@suse.de>
1628 PR tree-optimization/61194
1629 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1630 bool patterns ending in a COND_EXPR.
1632 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1634 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1636 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1638 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1639 where we were unable to cost an RTX.
1641 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1643 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1646 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1647 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1649 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1651 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1652 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1654 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1655 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1657 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1658 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1660 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1663 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1664 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1666 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
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_arith_op_extract_p): New.
1673 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
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): Improve costs for
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): Cost
1685 ZERO_EXTEND and SIGN_EXTEND better.
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_costs): Improve cost for
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): Use address
1697 costs when costing loads and stores to memory.
1699 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1700 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1702 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1705 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1707 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1709 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1710 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1712 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1714 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1716 (aarch64_rtx_mult_cost): New.
1717 (aarch64_rtx_costs): Use it, refactor as appropriate.
1719 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1720 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1722 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1723 emit instructions, return number of instructions which would
1725 (aarch64_add_constant): Update call to aarch64_build_constant.
1726 (aarch64_output_mi_thunk): Likewise.
1727 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1730 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1732 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1733 (TARGET_RTX_COSTS): Call it.
1735 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1737 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1738 (cortexa57_vector_cost): Likewise.
1739 (cortexa57_tunings): Use them.
1741 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1743 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1744 (cpu_addrcost_table): Use it.
1745 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1746 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1749 2014-05-16 Richard Biener <rguenther@suse.de>
1751 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1752 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1753 (visit_phi): Ignore edges marked as not executable.
1754 (class cond_dom_walker): New.
1755 (cond_dom_walker::before_dom_children): Value-number
1756 control statements and mark successor edges as not
1757 executable if possible.
1758 (run_scc_vn): First walk all control statements in
1759 dominator order, marking edges as not executable.
1760 * tree-inline.c (copy_edges_for_bb): Be not confused
1761 about random edge flags.
1763 2014-05-16 Richard Biener <rguenther@suse.de>
1765 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1767 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1770 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1771 (__TM_simple_begin): Use it.
1772 (__TM_begin): Likewise.
1774 2014-05-15 Martin Jambor <mjambor@suse.cz>
1777 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1778 type_preserved flag when the indirect edge is polymorphic.
1780 2014-05-15 Martin Jambor <mjambor@suse.cz>
1782 PR tree-optimization/61090
1783 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1784 build_ref_for_model.
1786 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1788 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1789 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1791 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1793 PR tree-optimization/61158
1794 * fold-const.c (fold_binary_loc): If X is zero-extended and
1795 shiftc >= prec, make sure zerobits is all ones instead of
1796 invoking undefined behavior.
1798 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1800 * regcprop.h: New file.
1801 * regcprop.c (skip_debug_insn_p): New decl.
1802 (replace_oldest_value_reg): Check skip_debug_insn_p.
1803 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1804 * shrink-wrap.c: Include regcprop.h.
1805 (prepare_shrink_wrap): Call
1806 copyprop_hardreg_forward_bb_without_debug_insn.
1808 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1810 * shrink-wrap.h: Update comment.
1811 * shrink-wrap.c: Update comment.
1812 (next_block_for_reg): Rename to live_edge_for_reg.
1813 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1814 (move_insn_for_shrink_wrap): Split live_edge.
1815 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1817 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1819 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1821 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1822 * config/sparc/sparc.md (fptype_ut699): New attribute.
1823 (in_branch_delay): Return false if -mfix-ut699 is specified and
1824 fptype_ut699 is set to single.
1825 (truncdfsf2): Add fptype_ut699 attribute.
1826 (fix_truncdfsi2): Likewise.
1827 (floatsisf2): Change fptype attribute.
1828 (fix_truncsfsi2): Likewise.
1829 (negtf2_notv9): Delete.
1830 (negtf2_v9): Likewise.
1831 (negtf2_hq): New instruction.
1832 (negtf2): New instruction and splitter.
1833 (negdf2_notv9): Rewrite.
1834 (abstf2_notv9): Delete.
1835 (abstf2_hq_v9): Likewise.
1836 (abstf2_v9): Likewise.
1837 (abstf2_hq): New instruction.
1838 (abstf2): New instruction and splitter.
1839 (absdf2_notv9): Rewrite.
1841 2014-05-14 Cary Coutant <ccoutant@google.com>
1844 * opts.c (common_handle_option): Don't special-case "-g".
1845 (set_debug_level): Default to at least level 2 with "-g".
1847 2014-05-14 DJ Delorie <dj@redhat.com>
1849 * config/msp430/msp430.c (msp430_builtin): Add
1850 MSP430_BUILTIN_DELAY_CYCLES.
1851 (msp430_init_builtins): Register void __delay_cycles(long long).
1852 (msp430_builtin_decl): Add it.
1853 (cg_magic_constant): New.
1854 (msp430_expand_delay_cycles): New.
1855 (msp430_expand_builtin): Call it.
1856 (msp430_print_operand_raw): Change integer printing from "int" to
1858 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1859 (delay_cycles_start): New.
1860 (delay_cycles_end): New.
1861 (delay_cycles_32): New.
1862 (delay_cycles_32x): New.
1863 (delay_cycles_16): New.
1864 (delay_cycles_16x): New.
1865 (delay_cycles_2): New.
1866 (delay_cycles_1): New.
1867 * doc/extend.texi: Document __delay_cycles().
1869 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1871 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1872 length attribute computation.
1874 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
1877 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
1879 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
1882 * config/sparc/sparc.md: Fix types of low and high in DI constant
1883 splitter. Use gen_int_mode in some other splitters.
1885 2014-05-14 Martin Jambor <mjambor@suse.cz>
1888 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
1890 2014-05-14 James Norris <jnorris@codesourcery.com>
1892 * omp-low.c (expand_parallel_call): Remove shadow variable.
1893 (expand_omp_taskreg): Likewise.
1895 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
1897 * common/config/i386/i386-common.c
1898 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
1899 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
1900 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
1901 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
1902 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
1903 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
1904 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
1905 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
1906 xsavecintrin.h, xsavesintrin.h.
1907 (x86_64-*-*): Ditto.
1908 * config/i386/clflushoptintrin.h: New.
1909 * config/i386/xsavecintrin.h: Ditto.
1910 * config/i386/xsavesintrin.h: Ditto.
1911 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
1912 (bit_XSAVES): Ditto.
1913 (bit_XSAVES): Ditto.
1914 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
1915 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
1917 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
1918 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
1919 OPTION_MASK_ISA_XSAVES.
1920 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
1922 (PTA_CLFLUSHOPT) Define.
1923 (PTA_XSAVEC): Ditto.
1924 (PTA_XSAVES): Ditto.
1925 (ix86_option_override_internal): Handle new options.
1926 (ix86_valid_target_attribute_inner_p): Ditto.
1927 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
1928 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
1929 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
1930 (bdesc_special_args): Add __builtin_ia32_xsaves,
1931 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
1932 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
1933 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
1934 (ix86_expand_builtin): Handle new builtins.
1935 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
1936 (TARGET_CLFLUSHOPT_P): Ditto.
1937 (TARGET_XSAVEC): Ditto.
1938 (TARGET_XSAVEC_P): Ditto.
1939 (TARGET_XSAVES): Ditto.
1940 (TARGET_XSAVES_P): Ditto.
1941 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
1942 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
1943 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
1945 (ANY_XRSTOR64): Ditto.
1947 (xrstor): Change into <xrstor>.
1948 (xrstor_rex64): Change into <xrstor>_rex64.
1949 (xrstor64): Change into <xrstor>64
1951 * config/i386/i386.opt (mclflushopt): New.
1954 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
1956 * doc/invoke.texi: Document new options.
1958 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1960 PR rtl-optimization/60866
1961 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
1962 Default it to -1. Pass it down to init_simplejump_data.
1963 (init_simplejump_data): New parameter old_seqno. Pass it down
1964 to get_seqno_for_a_jump.
1965 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
1966 initializing new jump seqno as a last resort. Add comment.
1967 (sel_redirect_edge_and_branch): Save old seqno of the conditional
1968 jump and pass it down to sel_init_new_insn.
1969 (sel_redirect_edge_and_branch_force): Likewise.
1971 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
1973 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
1974 shifted values to avoid build warning.
1976 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1978 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
1979 * cfgrtl.c (rtl_merge_blocks): Fix comment.
1980 (cfg_layout_merge_blocks): Likewise.
1981 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
1983 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
1985 PR rtl-optimization/60901
1986 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
1987 bb predecessor belongs to the same scheduling region. Adjust comment.
1989 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
1991 * doc/sourcebuild.texi: (dfp_hw): Document.
1992 (p8vector_hw): Likewise.
1993 (powerpc_eabi_ok): Likewise.
1994 (powerpc_elfv2): Likewise.
1995 (powerpc_htm_ok): Likewise.
1996 (ppc_recip_hw): Likewise.
1999 2014-05-13 Cary Coutant <ccoutant@google.com>
2001 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2003 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2005 * gengtype-parse.c (require3): Eliminate in favor of...
2007 (require_template_declaration): Update to support optional single *
2010 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2011 (create_user_defined_type): Handle a single level of explicit
2012 pointerness within template arguments.
2013 (struct write_types_data): Add field "kind".
2014 (filter_type_name): Handle "*" character.
2015 (write_user_func_for_structure_ptr): Require a write_types_data
2016 rather than just a prefix string, so that we can look up the kind
2017 of the wtd and use it as an index into wrote_user_func_for_ptr,
2018 ensuring that such functions are written at most once. Support
2019 subclasses by invoking the marking function of the ultimate base class.
2020 (write_user_func_for_structure_body): Require a write_types_data
2021 rather than just a prefix string, so that we can pass this to
2022 write_user_func_for_structure_ptr.
2023 (write_func_for_structure): Likewise.
2024 (ggc_wtd): Add initializer of new "kind" field.
2025 (pch_wtd): Likewise.
2027 * gengtype.h (enum write_types_kinds): New.
2028 (struct type): Add field wrote_user_func_for_ptr to the "s"
2031 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2033 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2034 instead of const_binop.
2035 (fold_binary_loc): Likewise.
2037 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2039 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2040 calculation to match get_ref_base_and_extent.
2042 2014-05-13 Catherine Moore <clm@codesourcery.com>
2043 Sandra Loosemore <sandra@codesourcery.com>
2045 * configure.ac: Fix assembly for explicit JALR relocation check.
2046 * configure: Regenerate.
2048 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2050 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2051 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2052 Remove associated type declarations and initialisations.
2053 (arm_expand_neon_builtin): Likewise.
2054 (neon_emit_pair_result_insn): Delete.
2055 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2056 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2057 (neon_vzip<mode>): Likewise.
2058 (neon_vuzp<mode>): Likewise.
2060 2014-05-13 Richard Biener <rguenther@suse.de>
2063 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2064 it needs revisiting whether the call still may be tail-called.
2066 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2068 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2069 * rtl.h (block_symbol): Reduce number of fields to 2.
2070 (rtx_def): Add u2.symbol_ref_flags.
2071 (SYMBOL_REF_FLAGS): Use it.
2072 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2073 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2074 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2075 Lower index of SYMBOL_REF_DATA.
2076 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2077 Print SYMBOL_REF_FLAGS at the same time.
2078 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2080 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2082 * rtl.def (VAR_LOCATION): Remove "i" field.
2083 * rtl.h (rtx_def): Add u2.var_location_status.
2084 (PAT_VAR_LOCATION_STATUS): Use it.
2085 (gen_rtx_VAR_LOCATION): Declare.
2086 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2087 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2088 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2090 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2092 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2093 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2095 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2097 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2098 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2099 * rtl.h (rtx_def): Add insn_uid to u2 field.
2100 (RTX_FLAG_CHECK8): Delete in favor of...
2101 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2102 (INSN_DELETED_P): Update accordingly.
2103 (INSN_UID): Use u2.insn_uid.
2104 (INSN_CHAIN_CODE_P): Define.
2105 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2106 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2107 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2108 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2109 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2110 indices accordingly.
2111 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2112 Update indices for insn-chain rtxes.
2113 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2114 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2115 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2116 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2117 * combine.c (try_combine): Likewise.
2118 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2120 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2122 * rtl.def (REG): Remove middle field.
2123 * rtl.h (rtx_def): Add orignal_regno to u2.
2124 (ORIGINAL_REGNO): Use it instead of field 1.
2125 (REG_ATTRS): Lower field index accordingly.
2126 * gengtype.c (adjust_field_rtx_def): Remove handling of
2127 ORIGINAL_REGNO. Move REG_ATTRS index down.
2128 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2129 code that prints the REGNO.
2131 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2133 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2136 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2138 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2140 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2142 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2143 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2145 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2147 * config/aarch64/aarch64-protos.h
2148 (aarch64_hard_regno_caller_save_mode): New prototype.
2149 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2151 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2153 2014-05-13 Christian Bruel <christian.bruel@st.com>
2155 * target.def (mode_switching): New hook vector.
2156 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2157 (mode_exit, modepriority_to_mode): Likewise.
2158 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2159 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2160 * target.h: Include tm.h and hard-reg-set.h.
2161 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2162 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2163 * doc/tm.texi Regenerate.
2164 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2165 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2166 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2167 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2168 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2169 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2170 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2171 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2172 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2173 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2174 (ix86_emit_mode_set): Hookify.
2175 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2177 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2178 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2179 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2180 (epiphany_mode_priority_to_mode): Remove declaration.
2181 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2182 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2183 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2185 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2186 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2187 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2189 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2192 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2193 is const0_rtx, return immediately. Don't test count == 0 when
2196 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2198 * Makefile.in: add shrink-wrap.o.
2199 * config/i386/i386.c: include "shrink-wrap.h"
2200 * function.c: Likewise.
2201 (requires_stack_frame_p, next_block_for_reg,
2202 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2203 dup_block_and_redirect): Move to shrink-wrap.c
2204 (thread_prologue_and_epilogue_insns): Extract three code segments
2205 as functions in shrink-wrap.c
2206 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2208 * shrink-wrap.c: New file.
2209 * shrink-wrap.h: New file.
2211 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2213 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2214 reference to Solaris.
2216 2014-05-12 Mike Stump <mikestump@comcast.net>
2219 * genattrtab.c (filename): Add.
2220 (convert_set_attr_alternative): Improve error message.
2221 (check_defs): Restore read_md_filename for error messages.
2222 (gen_insn): Save filename.
2224 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2226 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2227 -fno-local-ivars and -fivar-visibility.
2228 * c-family/c.opt: Make -Wshadow also implicitly enable
2231 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2233 * doc/tm.texi: Remove reference to deleted macro.
2234 * doc/tm.texi.in: Likewise.
2236 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2239 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2242 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2244 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2245 * config/arm/aout.h (License): Same.
2246 * config/arm/bpabi.h (License): Same.
2247 * config/arm/elf.h (License): Same.
2248 * config/arm/linux-elf.h (License): Same.
2249 * config/arm/linux-gas.h (License): Same.
2250 * config/arm/netbsd-elf.h (License): Same.
2251 * config/arm/uclinux-eabi.h (License): Same.
2252 * config/arm/uclinux-elf.h (License): Same.
2253 * config/arm/vxworks.h (License): Same.
2255 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2257 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2258 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2259 number of operands to 3.
2260 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2261 * tree-nested.c (convert_nonlocal_omp_clauses,
2262 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2263 * gimplify.c (gimplify_scan_omp_clauses): Handle
2264 OMP_CLAUSE_LINEAR_STMT.
2265 * omp-low.c (lower_rec_input_clauses): Fix typo.
2266 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2267 cast between Fortran boolean_type_node and C _Bool if
2270 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2272 PR tree-optimization/61136
2273 * wide-int.h (multiple_of_p): Define a version that doesn't return
2275 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2276 integer_zerop/const_binop pair.
2277 (multiple_of_p): Likewise, converting both operands to widest_int
2280 2014-05-09 Teresa Johnson <tejohnson@google.com>
2282 * cgraphunit.c (analyze_functions): Use correct dump file.
2284 2014-05-09 Florian Weimer <fweimer@redhat.com>
2286 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2288 (stack_protect_return_slot_p): New function.
2289 (expand_used_vars): Call stack_protect_decl_p and
2290 stack_protect_return_slot_p for -fstack-protector-strong.
2292 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2295 * fold-const.c (fold_binary_loc): Changed width of mask.
2297 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2299 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2300 unsigned int initializers for regno_in, regno_out.
2302 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2305 * config/avr/avr.md (cc): Add new attribute set_vzn.
2306 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2307 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2308 with INC, DEC or NEG.
2309 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2310 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2311 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2313 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2316 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2318 * wide-int.cc (UTItype): Define.
2319 (UDWtype): Define for appropriate W_TYPE_SIZE.
2321 2014-05-09 Richard Biener <rguenther@suse.de>
2323 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2324 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2325 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2326 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2327 ssa_propagate): Adjust.
2329 2014-05-08 Jeff Law <law@redhat.com>
2331 PR tree-optimization/61009
2332 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2333 tri-state rather than a boolean. When a block is too big to
2334 thread through, inform caller via negative return value.
2335 (thread_across_edge): If a block was too big for normal threading,
2336 then it's too big for a joiner too, so remove temporary equivalences
2337 and return immediately.
2339 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2340 Matthias Klose <doko@ubuntu.com>
2343 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2345 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2348 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2350 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2353 * config/alpha/alpha.c: Include gimple-iterator.h.
2354 (alpha_gimple_fold_builtin): New function. Move
2355 ALPHA_BUILTIN_UMULH folding from ...
2356 (alpha_fold_builtin): ... here.
2357 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2359 2014-05-08 Wei Mi <wmi@google.com>
2362 * config/i386/i386.c (ix86_compute_frame_layout): Update
2363 preferred_stack_boundary for call, expanded from tls descriptor.
2364 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2365 to depend on SP register.
2366 (*tls_local_dynamic_base_32_gnu): Ditto.
2367 (*tls_local_dynamic_32_once): Ditto.
2368 (tls_global_dynamic_64_<mode>): Set
2369 ix86_tls_descriptor_calls_expanded_in_cfun.
2370 (tls_local_dynamic_base_64_<mode>): Ditto.
2371 (tls_global_dynamic_32): Set
2372 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2373 to depend on SP register.
2374 (tls_local_dynamic_base_32): Ditto.
2376 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2378 * config/arm/arm_neon.h: Update comment.
2379 * config/arm/neon-docgen.ml: Delete.
2380 * config/arm/neon-gen.ml: Delete.
2381 * doc/arm-neon-intrinsics.texi: Update comment.
2383 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2385 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2387 (vand, vorr, veor, vorn, vbic): Remove.
2388 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2390 (neon_vsub_unspec): Likewise.
2391 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2393 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2395 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2396 (vadd_s16): Likewise.
2397 (vadd_s32): Likewise.
2398 (vadd_f32): Likewise.
2399 (vadd_u8): Likewise.
2400 (vadd_u16): Likewise.
2401 (vadd_u32): Likewise.
2402 (vadd_s64): Likewise.
2403 (vadd_u64): Likewise.
2404 (vaddq_s8): Likewise.
2405 (vaddq_s16): Likewise.
2406 (vaddq_s32): Likewise.
2407 (vaddq_s64): Likewise.
2408 (vaddq_f32): Likewise.
2409 (vaddq_u8): Likewise.
2410 (vaddq_u16): Likewise.
2411 (vaddq_u32): Likewise.
2412 (vaddq_u64): Likewise.
2413 (vmul_s8): Likewise.
2414 (vmul_s16): Likewise.
2415 (vmul_s32): Likewise.
2416 (vmul_f32): Likewise.
2417 (vmul_u8): Likewise.
2418 (vmul_u16): Likewise.
2419 (vmul_u32): Likewise.
2420 (vmul_p8): Likewise.
2421 (vmulq_s8): Likewise.
2422 (vmulq_s16): Likewise.
2423 (vmulq_s32): Likewise.
2424 (vmulq_f32): Likewise.
2425 (vmulq_u8): Likewise.
2426 (vmulq_u16): Likewise.
2427 (vmulq_u32): Likewise.
2428 (vsub_s8): Likewise.
2429 (vsub_s16): Likewise.
2430 (vsub_s32): Likewise.
2431 (vsub_f32): Likewise.
2432 (vsub_u8): Likewise.
2433 (vsub_u16): Likewise.
2434 (vsub_u32): Likewise.
2435 (vsub_s64): Likewise.
2436 (vsub_u64): Likewise.
2437 (vsubq_s8): Likewise.
2438 (vsubq_s16): Likewise.
2439 (vsubq_s32): Likewise.
2440 (vsubq_s64): Likewise.
2441 (vsubq_f32): Likewise.
2442 (vsubq_u8): Likewise.
2443 (vsubq_u16): Likewise.
2444 (vsubq_u32): Likewise.
2445 (vsubq_u64): Likewise.
2446 (vand_s8): Likewise.
2447 (vand_s16): Likewise.
2448 (vand_s32): Likewise.
2449 (vand_u8): Likewise.
2450 (vand_u16): Likewise.
2451 (vand_u32): Likewise.
2452 (vand_s64): Likewise.
2453 (vand_u64): Likewise.
2454 (vandq_s8): Likewise.
2455 (vandq_s16): Likewise.
2456 (vandq_s32): Likewise.
2457 (vandq_s64): Likewise.
2458 (vandq_u8): Likewise.
2459 (vandq_u16): Likewise.
2460 (vandq_u32): Likewise.
2461 (vandq_u64): Likewise.
2462 (vorr_s8): Likewise.
2463 (vorr_s16): Likewise.
2464 (vorr_s32): Likewise.
2465 (vorr_u8): Likewise.
2466 (vorr_u16): Likewise.
2467 (vorr_u32): Likewise.
2468 (vorr_s64): Likewise.
2469 (vorr_u64): Likewise.
2470 (vorrq_s8): Likewise.
2471 (vorrq_s16): Likewise.
2472 (vorrq_s32): Likewise.
2473 (vorrq_s64): Likewise.
2474 (vorrq_u8): Likewise.
2475 (vorrq_u16): Likewise.
2476 (vorrq_u32): Likewise.
2477 (vorrq_u64): Likewise.
2478 (veor_s8): Likewise.
2479 (veor_s16): Likewise.
2480 (veor_s32): Likewise.
2481 (veor_u8): Likewise.
2482 (veor_u16): Likewise.
2483 (veor_u32): Likewise.
2484 (veor_s64): Likewise.
2485 (veor_u64): Likewise.
2486 (veorq_s8): Likewise.
2487 (veorq_s16): Likewise.
2488 (veorq_s32): Likewise.
2489 (veorq_s64): Likewise.
2490 (veorq_u8): Likewise.
2491 (veorq_u16): Likewise.
2492 (veorq_u32): Likewise.
2493 (veorq_u64): Likewise.
2494 (vbic_s8): Likewise.
2495 (vbic_s16): Likewise.
2496 (vbic_s32): Likewise.
2497 (vbic_u8): Likewise.
2498 (vbic_u16): Likewise.
2499 (vbic_u32): Likewise.
2500 (vbic_s64): Likewise.
2501 (vbic_u64): Likewise.
2502 (vbicq_s8): Likewise.
2503 (vbicq_s16): Likewise.
2504 (vbicq_s32): Likewise.
2505 (vbicq_s64): Likewise.
2506 (vbicq_u8): Likewise.
2507 (vbicq_u16): Likewise.
2508 (vbicq_u32): Likewise.
2509 (vbicq_u64): Likewise.
2510 (vorn_s8): Likewise.
2511 (vorn_s16): Likewise.
2512 (vorn_s32): Likewise.
2513 (vorn_u8): Likewise.
2514 (vorn_u16): Likewise.
2515 (vorn_u32): Likewise.
2516 (vorn_s64): Likewise.
2517 (vorn_u64): Likewise.
2518 (vornq_s8): Likewise.
2519 (vornq_s16): Likewise.
2520 (vornq_s32): Likewise.
2521 (vornq_s64): Likewise.
2522 (vornq_u8): Likewise.
2523 (vornq_u16): Likewise.
2524 (vornq_u32): Likewise.
2525 (vornq_u64): Likewise.
2527 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2529 * wide-int.cc (UTItype): Define.
2530 (UDWtype): Define for appropriate W_TYPE_SIZE.
2532 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2534 PR tree-optimization/59100
2535 * tree-ssa-phiopt.c: Include tree-inline.h.
2536 (neutral_element_p, absorbing_element_p): New functions.
2537 (value_replacement): Handle conditional binary operations with a
2538 neutral or absorbing element.
2540 2014-05-08 Richard Biener <rguenther@suse.de>
2542 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2543 folding the expression.
2544 (valueize_expr): Remove.
2545 (visit_reference_op_load): Do not valueize the result of
2547 (simplify_binary_expression): Likewise.
2548 (simplify_unary_expression): Likewise.
2550 2014-05-08 Richard Biener <rguenther@suse.de>
2552 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2553 looking at TYPE_ARG_TYPES.
2555 2014-05-08 Richard Biener <rguenther@suse.de>
2557 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2558 pointer propagation special-case.
2560 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2562 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2563 core part of address expressions.
2565 2014-05-08 Alan Modra <amodra@gmail.com>
2568 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2569 loads and stores when -mno-strict-align at any alignment.
2570 (expand_block_clear): Similarly. Also correct calculation of
2573 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2576 * tree-complex.c (expand_complex_move): Keep line info when expanding
2578 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2579 of complex expression. Use new argument to display correct location
2580 for values coming from phi statement.
2581 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2582 (warn_uninitialized_phi): Pass location of phi argument to
2584 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2585 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2587 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2589 * config/rs6000/predicates.md (indexed_address_mem): New.
2590 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2591 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2592 fpstore_ux, fpstore_u.
2593 (sign_extend, indexed, update): New.
2594 (cell_micro): Adjust.
2595 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2596 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2597 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2598 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2599 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2600 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2601 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2602 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2603 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2604 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2605 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2606 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2607 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2608 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2609 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2611 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2612 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2613 *vsx_extract_<mode>_store): Adjust.
2614 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2615 is_cracked_insn, insn_must_be_first_in_group,
2616 insn_must_be_last_in_group): Adjust.
2618 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2620 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2621 ppc440-fpstore): Adjust.
2622 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2623 ppc476-fpstore): Adjust.
2624 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2625 ppc601-fpstore): Adjust.
2626 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2628 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2630 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2631 ppc7450-fpstore): Adjust.
2632 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2633 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2634 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2636 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2637 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2638 cell-fpstore, cell-fpstore-update): Adjust.
2639 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2640 ppce300c3_store, ppce300c3_fpstore): Adjust.
2641 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2642 e500mc_fpstore): Adjust.
2643 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2644 e500mc64_store, e500mc64_fpstore): Adjust.
2645 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2646 e5500_fpstore): Adjust.
2647 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2648 e6500_fpstore): Adjust.
2649 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2651 * config/rs6000/power4.md (power4-load, power4-load-ext,
2652 power4-load-ext-update, power4-load-ext-update-indexed,
2653 power4-load-update-indexed, power4-load-update, power4-fpload,
2654 power4-fpload-update, power4-store, power4-store-update,
2655 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2657 * config/rs6000/power5.md (power5-load, power5-load-ext,
2658 power5-load-ext-update, power5-load-ext-update-indexed,
2659 power5-load-update-indexed, power5-load-update, power5-fpload,
2660 power5-fpload-update, power5-store, power5-store-update,
2661 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2663 * config/rs6000/power6.md (power6-load, power6-load-ext,
2664 power6-load-update, power6-load-update-indexed,
2665 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2666 power6-fpload-update, power6-store, power6-store-update,
2667 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2669 * config/rs6000/power7.md (power7-load, power7-load-ext,
2670 power7-load-update, power7-load-update-indexed,
2671 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2672 power7-fpload-update, power7-store, power7-store-update,
2673 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2675 * config/rs6000/power8.md (power8-load, power8-load-update,
2676 power8-load-ext, power8-load-ext-update, power8-fpload,
2677 power8-fpload-update, power8-store, power8-store-update-indexed,
2678 power8-fpstore, power8-fpstore-update): Adjust.
2679 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2681 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2682 titan_lsu_store, titan_lsu_fpstore): Adjust.
2683 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2685 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2688 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2689 unrolled byte insns. Emit address increments after move insns.
2691 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2693 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2694 const_gimple, rather than a gimple.
2695 (gimple_call_builtin_p): Likewise, for the three variants.
2697 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2698 (gimple_call_builtin_p): Likewise, for the three variants.
2700 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2702 PR tree-optimization/61095
2703 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2705 2014-05-07 Richard Biener <rguenther@suse.de>
2707 PR tree-optimization/61034
2708 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2709 (maybe_skip_until): Use translate to take into account
2710 lattices when trying to do disambiguations.
2711 (get_continuation_for_phi_1): Likewise.
2712 (get_continuation_for_phi): Adjust for added translate arguments.
2713 (walk_non_aliased_vuses): Likewise.
2714 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2715 (walk_non_aliased_vuses): Likewise.
2716 (call_may_clobber_ref_p_1): Declare.
2717 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2718 calls. Stop early if we are only supposed to disambiguate.
2719 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2721 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2723 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2724 Emit an error when the function has arguments.
2726 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2728 * cfgloop.h (unswitch_loops): Remove.
2729 * doc/passes.texi: Remove references to loop-unswitch.c
2730 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2732 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2734 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2735 check for loads group of length 3.
2736 (vect_permute_load_chain): New permutations for loads group of
2738 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2739 of vec_perm_shuffle for the new permutations.
2741 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2743 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2744 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2745 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2746 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2747 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2748 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2749 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2750 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2752 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2754 * loop-unswitch.c: Delete.
2756 2014-05-07 Richard Biener <rguenther@suse.de>
2758 * config.gcc: Always set need_64bit_hwint to yes.
2760 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2762 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2763 of using optimize_size.
2765 2014-05-06 Mike Stump <mikestump@comcast.net>
2767 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2769 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2771 * config/i386/sse.md (*mov<mode>_internal)
2772 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2773 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2774 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2775 (*<code><mode>3, *andnot<mode>3<mask_name>)
2776 (<mask_codefor><code><mode>3<mask_name>): Only consider
2777 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2779 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2782 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2784 * lra-constraints.c (valid_address_p): Move earlier in file.
2785 Add a constraint argument to the address_info version.
2786 (satisfies_memory_constraint_p): New function.
2787 (satisfies_address_constraint_p): Likewise.
2788 (process_alt_operands, curr_insn_transform): Use them.
2789 (process_address): Pass the constraint to valid_address_p when
2790 checking address operands.
2792 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2794 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2795 to their respective blocks. Fix inadvertent use of "node".
2797 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2799 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2801 (init_emit_once): ...here.
2802 * rtl.h (init_derived_machine_modes): Declare.
2803 * toplev.c (do_compile): Call it even if no_backend.
2805 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2806 Mike Stump <mikestump@comcast.net>
2807 Richard Sandiford <rdsandiford@googlemail.com>
2808 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2810 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2811 (rtx_equal_for_memref_p): Update comment.
2812 (adjust_offset_for_component_ref): Use wide-int interfaces.
2813 * builtins.c (get_object_alignment_2): Likewise.
2814 (c_readstr): Likewise.
2815 (target_char_cast): Add comment.
2816 (determine_block_size): Use wide-int interfaces.
2817 (expand_builtin_signbit): Likewise.
2818 (fold_builtin_int_roundingfn): Likewise.
2819 (fold_builtin_bitop): Likewise.
2820 (fold_builtin_bswap): Likewise.
2821 (fold_builtin_logarithm): Use signop.
2822 (fold_builtin_pow): Likewise.
2823 (fold_builtin_memory_op): Use wide-int interfaces.
2824 (fold_builtin_object_size): Likewise.
2825 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2826 nb_iterations_estimate.
2827 (record_niter_bound): Use wide-int interfaces.
2828 (get_estimated_loop_iterations_int): Likewise.
2829 (get_estimated_loop_iterations): Likewise.
2830 (get_max_loop_iterations): Likewise.
2831 * cfgloop.h: Include wide-int.h.
2832 (struct nb_iter_bound): Change bound to widest_int.
2833 (struct loop): Change nb_iterations_upper_bound and
2834 nb_iterations_estimate to widest_int.
2835 (record_niter_bound): Switch to use widest_int.
2836 (get_estimated_loop_iterations): Likewise.
2837 (get_max_loop_iterations): Likewise.
2838 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2839 update for wide-int.
2840 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2841 * combine.c (try_combine): Likewise.
2842 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2843 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2845 (aarch64_float_const_representable_p): Likewise.
2846 * config/arc/arc.c: Include wide-int.h.
2847 (arc_can_use_doloop_p): Use wide-int interfaces.
2848 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2849 (vfp3_const_double_index): Likewise.
2850 * config/avr/avr.c (avr_out_round): Likewise.
2851 (avr_fold_builtin): Likewise.
2852 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2853 (bfin_can_use_doloop_p): Likewise.
2854 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2855 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2856 * config/i386/i386.c: Include wide-int.h.
2857 (ix86_data_alignment): Use wide-int interfaces.
2858 (ix86_local_alignment): Likewise.
2859 (ix86_emit_swsqrtsf): Update real_from_integer.
2860 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2861 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2862 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2863 (zero_constant): Likewise.
2864 (input_operand): Likewise.
2865 (splat_input_operand): Likewise.
2866 (non_logical_cint_operand): Change const_double to const_wide_int.
2867 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2868 (easy_altivec_constant): Remove comment.
2869 (paired_expand_vector_init): Use CONSTANT_P.
2870 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2871 (rs6000_emit_move): Update checks.
2872 (rs6000_aggregate_candidate): Use wide-int interfaces.
2873 (rs6000_expand_ternop_builtin): Likewise.
2874 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
2875 (rs6000_assemble_integer): Likewise.
2876 (rs6000_hash_constant): Likewise.
2877 (output_toc): Likewise.
2878 (rs6000_rtx_costs): Likewise.
2879 (rs6000_emit_swrsqrt); Update call to real_from_integer.
2880 * config/rs6000/rs6000-c.c: Include wide-int.h.
2881 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
2882 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
2883 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
2884 Handle CONST_WIDE_INT.
2885 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
2886 Use tree_fits_uhwi_p.
2887 * config/sparc/sparc.c: Include wide-int.h.
2888 (sparc_fold_builtin): Use wide-int interfaces.
2889 * config/vax/vax.c: Include wide-int.h.
2890 (vax_float_literal): Use real_from_integer.
2891 * coretypes.h (struct hwivec_def): New.
2893 (const_hwivec): New.
2894 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
2895 (equiv_constant): Handle CONST_WIDE_INT.
2896 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
2897 (cselib_hash_rtx): Handle CONST_WIDE_INT.
2898 * dbxout.c (stabstr_U): Use wide-int interfaces.
2899 (dbxout_type): Update to use cst_fits_shwi_p.
2900 * defaults.h (LOG2_BITS_PER_UNIT): Define.
2901 (TARGET_SUPPORTS_WIDE_INT): Add default.
2902 * dfp.c: Include wide-int.h.
2903 (decimal_real_to_integer2): Use wide-int interfaces and rename to
2904 decimal_real_to_integer.
2905 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
2906 decimal_real_to_integer.
2907 * doc/generic.texi (Constant expressions): Update for wide_int.
2908 * doc/rtl.texi (const_double): Likewise.
2909 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
2910 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
2911 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
2912 (REAL_VALUE_FROM_INT): Remove.
2913 (TARGET_SUPPORTS_WIDE_INT): New.
2914 * doc/tm.texi: Regenerate.
2915 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
2916 * double-int.h: Include wide-int.h.
2917 (struct wi::int_traits): New.
2918 * dwarf2out.c (get_full_len): New.
2919 (dw_val_equal_p): Add case dw_val_class_wide_int.
2920 (size_of_loc_descr): Likewise.
2921 (output_loc_operands): Likewise.
2922 (insert_double): Remove.
2923 (insert_wide_int): New.
2925 (print_die): Add case dw_val_class_wide_int.
2926 (attr_checksum): Likewise.
2927 (attr_checksum_ordered): Likewise.
2928 (same_dw_val_p): Likewise.
2929 (size_of_die): Likewise.
2930 (value_format): Likewise.
2931 (output_die): Likewise.
2932 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
2934 (clz_loc_descriptor): Use wide-int interfaces.
2935 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
2936 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
2937 (round_up_to_align): Use wide-int interfaces.
2938 (field_byte_offset): Likewise.
2939 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
2940 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
2941 CONST_DOUBLE handling. Use wide-int interfaces.
2942 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
2943 (gen_enumeration_type_die): Use add_AT_wide.
2944 (hash_loc_operands): Add case dw_val_class_wide_int.
2945 (compare_loc_operands): Likewise.
2946 * dwarf2out.h: Include wide-int.h.
2947 (wide_int_ptr): New.
2948 (enum dw_val_class): Add dw_val_class_wide_int.
2949 (struct dw_val_struct): Add val_wide.
2950 * emit-rtl.c (const_wide_int_htab): New.
2951 (const_wide_int_htab_hash): New.
2952 (const_wide_int_htab_eq): New.
2953 (lookup_const_wide_int): New.
2954 (const_double_htab_hash): Use wide-int interfaces.
2955 (const_double_htab_eq): Likewise.
2956 (rtx_to_double_int): Conditionally compile for wide-int.
2957 (immed_double_int_const): Rename to immed_wide_int_const and
2958 update for wide-int.
2959 (immed_double_const): Conditionally compile for wide-int.
2960 (init_emit_once): Use wide-int interfaces.
2961 * explow.c (plus_constant): Likewise.
2962 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
2963 (lshift_value): Use wide-int interfaces.
2964 (expand_mult): Likewise.
2965 (choose_multiplier): Likewise.
2966 (expand_smod_pow2): Likewise.
2967 (make_tree): Likewise.
2968 * expr.c (convert_modes): Consolidate handling of constants.
2969 Use wide-int interfaces.
2970 (emit_group_load_1): Add note.
2971 (store_expr): Update comment.
2972 (get_inner_reference): Use wide-int interfaces.
2973 (expand_constructor): Update comment.
2974 (expand_expr_real_2): Use wide-int interfaces.
2975 (expand_expr_real_1): Likewise.
2976 (reduce_to_bit_field_precision): Likewise.
2977 (const_vector_from_tree): Likewise.
2978 * final.c: Include wide-int-print.h.
2979 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
2980 * fixed-value.c: Include wide-int.h.
2981 (fixed_from_string): Use wide-int interfaces.
2982 (fixed_to_decimal): Likewise.
2983 (fixed_convert_from_real): Likewise.
2984 (real_convert_from_fixed): Likewise.
2985 * fold-const.h (mem_ref_offset): Return an offset_int.
2986 (div_if_zero_remainder): Remove code parameter.
2987 * fold-const.c (div_if_zero_remainder): Remove code parameter.
2988 Use wide-int interfaces.
2989 (may_negate_without_overflow_p): Use wide-int interfaces.
2990 (negate_expr_p): Likewise.
2991 (fold_negate_expr): Likewise.
2992 (int_const_binop_1): Likewise.
2993 (const_binop): Likewise.
2994 (fold_convert_const_int_from_int): Likewise.
2995 (fold_convert_const_int_from_real): Likewise.
2996 (fold_convert_const_int_from_fixed): Likewise.
2997 (fold_convert_const_fixed_from_int): Likewise.
2998 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
2999 (sign_bit_p): Use wide-int interfaces.
3000 (make_range_step): Likewise.
3001 (build_range_check): Likewise. Pass an integer of the correct type
3002 instead of using integer_one_node.
3003 (range_predecessor): Pass an integer of the correct type instead
3004 of using integer_one_node.
3005 (range_successor): Likewise.
3006 (merge_ranges): Likewise.
3007 (unextend): Use wide-int interfaces.
3008 (extract_muldiv_1): Likewise.
3009 (fold_div_compare): Likewise.
3010 (fold_single_bit_test): Likewise.
3011 (fold_sign_changed_comparison): Likewise.
3012 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3013 (fold_plusminus_mult_expr): Use wide-int interfaces.
3014 (native_encode_int): Likewise.
3015 (native_interpret_int): Likewise.
3016 (fold_unary_loc): Likewise.
3017 (pointer_may_wrap_p): Likewise.
3018 (size_low_cst): Likewise.
3019 (mask_with_tz): Likewise.
3020 (fold_binary_loc): Likewise.
3021 (fold_ternary_loc): Likewise.
3022 (multiple_of_p): Likewise.
3023 (tree_call_nonnegative_warnv_p): Update calls to
3024 tree_int_cst_min_precision and real_from_integer.
3025 (fold_negate_const): Use wide-int interfaces.
3026 (fold_abs_const): Likewise.
3027 (fold_relational_const): Use tree_int_cst_lt.
3028 (round_up_loc): Use wide-int interfaces.
3029 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3030 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3031 * gengtype.c: Remove include of double-int.h.
3032 (do_typedef): Use wide-int interfaces.
3033 (open_base_files): Add wide-int.h.
3034 (main): Add offset_int and widest_int typedefs.
3035 * gengtype-lex.l: Handle "^".
3036 (CXX_KEYWORD): Add "static".
3037 * gengtype-parse.c (require3): New.
3038 (require_template_declaration): Handle constant template arguments
3039 and nested templates.
3040 * gengtype-state.c: Don't include "double-int.h".
3041 * genpreds.c (write_one_predicate_function): Update comment.
3042 (write_tm_constrs_h): Add check for hval and lval use in
3044 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3045 (add_to_sequence): Likewise.
3046 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3047 and const_double_operand.
3048 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3050 * gimple-fold.c (get_base_constructor): Likewise.
3051 (fold_array_ctor_reference): Likewise.
3052 (fold_nonarray_ctor_reference): Likewise.
3053 (fold_const_aggregate_ref_1): Likewise.
3054 (gimple_val_nonnegative_real_p): Likewise.
3055 (gimple_fold_indirect_ref): Likewise.
3056 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3057 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3058 (struct slsr_cand_d): Change index to be widest_int.
3059 (struct incr_info_d): Change incr to be widest_int.
3060 (alloc_cand_and_find_basis): Use wide-int interfaces.
3061 (slsr_process_phi): Likewise.
3062 (backtrace_base_for_ref): Likewise. Return a widest_int.
3063 (restructure_reference): Take a widest_int instead of a double_int.
3064 (slsr_process_ref): Use wide-int interfaces.
3065 (create_mul_ssa_cand): Likewise.
3066 (create_mul_imm_cand): Likewise.
3067 (create_add_ssa_cand): Likewise.
3068 (create_add_imm_cand): Take a widest_int instead of a double_int.
3069 (slsr_process_add): Use wide-int interfaces.
3070 (slsr_process_cast): Likewise.
3071 (slsr_process_copy): Likewise.
3072 (dump_candidate): Likewise.
3073 (dump_incr_vec): Likewise.
3074 (replace_ref): Likewise.
3075 (cand_increment): Likewise. Return a widest_int.
3076 (cand_abs_increment): Likewise.
3077 (replace_mult_candidate): Take a widest_int instead of a double_int.
3078 (replace_unconditional_candidate): Use wide-int interfaces.
3079 (incr_vec_index): Take a widest_int instead of a double_int.
3080 (create_add_on_incoming_edge): Likewise.
3081 (create_phi_basis): Use wide-int interfaces.
3082 (replace_conditional_candidate): Likewise.
3083 (record_increment): Take a widest_int instead of a double_int.
3084 (record_phi_increments): Use wide-int interfaces.
3085 (phi_incr_cost): Take a widest_int instead of a double_int.
3086 (lowest_cost_path): Likewise.
3087 (total_savings): Likewise.
3088 (analyze_increments): Use wide-int interfaces.
3089 (ncd_with_phi): Take a widest_int instead of a double_int.
3090 (ncd_of_cand_and_phis): Likewise.
3091 (nearest_common_dominator_for_cands): Likewise.
3092 (insert_initializers): Use wide-int interfaces.
3093 (all_phi_incrs_profitable): Likewise.
3094 (replace_one_candidate): Likewise.
3095 (replace_profitable_candidates): Likewise.
3096 * godump.c: Include wide-int-print.h.
3097 (go_output_typedef): Use wide-int interfaces.
3098 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3099 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3100 (build_loop_iteration_domains): Likewise.
3101 * hooks.h: Include wide-int.h rather than double-int.h.
3102 (hook_bool_dint_dint_uint_bool_true): Delete.
3103 (hook_bool_wint_wint_uint_bool_true): Declare.
3104 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3105 (hook_bool_wint_wint_uint_bool_true): New.
3106 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3108 (ubsan_expand_si_overflow_mul_check): Likewise.
3109 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3110 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3111 (get_ancestor_addr_info): Likewise.
3112 (ipa_modify_call_arguments): Likewise.
3113 * loop-doloop.c (doloop_modify): Likewise.
3114 (doloop_optimize): Likewise.
3115 * loop-iv.c (iv_number_of_iterations): Likewise.
3116 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3117 (unroll_loop_constant_iterations): Likewise.
3118 (decide_unroll_runtime_iterations): Likewise.
3119 (unroll_loop_runtime_iterations): Likewise.
3120 (decide_peel_simple): Likewise.
3121 (decide_unroll_stupid): Likewise.
3122 * lto-streamer-in.c (streamer_read_wi): Add.
3123 (input_cfg): Use wide-int interfaces.
3124 (lto_input_tree_1): Likewise.
3125 * lto-streamer-out.c (streamer_write_wi): Add.
3126 (hash_tree): Use wide-int interfaces.
3127 (output_cfg): Likewise.
3128 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3129 (GTFILES): Add wide-int.h and signop.h.
3130 (TAGS): Look for .cc files too.
3131 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3132 * optabs.c (expand_subword_shift): Likewise.
3133 (expand_doubleword_shift): Likewise.
3134 (expand_absneg_bit): Likewise.
3135 (expand_copysign_absneg): Likewise.
3136 (expand_copysign_bit): Likewise.
3137 * postreload.c (reload_cse_simplify_set): Likewise.
3138 * predict.c (predict_iv_comparison): Likewise.
3139 * pretty-print.h: Include wide-int-print.h.
3141 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3142 * print-tree.c: Include wide-int-print.h.
3143 (print_node_brief): Use wide-int interfaces.
3144 (print_node): Likewise.
3145 * read-rtl.c (validate_const_wide_int): New.
3146 (read_rtx_code): Add CONST_WIDE_INT case.
3147 * real.c: Include wide-int.h.
3148 (real_to_integer2): Delete.
3149 (real_to_integer): New function, returning a wide_int.
3150 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3151 (ten_to_ptwo): Update call to real_from_integer.
3152 (real_digit): Likewise.
3153 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3154 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3155 (REAL_VALUE_TO_INT): Delete.
3156 (real_to_integer): Declare a wide-int form.
3157 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3158 * recog.c (const_int_operand): Improve comment.
3159 (const_scalar_int_operand): New.
3160 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3161 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3162 (split_double): Likewise.
3163 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3164 (rtx_size): Likewise.
3165 (rtx_alloc_stat_v): New.
3166 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3167 (cwi_output_hex): New.
3168 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3169 (cwi_check_failed_bounds): New.
3170 * rtl.def (CONST_WIDE_INT): New.
3171 * rtl.h: Include <utility> and wide-int.h.
3172 (struct hwivec_def): New.
3173 (CWI_GET_NUM_ELEM): New.
3174 (CWI_PUT_NUM_ELEM): New.
3175 (struct rtx_def): Add num_elem and hwiv.
3176 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3177 (CASE_CONST_UNIQUE): Likewise.
3178 (CASE_CONST_ANY): Likewise.
3179 (CONST_SCALAR_INT_P): Likewise.
3180 (CONST_WIDE_INT_P): New.
3182 (HWIVEC_CHECK): New.
3183 (cwi_check_failed_bounds): New.
3185 (HWIVEC_CHECK): New.
3186 (CONST_WIDE_INT_VEC) New.
3187 (CONST_WIDE_INT_NUNITS) New.
3188 (CONST_WIDE_INT_ELT) New.
3189 (rtx_mode_t): New type.
3190 (wi::int_traits <rtx_mode_t>): New.
3192 (wi::min_value): New.
3193 (wi::max_value): New.
3195 (const_wide_int_alloc): New.
3196 (immed_wide_int_const): New.
3197 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3198 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3199 * signop.h: New file.
3200 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3201 (simplify_const_unary_operation): Use wide-int interfaces.
3202 (simplify_binary_operation_1): Likewise.
3203 (simplify_const_binary_operation): Likewise.
3204 (simplify_const_relational_operation): Likewise.
3205 (simplify_immed_subreg): Likewise.
3206 * stmt.c (expand_case): Likewise.
3207 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3208 signop rather than a bool.
3209 * stor-layout.c (layout_type): Use wide-int interfaces.
3210 (initialize_sizetypes): Update calls to
3211 set_min_and_max_values_for_integral_type.
3212 (set_min_and_max_values_for_integral_type): Take a signop rather
3213 than a bool. Use wide-int interfaces.
3214 (fixup_signed_type): Update accordingly. Remove
3215 HOST_BITS_PER_DOUBLE_INT limit.
3216 (fixup_unsigned_type): Likewise.
3217 * system.h (STATIC_CONSTANT_P): New.
3218 (STATIC_ASSERT): New.
3219 * target.def (can_use_doloop_p): Take widest_ints rather than
3221 * target.h: Include wide-int.h rather than double-int.h.
3222 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3224 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3225 rather than INT_CST_LT_UNSIGNED.
3226 (can_use_doloop_if_innermost): Take widest_ints rather than
3228 * tree-affine.c: Include wide-int-print.h.
3229 (double_int_ext_for_comb): Delete.
3230 (wide_int_ext_for_comb): New.
3231 (aff_combination_zero): Use wide-int interfaces.
3232 (aff_combination_const): Take a widest_int instead of a double_int.
3233 (aff_combination_elt): Use wide-int interfaces.
3234 (aff_combination_scale): Take a widest_int instead of a double_int.
3235 (aff_combination_add_elt): Likewise.
3236 (aff_combination_add_cst): Likewise.
3237 (aff_combination_add): Use wide-int interfaces.
3238 (aff_combination_convert): Likewise.
3239 (tree_to_aff_combination): Likewise.
3240 (add_elt_to_tree): Take a widest_int instead of a double_int.
3241 (aff_combination_to_tree): Use wide-int interfaces.
3242 (aff_combination_remove_elt): Likewise.
3243 (aff_combination_add_product): Take a widest_int instead of
3245 (aff_combination_mult): Use wide-int interfaces.
3246 (aff_combination_expand): Likewise.
3247 (double_int_constant_multiple_p): Delete.
3248 (wide_int_constant_multiple_p): New.
3249 (aff_combination_constant_multiple_p): Take a widest_int pointer
3250 instead of a double_int pointer.
3251 (print_aff): Use wide-int interfaces.
3252 (get_inner_reference_aff): Take a widest_int pointer
3253 instead of a double_int pointer.
3254 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3255 * tree-affine.h: Include wide-int.h.
3256 (struct aff_comb_elt): Change type of coef to widest_int.
3257 (struct affine_tree_combination): Change type of offset to widest_int.
3258 (double_int_ext_for_comb): Delete.
3259 (wide_int_ext_for_comb): New.
3260 (aff_combination_const): Use widest_int instead of double_int.
3261 (aff_combination_scale): Likewise.
3262 (aff_combination_add_elt): Likewise.
3263 (aff_combination_constant_multiple_p): Likewise.
3264 (get_inner_reference_aff): Likewise.
3265 (aff_comb_cannot_overlap_p): Likewise.
3266 (aff_combination_zero_p): Use wide-int interfaces.
3267 * tree.c: Include tree.h.
3268 (init_ttree): Use make_int_cst.
3269 (tree_code_size): Removed code for INTEGER_CST case.
3270 (tree_size): Add INTEGER_CST case.
3271 (make_node_stat): Update comment.
3272 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3273 (build_int_cst_type): Use wide-int interfaces.
3274 (double_int_to_tree): Likewise.
3275 (double_int_fits_to_tree_p): Delete.
3276 (force_fit_type_double): Delete.
3277 (force_fit_type): New.
3278 (int_cst_hash_hash): Use wide-int interfaces.
3279 (int_cst_hash_eq): Likewise.
3280 (build_int_cst_wide): Delete.
3281 (wide_int_to_tree): New.
3282 (cache_integer_cst): Use wide-int interfaces.
3283 (build_low_bits_mask): Likewise.
3284 (cst_and_fits_in_hwi): Likewise.
3285 (real_value_from_int_cst): Likewise.
3286 (make_int_cst_stat): New.
3287 (integer_zerop): Use wide_int interfaces.
3288 (integer_onep): Likewise.
3289 (integer_all_onesp): Likewise.
3290 (integer_pow2p): Likewise.
3291 (integer_nonzerop): Likewise.
3292 (tree_log2): Likewise.
3293 (tree_floor_log2): Likewise.
3294 (tree_ctz): Likewise.
3295 (int_size_in_bytes): Likewise.
3296 (mem_ref_offset): Return an offset_int rather than a double_int.
3297 (build_type_attribute_qual_variant): Use wide_int interfaces.
3298 (type_hash_eq): Likewise
3299 (tree_int_cst_equal): Likewise.
3300 (tree_int_cst_lt): Delete.
3301 (tree_int_cst_compare): Likewise.
3302 (tree_fits_shwi_p): Use wide_int interfaces.
3303 (tree_fits_uhwi_p): Likewise.
3304 (tree_int_cst_sign_bit): Likewise.
3305 (tree_int_cst_sgn): Likewise.
3306 (tree_int_cst_min_precision): Take a signop rather than a bool.
3307 (simple_cst_equal): Use wide_int interfaces.
3308 (compare_tree_int): Likewise.
3309 (iterative_hash_expr): Likewise.
3310 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3312 (get_type_static_bounds): Use wide_int interfaces.
3313 (tree_int_cst_elt_check_failed): New.
3314 (build_common_tree_nodes): Reordered to set prec before filling in
3316 (int_cst_value): Check cst_and_fits_in_hwi.
3317 (widest_int_cst_value): Use wide_int interfaces.
3318 (upper_bound_in_type): Likewise.
3319 (lower_bound_in_type): Likewise.
3320 (num_ending_zeros): Likewise.
3321 (drop_tree_overflow): Likewise.
3322 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3323 (gen_conditions_for_pow_cst_base): Likewise.
3324 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3325 (group_case_labels_stmt): Use wide-int interfaces.
3326 (verify_gimple_assign_binary): Likewise.
3327 (print_loop): Likewise.
3328 * tree-chrec.c (tree_fold_binomial): Likewise.
3329 * tree-core.h (struct tree_base): Add int_length.
3330 (struct tree_int_cst): Change rep of value.
3331 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3332 (dr_may_alias_p): Likewise.
3333 (max_stmt_executions_tree): Likewise.
3334 * tree.def (INTEGER_CST): Update comment.
3335 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3336 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3337 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3338 (dequeue_and_dump): Use wide-int interfaces.
3339 * tree.h: Include wide-int.h.
3340 (NULL_TREE): Moved to earlier loc in file.
3341 (TREE_INT_CST_ELT_CHECK): New.
3342 (tree_int_cst_elt_check_failed): New.
3344 (TREE_INT_CST): Delete.
3345 (TREE_INT_CST_LOW): Use wide-int interfaces.
3346 (TREE_INT_CST_HIGH): Delete.
3347 (TREE_INT_CST_NUNITS): New.
3348 (TREE_INT_CST_EXT_NUNITS): Likewise.
3349 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3350 (TREE_INT_CST_ELT): Likewise.
3351 (INT_CST_LT): Delete.
3352 (tree_int_cst_elt_check): New (two forms).
3353 (type_code_size): Update comment.
3354 (make_int_cst_stat, make_int_cst): New.
3355 (tree_to_double_int): Delete.
3356 (double_int_fits_to_tree_p): Delete.
3357 (force_fit_type_double): Delete.
3358 (build_int_cstu): Replace with out-of-line function.
3359 (build_int_cst_wide): Delete.
3360 (tree_int_cst_lt): Define inline.
3361 (tree_int_cst_le): New.
3362 (tree_int_cst_compare): Define inline.
3363 (tree_int_cst_min_precision): Take a signop rather than a bool.
3364 (wi::int_traits <const_tree>): New.
3365 (wi::int_traits <tree>): New.
3366 (wi::extended_tree): New.
3367 (wi::int_traits <wi::extended_tree>): New.
3368 (wi::to_widest): New.
3369 (wi::to_offset): New.
3370 (wi::fits_to_tree_p): New.
3371 (wi::min_value): New.
3372 (wi::max_value): New.
3373 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3374 (copy_tree_body_r): Likewise.
3375 * tree-object-size.c (compute_object_offset): Likewise.
3376 (addr_object_size): Likewise.
3377 * tree-predcom.c: Include wide-int-print.h.
3378 (struct dref_d): Change type of offset to widest_int.
3379 (dump_dref): Call wide-int printer.
3380 (aff_combination_dr_offset): Use wide-int interfaces.
3381 (determine_offset): Take a widest_int pointer rather than a
3383 (split_data_refs_to_components): Use wide-int interfaces.
3384 (suitable_component_p): Likewise.
3385 (order_drefs): Likewise.
3386 (add_ref_to_chain): Likewise.
3387 (valid_initializer_p): Likewise.
3388 (determine_roots_comp): Likewise.
3389 * tree-pretty-print.c: Include wide-int-print.h.
3390 (dump_generic_node): Use wide-int interfaces.
3391 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3392 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3393 (move_fixed_address_to_symbol): Likewise.
3394 (move_hint_to_base): Likewise.
3395 (move_pointer_to_base): Likewise.
3396 (move_variant_to_index): Likewise.
3397 (most_expensive_mult_to_index): Likewise.
3398 (addr_to_parts): Likewise.
3399 (copy_ref_info): Likewise.
3400 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3401 (indirect_refs_may_alias_p): Likewise.
3402 (stmt_kills_ref_p_1): Likewise.
3403 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3404 * tree-ssa-ccp.c: Update comment at top of file. Include
3406 (struct prop_value_d): Change type of mask to widest_int.
3407 (extend_mask): New function.
3408 (dump_lattice_value): Use wide-int interfaces.
3409 (get_default_value): Likewise.
3410 (set_constant_value): Likewise.
3411 (set_value_varying): Likewise.
3412 (valid_lattice_transition): Likewise.
3413 (set_lattice_value): Likewise.
3414 (value_to_double_int): Delete.
3415 (value_to_wide_int): New.
3416 (get_value_from_alignment): Use wide-int interfaces.
3417 (get_value_for_expr): Likewise.
3418 (do_dbg_cnt): Likewise.
3419 (ccp_finalize): Likewise.
3420 (ccp_lattice_meet): Likewise.
3421 (bit_value_unop_1): Use widest_ints rather than double_ints.
3422 (bit_value_binop_1): Likewise.
3423 (bit_value_unop): Use wide-int interfaces.
3424 (bit_value_binop): Likewise.
3425 (bit_value_assume_aligned): Likewise.
3426 (evaluate_stmt): Likewise.
3427 (ccp_fold_stmt): Likewise.
3428 (visit_cond_stmt): Likewise.
3429 (ccp_visit_stmt): Likewise.
3430 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3431 (constant_pointer_difference): Likewise.
3432 (associate_pointerplus): Likewise.
3433 (combine_conversions): Likewise.
3434 * tree-ssa-loop.h: Include wide-int.h.
3435 (struct tree_niter_desc): Change type of max to widest_int.
3436 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3437 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3438 (remove_redundant_iv_tests): Likewise.
3439 (canonicalize_loop_induction_variables): Likewise.
3440 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3441 (constant_multiple_of): Take a widest_int pointer instead of
3442 a double_int pointer.
3443 (get_computation_aff): Use wide-int interfaces.
3444 (ptr_difference_cost): Likewise.
3445 (difference_cost): Likewise.
3446 (get_loop_invariant_expr_id): Likewise.
3447 (get_computation_cost_at): Likewise.
3448 (iv_elimination_compare_lt): Likewise.
3449 (may_eliminate_iv): Likewise.
3450 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3451 instead of double_int.
3452 (max_loop_iterations): Likewise.
3453 (max_stmt_executions): Likewise.
3454 (estimated_stmt_executions): Likewise.
3455 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3456 (split_to_var_and_offset): Use wide-int interfaces.
3457 (determine_value_range): Likewise.
3458 (bound_difference_of_offsetted_base): Likewise.
3459 (bounds_add): Take a widest_int instead of a double_int.
3460 (number_of_iterations_ne_max): Use wide-int interfaces.
3461 (number_of_iterations_ne): Likewise.
3462 (number_of_iterations_lt_to_ne): Likewise.
3463 (assert_loop_rolls_lt): Likewise.
3464 (number_of_iterations_lt): Likewise.
3465 (number_of_iterations_le): Likewise.
3466 (number_of_iterations_cond): Likewise.
3467 (number_of_iterations_exit): Likewise.
3468 (finite_loop_p): Likewise.
3469 (derive_constant_upper_bound_assign): Likewise.
3470 (derive_constant_upper_bound): Return a widest_int.
3471 (derive_constant_upper_bound_ops): Likewise.
3472 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3473 (record_estimate): Take a widest_int rather than a double_int.
3474 (record_nonwrapping_iv): Use wide-int interfaces.
3475 (double_int_cmp): Delete.
3476 (wide_int_cmp): New.
3477 (bound_index): Take a widest_int rather than a double_int.
3478 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3479 (maybe_lower_iteration_bound): Likewise.
3480 (estimate_numbers_of_iterations_loop): Likewise.
3481 (estimated_loop_iterations): Take a widest_int pointer than than
3482 a double_int pointer.
3483 (estimated_loop_iterations_int): Use wide-int interfaces.
3484 (max_loop_iterations): Take a widest_int pointer than than
3485 a double_int pointer.
3486 (max_loop_iterations_int): Use wide-int interfaces.
3487 (max_stmt_executions): Take a widest_int pointer than than
3488 a double_int pointer.
3489 (estimated_stmt_executions): Likewise.
3490 (n_of_executions_at_most): Use wide-int interfaces.
3491 (scev_probably_wraps_p): Likewise.
3492 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3494 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3496 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3497 double_ints. Adjust for trailing_wide_ints <3> representation.
3498 (set_nonzero_bits): Likewise.
3499 (get_range_info): Return wide_ints rather than double_ints.
3500 Adjust for trailing_wide_ints <3> representation.
3501 (get_nonzero_bits): Likewise.
3502 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3504 * tree-ssanames.h (struct range_info_def): Replace min, max and
3505 nonzero_bits with a trailing_wide_ints <3>.
3506 (set_range_info): Use wide_int_refs rather than double_ints.
3507 (set_nonzero_bits): Likewise.
3508 (get_range_info): Return wide_ints rather than double_ints.
3509 (get_nonzero_bits): Likewise.
3510 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3511 * tree-ssa-pre.c (phi_translate_1): Likewise.
3512 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3513 (acceptable_pow_call): Likewise.
3514 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3516 (vn_reference_fold_indirect): Likewise.
3517 (vn_reference_maybe_forwprop_address): Likewise.
3518 (valueize_refs_1): Likewise.
3519 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3520 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3521 tree_int_cst_lt and tree_int_cst_le.
3522 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3524 (streamer_alloc_tree): Likewise.
3525 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3526 (streamer_write_tree_header): Likewise.
3527 (streamer_write_integer_cst): Likewise.
3528 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3529 (build_constructors): Likewise.
3530 (array_value_type): Likewise.
3531 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3532 (vect_check_gather): Likewise.
3533 * tree-vect-generic.c (build_replicated_const): Likewise.
3534 (expand_vector_divmod): Likewise.
3535 * tree-vect-loop.c (vect_transform_loop): Likewise.
3536 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3537 (vect_do_peeling_for_alignment): Likewise.
3538 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3539 * tree-vrp.c: Include wide-int.h.
3540 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3541 (extract_range_from_assert): Use wide-int interfaces.
3542 (vrp_int_const_binop): Likewise.
3543 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3544 double_int pointers.
3545 (ranges_from_anti_range): Use wide-int interfaces.
3546 (quad_int_cmp): Delete.
3547 (quad_int_pair_sort): Likewise.
3548 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3549 (extract_range_from_unary_expr_1): Likewise.
3550 (adjust_range_with_scev): Likewise.
3551 (masked_increment): Take and return wide_ints rather than double_ints.
3552 (register_edge_assert_for_2): Use wide-int interfaces.
3553 (check_array_ref): Likewise.
3554 (search_for_addr_array): Likewise.
3555 (maybe_set_nonzero_bits): Likewise.
3556 (union_ranges): Pass an integer of the correct type instead of
3557 using integer_one_node.
3558 (intersect_ranges): Likewise.
3559 (simplify_truth_ops_using_ranges): Likewise.
3560 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3561 (range_fits_type_p): Likewise.
3562 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3564 (simplify_conversion_using_ranges): Use wide-int interfaces.
3565 (simplify_float_conversion_using_ranges): Likewise.
3566 (vrp_finalize): Likewise.
3567 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3568 (gimple_stringops_transform): Likewise.
3569 * varasm.c (decode_addr_const): Likewise.
3570 (const_hash_1): Likewise.
3571 (const_rtx_hash_1): Likewise
3572 (output_constant): Likewise.
3573 (array_size_for_constructor): Likewise.
3574 (output_constructor_regular_field): Likewise.
3575 (output_constructor_bitfield): Likewise.
3576 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3577 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3579 * gencheck.c: Define BITS_PER_UNIT.
3582 * wide-int-print.cc: New.
3583 * wide-int-print.h: New.
3585 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3587 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3589 2014-05-06 Richard Biener <rguenther@suse.de>
3591 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3592 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3593 (TODO_verify_all): Adjust.
3594 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3595 TODO_verify_stmts and TODO_verify_rtl_sharing.
3596 * bb-reorder.c: Likewise.
3597 * cfgexpand.c: Likewise.
3598 * cprop.c: Likewise.
3600 * function.c: Likewise.
3601 * fwprop.c: Likewise.
3603 * gimple-ssa-isolate-paths.c: Likewise.
3604 * gimple-ssa-strength-reduction.c: Likewise.
3605 * ipa-split.c: Likewise.
3606 * loop-init.c: Likewise.
3607 * loop-unroll.c: Likewise.
3608 * lower-subreg.c: Likewise.
3609 * modulo-sched.c: Likewise.
3610 * postreload-gcse.c: Likewise.
3611 * predict.c: Likewise.
3612 * recog.c: Likewise.
3613 * sched-rgn.c: Likewise.
3614 * store-motion.c: Likewise.
3615 * tracer.c: Likewise.
3616 * trans-mem.c: Likewise.
3617 * tree-call-cdce.c: Likewise.
3618 * tree-cfg.c: Likewise.
3619 * tree-cfgcleanup.c: Likewise.
3620 * tree-complex.c: Likewise.
3621 * tree-eh.c: Likewise.
3622 * tree-emutls.c: Likewise.
3623 * tree-if-conv.c: Likewise.
3624 * tree-into-ssa.c: Likewise.
3625 * tree-loop-distribution.c: Likewise.
3626 * tree-object-size.c: Likewise.
3627 * tree-parloops.c: Likewise.
3628 * tree-pass.h: Likewise.
3629 * tree-sra.c: Likewise.
3630 * tree-ssa-ccp.c: Likewise.
3631 * tree-ssa-copy.c: Likewise.
3632 * tree-ssa-copyrename.c: Likewise.
3633 * tree-ssa-dce.c: Likewise.
3634 * tree-ssa-dom.c: Likewise.
3635 * tree-ssa-dse.c: Likewise.
3636 * tree-ssa-forwprop.c: Likewise.
3637 * tree-ssa-ifcombine.c: Likewise.
3638 * tree-ssa-loop-ch.c: Likewise.
3639 * tree-ssa-loop-ivcanon.c: Likewise.
3640 * tree-ssa-loop.c: Likewise.
3641 * tree-ssa-math-opts.c: Likewise.
3642 * tree-ssa-phiopt.c: Likewise.
3643 * tree-ssa-phiprop.c: Likewise.
3644 * tree-ssa-pre.c: Likewise.
3645 * tree-ssa-reassoc.c: Likewise.
3646 * tree-ssa-sink.c: Likewise.
3647 * tree-ssa-strlen.c: Likewise.
3648 * tree-ssa-tail-merge.c: Likewise.
3649 * tree-ssa-uncprop.c: Likewise.
3650 * tree-switch-conversion.c: Likewise.
3651 * tree-tailcall.c: Likewise.
3652 * tree-vect-generic.c: Likewise.
3653 * tree-vectorizer.c: Likewise.
3654 * tree-vrp.c: Likewise.
3656 * var-tracking.c: Likewise.
3657 * bt-load.c: Likewise.
3658 * cfgcleanup.c: Likewise.
3659 * combine-stack-adj.c: Likewise.
3660 * combine.c: Likewise.
3661 * compare-elim.c: Likewise.
3662 * config/epiphany/resolve-sw-modes.c: Likewise.
3663 * config/i386/i386.c: Likewise.
3664 * config/mips/mips.c: Likewise.
3665 * config/s390/s390.c: Likewise.
3666 * config/sh/sh_treg_combine.cc: Likewise.
3667 * config/sparc/sparc.c: Likewise.
3670 * final.c: Likewise.
3671 * ifcvt.c: Likewise.
3672 * mode-switching.c: Likewise.
3673 * passes.c: Likewise.
3674 * postreload.c: Likewise.
3676 * reg-stack.c: Likewise.
3677 * regcprop.c: Likewise.
3678 * regrename.c: Likewise.
3681 2014-05-06 Richard Biener <rguenther@suse.de>
3684 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3685 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3687 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3690 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3692 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3693 Tom de Vries <tom@codesourcery.com>
3695 * target.def (call_fusage_contains_non_callee_clobbers): New
3697 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3699 (@node Miscellaneous Register Hooks): New node.
3700 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3701 * doc/tm.texi: Regenerate.
3703 2014-05-05 Marek Polacek <polacek@redhat.com>
3706 * opts.c (common_handle_option): Call error_at instead of warning_at.
3708 2014-05-05 Richard Biener <rguenther@suse.de>
3710 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3711 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3712 under the TODO_verify_il umbrella.
3714 2014-05-05 Richard Biener <rguenther@suse.de>
3716 * passes.c (execute_function_todo): Move TODO_verify_flow under
3717 the TODO_verify_ul umbrella.
3719 2014-05-05 Richard Biener <rguenther@suse.de>
3722 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3723 X & CST away from a CST that is the mask of a mode.
3725 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3727 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3728 int argument to enum machine_mode.
3729 (picochip_class_max_nregs): Ditto.
3730 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3731 (picochip_class_max_nregs): Ditto.
3733 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3735 * target.def: Add new target hook.
3736 * doc/tm.texi: Regenerate.
3737 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3738 * targhooks.c (default_keep_leaf_when_profiled): New function.
3740 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3741 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3743 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3745 PR tree-optimization/60363
3746 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3747 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3748 (update_destination_phis): New parameter.
3749 (create_edge_and_update_destination_phis): Ditto.
3750 (ssa_fix_duplicate_block_edges): Pass new arguments.
3751 (thread_single_edge): Ditto.
3753 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3755 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3756 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3757 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3758 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3759 Use RS6000_BTM_HARD_FLOAT.
3760 (BU_MISC_2): Likewise.
3761 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3762 RS6000_BTM_HARD_FLOAT.
3763 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3765 (rs6000_invalid_builtin): Add hard floating builtin support.
3766 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3767 hard float builtins.
3768 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3770 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3772 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3773 Add missing function* argument.
3775 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3777 * lra-constraints.c (valid_address_p): Move earlier in file.
3778 Add a constraint argument to the address_info version.
3779 (satisfies_memory_constraint_p): New function.
3780 (satisfies_address_constraint_p): Likewise.
3781 (process_alt_operands, curr_insn_transform): Use them.
3782 (process_address): Pass the constraint to valid_address_p when
3783 checking address operands.
3785 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3787 * config/mips/mips.c (mips_isa_rev): New variable.
3788 (mips_set_architecture): Set it.
3789 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3791 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3792 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3793 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3794 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3795 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3796 conditions in terms of mips_isa_rev.
3797 (mips_isa_rev): Declare.
3799 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3801 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3802 (prob_unlikely, prob_likely): Make variables const.
3804 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3806 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3808 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3810 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3812 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3814 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3815 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3816 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3818 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3819 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3821 Replace usage of ROUND_REG with sh_round_reg.
3822 Use CEIL instead of ROUND_ADVANCE.
3824 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3827 * config/sh/sh.c: Include stdlib headers before everything else.
3829 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3831 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3832 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3833 (gimplify_adjust_omp_clauses): Simd region is never
3834 directly nested in combined parallel. Instead, for linear
3835 with copyin/copyout, if in combined for simd loop, make decl
3836 firstprivate/lastprivate on OMP_FOR.
3837 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3838 expand_omp_for_static_chunk): When setting endvar, also set
3839 fd->loop.v to the same value.
3841 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3843 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3845 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3847 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3850 2014-05-02 Marek Polacek <polacek@redhat.com>
3852 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3854 2014-05-02 Kito Cheng <kito@0xlab.org>
3856 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3857 to a C expression marco.
3858 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3859 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3860 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3861 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3862 HONOR_REG_ALLOC_ORDER.
3863 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3865 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3867 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3869 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3871 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3873 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
3875 * tree-if-conv.c (is_cond_scalar_reduction): New function.
3876 (convert_scalar_cond_reduction): Likewise.
3877 (predicate_scalar_phi): Add recognition and transformation
3878 of simple conditioanl reduction to be vectorizable.
3880 2014-05-01 Marek Polacek <polacek@redhat.com>
3883 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
3885 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
3887 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
3888 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
3889 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
3890 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
3891 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
3892 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
3893 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
3894 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
3896 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3898 * config/arc/arc.opt (mlra): Move comment above option name
3899 to avoid mis-parsing as language options.
3901 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3903 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
3904 * config/sol2.h: ... here.
3905 * config/sol2-10.h: Remove.
3907 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
3908 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
3909 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
3910 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
3911 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
3912 * config/sol2.h: ... here.
3913 (SECTION_NAME_FORMAT): Don't redefine.
3914 (STARTFILE_ARCH32_SPEC): Rename to ...
3915 (STARTFILE_ARCH_SPEC): ... this.
3916 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
3917 * config/sparc/sol2.h: ... here.
3918 (SECTION_NAME_FORMAT): Don't undef.
3919 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
3920 (SUBTARGET_EXTRA_SPECS): Remove.
3921 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
3923 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
3924 (MD_STARTFILE_PREFIX): Remove.
3925 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
3926 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
3927 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
3928 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
3929 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
3930 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
3931 * config/i386/sol2.h: ... here.
3932 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
3933 * config/i386/sol2-bi.h: Remove.
3934 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
3935 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
3937 * config/i386/t-sol2-64: Rename to ...
3938 * config/i386/t-sol2: ... this.
3939 * config/sparc/t-sol2-64: Rename to ...
3940 * config/sparc/t-sol2: ... this.
3942 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
3943 sol2_tm_file_head, sol2_tm_file_tail.
3944 Include ${cpu_type}/sol2.h before sol2.h.
3946 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
3947 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
3948 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
3949 Reflect i386/t-sol2-64 renaming.
3950 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
3951 Reflect sparc/t-sol2-64 renaming.
3953 2014-04-30 Richard Biener <rguenther@suse.de>
3955 * passes.c (execute_function_todo): Move TODO_verify_stmts
3956 and TODO_verify_ssa under the TODO_verify_il umbrella.
3957 * tree-ssa.h (verify_ssa): Adjust prototype.
3958 * tree-ssa.c (verify_ssa): Add parameter to tell whether
3959 we should verify SSA operands.
3960 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
3961 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
3962 whether we should verify whether not throwing stmts have EH info.
3963 * graphite-scop-detection.c (create_sese_edges): Adjust.
3964 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
3965 * tree-eh.c (lower_try_finally_switch): Do not add the
3966 default case label twice.
3968 2014-04-30 Marek Polacek <polacek@redhat.com>
3970 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
3971 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
3972 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
3973 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
3975 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3977 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
3978 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
3979 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
3980 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
3981 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
3982 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
3983 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
3984 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
3986 2014-04-29 David Malcolm <dmalcolm@redhat.com>
3988 * tree-cfg.c (dump_function_to_file): Dump the return type of
3989 functions, in a line to itself before the function body, mimicking
3990 the layout of a C function.
3992 2014-04-29 Jakub Jelinek <jakub@redhat.com>
3994 PR tree-optimization/60971
3995 * tree-tailcall.c (process_assignment): Reject conversions which
3998 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4000 * calls.c (initialize_argument_information): Always treat
4001 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4002 (expand_call): Likewise.
4003 (emit_library_call_calue_1): Likewise.
4004 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4005 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4008 2014-04-29 Nick Clifton <nickc@redhat.com>
4010 * config/msp430/msp430.md (umulsidi): Fix typo.
4011 (mulhisi3): Enable even inside interrupt handlers.
4012 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4013 bigger return address pushed in large mode.
4015 2014-04-29 Nick Clifton <nickc@redhat.com>
4017 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4018 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4020 * config/m32r/m32r.c (init_reg_tables): Likewise.
4021 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4022 enum to hold the modes.
4024 2014-04-29 Richard Biener <rguenther@suse.de>
4026 * dominance.c (free_dominance_info): Add overload with
4028 (dom_info_state): Likewise.
4029 (dom_info_available_p): Likewise.
4030 * basic-block.h (free_dominance_info, dom_info_state,
4031 dom_info_available_p): Declare overloads.
4032 * passes.c (execute_function_todo): Verify that verifiers
4033 don't change dominator info state. Drop dominator info
4034 for IPA pass invocations.
4035 * cgraph.c (release_function_body): Restore asserts that
4036 dominator information is released.
4038 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4040 * doc/invoke.texi: Fix typo.
4041 * tree-vrp.c: Fix typos.
4042 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4044 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4046 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4048 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4050 * config/aarch64/aarch64-builtins.c
4051 (aarch64_types_storestruct_lane_qualifiers): New.
4052 (TYPES_STORESTRUCT_LANE): Likewise.
4053 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4054 (st3_lane): Likewise.
4055 (st4_lane): Likewise.
4056 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4057 (vec_store_lanesci_lane<mode>): Likewise.
4058 (vec_store_lanesxi_lane<mode>): Likewise.
4059 (aarch64_st2_lane<VQ:mode>): Likewise.
4060 (aarch64_st3_lane<VQ:mode>): Likewise.
4061 (aarch64_st4_lane<VQ:mode>): Likewise.
4062 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4063 * config/aarch64/arm_neon.h
4064 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4065 use new macro arguments.
4066 (__ST3_LANE_FUNC): Likewise.
4067 (__ST4_LANE_FUNC): Likewise.
4068 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4069 (V_THREE_ELEM): Likewise.
4070 (V_FOUR_ELEM): Likewise.
4072 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4074 * doc/gimple.texi: Replace the description of the now-defunct
4075 union gimple_statement_d with a diagram showing the
4076 gimple_statement_base class hierarchy and its relationships to
4077 the GSS_ and GIMPLE_ enums.
4079 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4081 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4082 * config/aarch64/aarch64.c
4083 (aarch64_cannot_change_mode_class): Weaken conditions.
4084 (aarch64_modes_tieable_p): New.
4085 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4087 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4089 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4090 (loadsync_<mode>): Change mode.
4091 (load_quadpti, store_quadpti): New.
4092 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4093 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4095 2014-04-28 Martin Jambor <mjambor@suse.cz>
4097 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4098 same alias type as the original statement.
4099 (subreplacement_assignment_data): New type.
4100 (handle_unscalarized_data_in_subtree): New type of parameter,
4101 generate new memory accesses with same alias type as the original
4103 (load_assign_lhs_subreplacements): Likewise.
4104 (sra_modify_constructor_assign): Generate new memory accesses with
4105 same alias type as the original statement.
4107 2014-04-28 Richard Biener <rguenther@suse.de>
4109 * tree-pass.h (TODO_verify_il): Define.
4110 (TODO_verify_all): Complete properly.
4111 * passes.c (execute_function_todo): Move existing loop-closed
4112 SSA verification under TODO_verify_il.
4113 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4114 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4115 Fix tree sharing issue.
4117 2014-04-28 Richard Biener <rguenther@suse.de>
4120 * builtins.def (DEF_C11_BUILTIN): Add.
4121 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4122 * coretypes.h (enum function_class): Add function_c11_misc.
4123 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4124 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4125 (call_may_clobber_ref_p_1): Likewise.
4126 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4127 (mark_all_reaching_defs_necessary_1): Likewise.
4128 (propagate_necessity): Likewise.
4129 (eliminate_unnecessary_stmts): Likewise.
4130 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4132 2014-04-28 Richard Biener <rguenther@suse.de>
4134 * tree-vrp.c (vrp_var_may_overflow): Remove.
4135 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4136 with overflow immediately bump to one before that value and
4137 let iteration figure out overflow status.
4139 2014-04-28 Richard Biener <rguenther@suse.de>
4141 * configure.ac: Do valgrind header checks unconditionally.
4142 Add --enable-valgrind-annotations.
4143 * system.h: Guard valgrind header inclusion with
4144 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4145 * alloc-pool.c (pool_alloc, pool_free): Use
4146 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4147 to guard possibly dead code.
4148 * config.in: Regenerated.
4149 * configure: Likewise.
4151 2014-04-28 Jeff Law <law@redhat.com>
4153 PR tree-optimization/60902
4154 * tree-ssa-threadedge.c
4155 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4156 over real defs when invalidating outputs from statements that do not
4157 produce useful outputs for threading.
4159 2014-04-28 Richard Biener <rguenther@suse.de>
4161 PR tree-optimization/60979
4162 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4163 SCOPs that end in a block with a successor with abnormal
4166 2014-04-28 Richard Biener <rguenther@suse.de>
4168 * tree-pass.h (execute_pass_list): Adjust prototype.
4169 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4170 (do_per_function): Change callback signature, push all actual
4171 work to the callbals.
4172 (do_per_function_toporder): Likewise.
4173 (execute_function_dump): Adjust.
4174 (execute_function_todo): Likewise.
4175 (clear_last_verified): Likewise.
4176 (verify_curr_properties): Likewise.
4177 (update_properties_after_pass): Likewise.
4178 (execute_pass_list_1): Split out from ...
4179 (execute_pass_list): ... here. Adjust.
4180 (execute_ipa_pass_list): Likewise.
4181 * cgraphunit.c (cgraph_add_new_function): Adjust.
4182 (analyze_function): Likewise.
4183 (expand_function): Likewise.
4184 * cgraph.c (release_function_body): Free dominance info
4185 here instead of asserting it was magically freed elsewhere.
4187 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4189 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4190 * configure: Regenerate.
4191 * config/sparc/sparc.opt (muser-mode): New option.
4192 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4194 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4195 * doc/invoke.texi (SPARC options): Document -muser-mode.
4197 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4199 * cselib.c (find_slot_memmode): Delete.
4200 (cselib_hasher): Change compare_type to a struct.
4201 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4203 (preserve_constants_and_equivs): Adjust for new compare_type.
4204 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4205 (wrap_constant): Delete.
4206 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4208 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4210 * doc/install.texi (Building with profile feedback): Remove
4213 2014-04-26 Tom de Vries <tom@codesourcery.com>
4215 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4218 2014-04-25 Cary Coutant <ccoutant@google.com>
4221 * dwarf2out.c (should_move_die_to_comdat): A type definition
4222 can contain a subprogram definition, but don't move it to a
4224 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4225 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4227 (clone_tree_hash): Rename to...
4228 (clone_tree_partial): ...this; change callers. Copy
4229 DW_TAG_subprogram DIEs as declarations.
4230 (copy_decls_walk): Don't copy children of a declaration into a
4233 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4236 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4239 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4241 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4242 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4244 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4247 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4249 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4251 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4253 PR tree-optimization/60930
4254 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4255 creating a multiply candidate by folding two constant
4256 multiplicands when the result overflows.
4258 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4260 PR tree-optimization/60960
4261 * tree-vect-generic.c (expand_vector_operation): Only call
4262 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4264 2014-04-25 Tom de Vries <tom@codesourcery.com>
4266 * expr.c (clobber_reg_mode): New function.
4267 * expr.h (clobber_reg): New function.
4269 2014-04-25 Tom de Vries <tom@codesourcery.com>
4271 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4274 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4275 Tom de Vries <tom@codesourcery.com>
4277 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4279 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4280 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4281 new argument to find_all_hard_reg_sets call.
4283 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4285 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4286 Use HOST_WIDE_INT_C for mask literal.
4287 (aarch_rev16_shleft_mask_imm_p): Likewise.
4289 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4292 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4294 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4296 PR preprocessor/56540
4297 * config/i386/i386-c.c (ix86_target_macros): Define
4298 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4300 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4302 * configure.ac (tga_func): Remove.
4303 (LIB_TLS_SPEC): Remove.
4304 * configure: Regenerate.
4305 * config.in: Regenerate.
4306 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4308 2014-04-25 Richard Biener <rguenther@suse.de>
4311 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4312 call stmt use/clobber sets during stmt walk instead of
4313 walking the possibly incomplete set of caller edges.
4315 2014-04-25 Richard Biener <rguenther@suse.de>
4318 * passes.c (apply_ipa_transforms): Inline into only caller ...
4319 (execute_one_pass): ... here. Properly bring in function
4320 bodies for nodes we want to apply IPA transforms to.
4322 2014-04-24 Cong Hou <congh@google.com>
4324 PR tree-optimization/60896
4325 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4326 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4327 (vect_mark_pattern_stmts): Set the def type of all statements in
4328 PATTERN_DEF_SEQ as vect_internal_def.
4330 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4332 * doc/extend.texi (PowerPC Built-in Functions): Document new
4333 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4334 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4336 * config/rs6000/predicates.md (const_0_to_3_operand): New
4337 predicate to match 0..3 integer constants.
4339 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4340 to support adding miscellaneous builtin functions.
4341 (BU_DFP_MISC_2): Likewise.
4342 (BU_P7_MISC_1): Likewise.
4343 (BU_P7_MISC_2): Likewise.
4344 (BU_P8V_MISC_3): Likewise.
4345 (BU_MISC_1): Likewise.
4346 (BU_MISC_2): Likewise.
4347 (DIVWE): Add extended divide builtin functions.
4350 (DIVWEUO): Likewise.
4354 (DIVDEUO): Likewise.
4355 (DXEX): Add decimal floating-point builtin functions.
4358 (DDEDPDQ): Likewise.
4360 (DENBCDQ): Likewise.
4367 (CDTBCD): Add new BCD builtin functions.
4371 (BCDADD_LT): Likewise.
4372 (BCDADD_EQ): Likewise.
4373 (BCDADD_GT): Likewise.
4374 (BCDADD_OV): Likewise.
4376 (BCDSUB_LT): Likewise.
4377 (BCDSUB_EQ): Likewise.
4378 (BCDSUB_GT): Likewise.
4379 (BCDSUB_OV): Likewise.
4380 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4381 (UNPACK_TD): Likewise.
4382 (PACK_TF): Likewise.
4383 (UNPACK_TF): Likewise.
4384 (UNPACK_TF_0): Likewise.
4385 (UNPACK_TF_1): Likewise.
4386 (PACK_V1TI): Likewise.
4387 (UNPACK_V1TI): Likewise.
4389 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4390 support for decimal floating point builtin functions.
4391 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4392 functions that take constant arguments.
4393 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4394 (rs6000_init_builtins): Setup long double, _Decimal64, and
4395 _Decimal128 types for new builtin functions.
4396 (builtin_function_type): Set the unsigned flags appropriately for
4397 the new builtin functions.
4398 (rs6000_opt_masks): Add support for decimal floating point builtin
4401 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4402 floating point builtin functions.
4403 (RS6000_BTM_COMMON): Likewise.
4404 (RS6000_BTI_long_double): Likewise.
4405 (RS6000_BTI_dfloat64): Likewise.
4406 (RS6000_BTI_dfloat128): Likewise.
4407 (long_double_type_internal_node): Likewise.
4408 (dfloat64_type_internal_node): Likewise.
4409 (dfloat128_type_internal_node): Likewise.
4411 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4412 2.07 bcd arithmetic instructions.
4413 (UNSPEC_BCDSUB): Likewise.
4414 (UNSPEC_BCD_OVERFLOW): Likewise.
4415 (UNSPEC_BCD_ADD_SUB): Likewise.
4416 (bcd_add_sub): Likewise.
4417 (BCD_TEST): Likewise.
4418 (bcd<bcd_add_sub>): Likewise.
4419 (bcd<bcd_add_sub>_test): Likewise.
4420 (bcd<bcd_add_sub>_test2): Likewise.
4421 (bcd<bcd_add_sub>_<code>): Likewise.
4422 (peephole2 for combined bcd ops): Likewise.
4424 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4425 decimal floating point builtin functions.
4426 (UNSPEC_DENBCD): Likewise.
4427 (UNSPEC_DXEX): Likewise.
4428 (UNSPEC_DIEX): Likewise.
4429 (UNSPEC_DSCLI): Likewise.
4430 (UNSPEC_DSCRI): Likewise.
4431 (D64_D128): Likewise.
4432 (dfp_suffix): Likewise.
4433 (dfp_ddedpd_<mode>): Likewise.
4434 (dfp_denbcd_<mode>): Likewise.
4435 (dfp_dxex_<mode>): Likewise.
4436 (dfp_diex_<mode>): Likewise.
4437 (dfp_dscli_<mode>): Likewise.
4438 (dfp_dscri_<mode>): Likewise.
4440 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4442 (UNSPEC_CDTBCD): Likewise.
4443 (UNSPEC_CBCDTD): Likewise.
4444 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4445 (UNSPEC_DIVEO): Likewise.
4446 (UNSPEC_DIVEU): Likewise.
4447 (UNSPEC_DIVEUO): Likewise.
4448 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4449 pack/unpack 128-bit types.
4450 (UNSPEC_PACK_128BIT): Likewise.
4451 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4452 (udiv<mode>3): Use idiv_ldiv mode attribute.
4453 (div<mode>3): Likewise.
4454 (addg6s): Add new BCD builtin functions.
4457 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4458 (div_extend): Likewise.
4459 (div<div_extend>_<mode>"): Likewise.
4460 (FP128_64): Add support for new builtin functions to pack/unpack
4462 (unpack<mode>): Likewise.
4463 (unpacktf_0): Likewise.
4464 (unpacktf_1): Likewise.
4465 (unpack<mode>_dm): Likewise.
4466 (unpack<mode>_nodm): Likewise.
4467 (pack<mode>): Likewise.
4468 (unpackv1ti): Likewise.
4469 (packv1ti): Likewise.
4471 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4473 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4476 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4478 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4479 * gimplify.c (omp_is_private): Change last argument's type to int.
4480 Only diagnose lastprivate if the simd argument is 1, only diagnose
4481 linear if the simd argument is 2.
4482 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4483 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4484 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4485 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4486 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4487 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4488 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4489 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4490 * tree-nested.c (convert_nonlocal_omp_clauses,
4491 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4493 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4496 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4499 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4501 * flag-types.h (enum ivar_visibility): Add.
4503 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4505 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4506 function * argument.
4508 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4510 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4512 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4513 Tom de Vries <tom@codesourcery.com>
4515 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4516 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4518 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4519 * emit-rtl.c (try_split): Same.
4521 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4522 Tom de Vries <tom@codesourcery.com>
4524 * common.opt (fuse-caller-save): New option.
4526 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4528 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4529 elements for big-endian.
4531 2014-04-24 Richard Biener <rguenther@suse.de>
4533 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4534 during TER and instead use the sepops interface for expanding
4535 non-GIMPLE_SINGLE_RHS.
4537 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4539 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4540 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4542 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4544 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4545 assembler 64-bit option.
4546 * configure: Regenerate.
4548 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4550 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4551 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4552 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4553 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4554 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4556 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4558 * config/aarch64/aarch64-builtins.c
4559 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4560 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4561 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4562 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4564 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4565 (Vrevsuff): New mode attribute.
4567 2014-04-24 Terry Guo <terry.guo@arm.com>
4569 * config/arm/arm.h (machine_function): Define variable
4570 after_arm_reorg here.
4571 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4572 (arm_split_constant): Update the way to access variable
4575 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4577 2014-04-23 Tom de Vries <tom@codesourcery.com>
4579 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4581 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4583 * is-a.h: Update comments to reflect the following changes to the
4584 "pointerness" of the API, making the template parameter match the
4585 return type, allowing use of is-a.h with typedefs of pointers.
4586 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4587 that the return type matches the parameter to the is_a_helper.
4589 (dyn_cast): Likewise.
4591 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4592 pointer from the is-a.h API.
4594 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4595 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4597 (is_a_helper <varpool_node>::test): Likewise, convert to...
4598 (is_a_helper <varpool_node *>::test): ...this.
4600 (varpool_first_variable): Update for removal of implicit pointer
4601 from the is-a.h API.
4602 (varpool_next_variable): Likewise.
4603 (varpool_first_static_initializer): Likewise.
4604 (varpool_next_static_initializer): Likewise.
4605 (varpool_first_defined_variable): Likewise.
4606 (varpool_next_defined_variable): Likewise.
4607 (cgraph_first_defined_function): Likewise.
4608 (cgraph_next_defined_function): Likewise.
4609 (cgraph_first_function): Likewise.
4610 (cgraph_next_function): Likewise.
4611 (cgraph_first_function_with_gimple_body): Likewise.
4612 (cgraph_next_function_with_gimple_body): Likewise.
4613 (cgraph_alias_target): Likewise.
4614 (varpool_alias_target): Likewise.
4615 (cgraph_function_or_thunk_node): Likewise.
4616 (varpool_variable_node): Likewise.
4617 (symtab_real_symbol_p): Likewise.
4618 * cgraphunit.c (referred_to_p): Likewise.
4619 (analyze_functions): Likewise.
4620 (handle_alias_pairs): Likewise.
4621 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4622 * gimple-ssa.h (gimple_vuse_op): Likewise.
4623 (gimple_vdef_op): Likewise.
4624 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4625 * gimple.c (gimple_build_asm_1): Likewise.
4626 (gimple_build_try): Likewise.
4627 (gimple_build_resx): Likewise.
4628 (gimple_build_eh_dispatch): Likewise.
4629 (gimple_build_omp_for): Likewise.
4630 (gimple_omp_for_set_clauses): Likewise.
4632 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4633 (is_a_helper <gimple_statement_asm *>::test): ...this.
4634 (is_a_helper <gimple_statement_bind>::test): Convert to...
4635 (is_a_helper <gimple_statement_bind *>::test): ...this.
4636 (is_a_helper <gimple_statement_call>::test): Convert to...
4637 (is_a_helper <gimple_statement_call *>::test): ...this.
4638 (is_a_helper <gimple_statement_catch>::test): Convert to...
4639 (is_a_helper <gimple_statement_catch *>::test): ...this.
4640 (is_a_helper <gimple_statement_resx>::test): Convert to...
4641 (is_a_helper <gimple_statement_resx *>::test): ...this.
4642 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4643 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4644 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4645 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4646 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4647 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4648 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4649 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4650 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4651 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4652 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4653 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4654 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4655 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4656 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4657 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4658 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4659 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4660 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4661 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4662 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4663 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4664 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4665 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4666 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4667 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4668 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4669 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4670 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4671 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4672 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4673 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4674 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4675 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4676 (is_a_helper <gimple_statement_phi>::test): Convert to...
4677 (is_a_helper <gimple_statement_phi *>::test): ...this.
4678 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4679 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4680 (is_a_helper <gimple_statement_try>::test): Convert to...
4681 (is_a_helper <gimple_statement_try *>::test): ...this.
4682 (is_a_helper <gimple_statement_wce>::test): Convert to...
4683 (is_a_helper <gimple_statement_wce *>::test): ...this.
4684 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4685 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4686 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4687 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4688 (is_a_helper <const gimple_statement_call>::test): Convert to...
4689 (is_a_helper <const gimple_statement_call *>::test): ...this.
4690 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4691 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4692 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4693 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4694 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4695 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4696 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4697 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4698 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4700 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4702 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4704 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4706 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4707 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4708 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4710 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4711 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4713 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4714 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4715 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4716 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4717 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4718 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4720 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4721 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4722 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4723 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4725 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4726 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4727 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4728 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4729 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4730 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4731 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4732 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4733 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4734 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4735 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4736 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4737 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4738 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4739 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4740 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4742 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4744 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4745 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4747 (gimple_use_ops): Update for removal of implicit pointer from the
4749 (gimple_set_use_ops): Likewise.
4750 (gimple_vuse): Likewise.
4751 (gimple_vdef): Likewise.
4752 (gimple_vuse_ptr): Likewise.
4753 (gimple_vdef_ptr): Likewise.
4754 (gimple_set_vuse): Likewise.
4755 (gimple_set_vdef): Likewise.
4756 (gimple_omp_return_set_lhs): Likewise.
4757 (gimple_omp_return_lhs): Likewise.
4758 (gimple_omp_return_lhs_ptr): Likewise.
4759 (gimple_call_fntype): Likewise.
4760 (gimple_call_set_fntype): Likewise.
4761 (gimple_call_set_internal_fn): Likewise.
4762 (gimple_call_use_set): Likewise.
4763 (gimple_call_clobber_set): Likewise.
4764 (gimple_bind_vars): Likewise.
4765 (gimple_bind_set_vars): Likewise.
4766 (gimple_bind_body_ptr): Likewise.
4767 (gimple_bind_set_body): Likewise.
4768 (gimple_bind_add_stmt): Likewise.
4769 (gimple_bind_block): Likewise.
4770 (gimple_bind_set_block): Likewise.
4771 (gimple_asm_ninputs): Likewise.
4772 (gimple_asm_noutputs): Likewise.
4773 (gimple_asm_nclobbers): Likewise.
4774 (gimple_asm_nlabels): Likewise.
4775 (gimple_asm_input_op): Likewise.
4776 (gimple_asm_input_op_ptr): Likewise.
4777 (gimple_asm_output_op): Likewise.
4778 (gimple_asm_output_op_ptr): Likewise.
4779 (gimple_asm_set_output_op): Likewise.
4780 (gimple_asm_clobber_op): Likewise.
4781 (gimple_asm_set_clobber_op): Likewise.
4782 (gimple_asm_label_op): Likewise.
4783 (gimple_asm_set_label_op): Likewise.
4784 (gimple_asm_string): Likewise.
4785 (gimple_catch_types): Likewise.
4786 (gimple_catch_types_ptr): Likewise.
4787 (gimple_catch_handler_ptr): Likewise.
4788 (gimple_catch_set_types): Likewise.
4789 (gimple_catch_set_handler): Likewise.
4790 (gimple_eh_filter_types): Likewise.
4791 (gimple_eh_filter_types_ptr): Likewise.
4792 (gimple_eh_filter_failure_ptr): Likewise.
4793 (gimple_eh_filter_set_types): Likewise.
4794 (gimple_eh_filter_set_failure): Likewise.
4795 (gimple_eh_must_not_throw_fndecl): Likewise.
4796 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4797 (gimple_eh_else_n_body_ptr): Likewise.
4798 (gimple_eh_else_e_body_ptr): Likewise.
4799 (gimple_eh_else_set_n_body): Likewise.
4800 (gimple_eh_else_set_e_body): Likewise.
4801 (gimple_try_eval_ptr): Likewise.
4802 (gimple_try_cleanup_ptr): Likewise.
4803 (gimple_try_set_eval): Likewise.
4804 (gimple_try_set_cleanup): Likewise.
4805 (gimple_wce_cleanup_ptr): Likewise.
4806 (gimple_wce_set_cleanup): Likewise.
4807 (gimple_phi_capacity): Likewise.
4808 (gimple_phi_num_args): Likewise.
4809 (gimple_phi_result): Likewise.
4810 (gimple_phi_result_ptr): Likewise.
4811 (gimple_phi_set_result): Likewise.
4812 (gimple_phi_arg): Likewise.
4813 (gimple_phi_set_arg): Likewise.
4814 (gimple_resx_region): Likewise.
4815 (gimple_resx_set_region): Likewise.
4816 (gimple_eh_dispatch_region): Likewise.
4817 (gimple_eh_dispatch_set_region): Likewise.
4818 (gimple_omp_critical_name): Likewise.
4819 (gimple_omp_critical_name_ptr): Likewise.
4820 (gimple_omp_critical_set_name): Likewise.
4821 (gimple_omp_for_clauses): Likewise.
4822 (gimple_omp_for_clauses_ptr): Likewise.
4823 (gimple_omp_for_set_clauses): Likewise.
4824 (gimple_omp_for_collapse): Likewise.
4825 (gimple_omp_for_index): Likewise.
4826 (gimple_omp_for_index_ptr): Likewise.
4827 (gimple_omp_for_set_index): Likewise.
4828 (gimple_omp_for_initial): Likewise.
4829 (gimple_omp_for_initial_ptr): Likewise.
4830 (gimple_omp_for_set_initial): Likewise.
4831 (gimple_omp_for_final): Likewise.
4832 (gimple_omp_for_final_ptr): Likewise.
4833 (gimple_omp_for_set_final): Likewise.
4834 (gimple_omp_for_incr): Likewise.
4835 (gimple_omp_for_incr_ptr): Likewise.
4836 (gimple_omp_for_set_incr): Likewise.
4837 (gimple_omp_for_pre_body_ptr): Likewise.
4838 (gimple_omp_for_set_pre_body): Likewise.
4839 (gimple_omp_parallel_clauses): Likewise.
4840 (gimple_omp_parallel_clauses_ptr): Likewise.
4841 (gimple_omp_parallel_set_clauses): Likewise.
4842 (gimple_omp_parallel_child_fn): Likewise.
4843 (gimple_omp_parallel_child_fn_ptr): Likewise.
4844 (gimple_omp_parallel_set_child_fn): Likewise.
4845 (gimple_omp_parallel_data_arg): Likewise.
4846 (gimple_omp_parallel_data_arg_ptr): Likewise.
4847 (gimple_omp_parallel_set_data_arg): Likewise.
4848 (gimple_omp_task_clauses): Likewise.
4849 (gimple_omp_task_clauses_ptr): Likewise.
4850 (gimple_omp_task_set_clauses): Likewise.
4851 (gimple_omp_task_child_fn): Likewise.
4852 (gimple_omp_task_child_fn_ptr): Likewise.
4853 (gimple_omp_task_set_child_fn): Likewise.
4854 (gimple_omp_task_data_arg): Likewise.
4855 (gimple_omp_task_data_arg_ptr): Likewise.
4856 (gimple_omp_task_set_data_arg): Likewise.
4857 (gimple_omp_taskreg_clauses): Likewise.
4858 (gimple_omp_taskreg_clauses_ptr): Likewise.
4859 (gimple_omp_taskreg_set_clauses): Likewise.
4860 (gimple_omp_taskreg_child_fn): Likewise.
4861 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4862 (gimple_omp_taskreg_set_child_fn): Likewise.
4863 (gimple_omp_taskreg_data_arg): Likewise.
4864 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4865 (gimple_omp_taskreg_set_data_arg): Likewise.
4866 (gimple_omp_task_copy_fn): Likewise.
4867 (gimple_omp_task_copy_fn_ptr): Likewise.
4868 (gimple_omp_task_set_copy_fn): Likewise.
4869 (gimple_omp_task_arg_size): Likewise.
4870 (gimple_omp_task_arg_size_ptr): Likewise.
4871 (gimple_omp_task_set_arg_size): Likewise.
4872 (gimple_omp_task_arg_align): Likewise.
4873 (gimple_omp_task_arg_align_ptr): Likewise.
4874 (gimple_omp_task_set_arg_align): Likewise.
4875 (gimple_omp_single_clauses): Likewise.
4876 (gimple_omp_single_clauses_ptr): Likewise.
4877 (gimple_omp_single_set_clauses): Likewise.
4878 (gimple_omp_target_clauses): Likewise.
4879 (gimple_omp_target_clauses_ptr): Likewise.
4880 (gimple_omp_target_set_clauses): Likewise.
4881 (gimple_omp_target_child_fn): Likewise.
4882 (gimple_omp_target_child_fn_ptr): Likewise.
4883 (gimple_omp_target_set_child_fn): Likewise.
4884 (gimple_omp_target_data_arg): Likewise.
4885 (gimple_omp_target_data_arg_ptr): Likewise.
4886 (gimple_omp_target_set_data_arg): Likewise.
4887 (gimple_omp_teams_clauses): Likewise.
4888 (gimple_omp_teams_clauses_ptr): Likewise.
4889 (gimple_omp_teams_set_clauses): Likewise.
4890 (gimple_omp_sections_clauses): Likewise.
4891 (gimple_omp_sections_clauses_ptr): Likewise.
4892 (gimple_omp_sections_set_clauses): Likewise.
4893 (gimple_omp_sections_control): Likewise.
4894 (gimple_omp_sections_control_ptr): Likewise.
4895 (gimple_omp_sections_set_control): Likewise.
4896 (gimple_omp_for_set_cond): Likewise.
4897 (gimple_omp_for_cond): Likewise.
4898 (gimple_omp_atomic_store_set_val): Likewise.
4899 (gimple_omp_atomic_store_val): Likewise.
4900 (gimple_omp_atomic_store_val_ptr): Likewise.
4901 (gimple_omp_atomic_load_set_lhs): Likewise.
4902 (gimple_omp_atomic_load_lhs): Likewise.
4903 (gimple_omp_atomic_load_lhs_ptr): Likewise.
4904 (gimple_omp_atomic_load_set_rhs): Likewise.
4905 (gimple_omp_atomic_load_rhs): Likewise.
4906 (gimple_omp_atomic_load_rhs_ptr): Likewise.
4907 (gimple_omp_continue_control_def): Likewise.
4908 (gimple_omp_continue_control_def_ptr): Likewise.
4909 (gimple_omp_continue_set_control_def): Likewise.
4910 (gimple_omp_continue_control_use): Likewise.
4911 (gimple_omp_continue_control_use_ptr): Likewise.
4912 (gimple_omp_continue_set_control_use): Likewise.
4913 (gimple_transaction_body_ptr): Likewise.
4914 (gimple_transaction_label): Likewise.
4915 (gimple_transaction_label_ptr): Likewise.
4916 (gimple_transaction_set_body): Likewise.
4917 (gimple_transaction_set_label): Likewise.
4919 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
4920 * ipa-inline-analysis.c (inline_write_summary): Likewise.
4921 * ipa-ref.c (ipa_record_reference): Likewise.
4922 * ipa-reference.c (analyze_function): Likewise.
4923 (ipa_reference_write_optimization_summary): Likewise.
4924 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
4925 (address_taken_from_non_vtable_p): Likewise.
4926 (comdat_can_be_unshared_p_1): Likewise.
4927 * lto-cgraph.c (lto_output_ref): Likewise.
4928 (add_references): Likewise.
4929 (compute_ltrans_boundary): Likewise.
4930 (output_symtab): Likewise.
4931 (input_ref): Likewise.
4932 (input_cgraph_1): Likewise.
4933 (output_cgraph_opt_summary): Likewise.
4934 * lto-streamer-out.c (lto_output): Likewise.
4935 (output_symbol_p): Likewise.
4936 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
4937 (lsei_start_function_in_partition): Likewise.
4938 (lsei_next_variable_in_partition): Likewise.
4939 (lsei_start_variable_in_partition): Likewise.
4940 * symtab.c (insert_to_assembler_name_hash): Likewise.
4941 (unlink_from_assembler_name_hash): Likewise.
4942 (symtab_unregister_node): Likewise.
4943 (symtab_remove_node): Likewise.
4944 (dump_symtab_node): Likewise.
4945 (verify_symtab_base): Likewise.
4946 (verify_symtab_node): Likewise.
4947 (symtab_make_decl_local): Likewise.
4948 (symtab_alias_ultimate_target): Likewise.
4949 (symtab_resolve_alias): Likewise.
4950 (symtab_get_symbol_partitioning_class): Likewise.
4951 * tree-phinodes.c (allocate_phi_node): Likewise.
4952 (reserve_phi_args_for_new_edge): Likewise.
4953 (remove_phi_args): Likewise.
4954 * varpool.c (varpool_node_for_asm): Likewise.
4955 (varpool_remove_unreferenced_decls): Likewise.
4957 2014-04-23 Jeff Law <law@redhat.com>
4959 PR tree-optimization/60902
4960 * tree-ssa-threadedge.c
4961 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
4962 invalidate outputs from statements that do not produce useful
4963 outputs for threading.
4965 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4967 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
4968 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
4969 machine descriptions for Stack Smashing Protector.
4971 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
4973 * aarch64.md (<optab>_rol<mode>3): New pattern.
4974 (<optab>_rolsi3_uxtw): Likewise.
4975 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
4977 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
4979 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
4980 (arm_cortex_a12_tune): Likewise.
4982 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4984 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
4986 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4988 * config/arm/arm.md (arm_rev16si2): New pattern.
4989 (arm_rev16si2_alt): Likewise.
4990 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
4992 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4994 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
4995 (rev16<mode>2_alt): Likewise.
4996 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
4997 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
4998 (aarch_rev16_shleft_mask_imm_p): Likewise.
4999 (aarch_rev16_p_1): Likewise.
5000 (aarch_rev16_p): Likewise.
5001 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5002 (aarch_rev16_shright_mask_imm_p): Likewise.
5003 (aarch_rev16_shleft_mask_imm_p): Likewise.
5005 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5007 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5008 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5010 (cortex_a53_extra_costs): Likewise.
5011 (cortex_a57_extra_costs): Likewise.
5012 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5013 (cortexa7_extra_costs): Likewise.
5014 (cortexa8_extra_costs): Likewise.
5015 (cortexa12_extra_costs): Likewise.
5016 (cortexa15_extra_costs): Likewise.
5017 (v7m_extra_costs): Likewise.
5018 (arm_new_rtx_costs): Handle BSWAP.
5020 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5022 * config/arm/arm.c (cortexa8_extra_costs): New table.
5023 (arm_cortex_a8_tune): New tuning struct.
5024 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5026 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5028 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5030 2014-04-23 Richard Biener <rguenther@suse.de>
5032 * Makefile.in (OBJS): Remove loop-unswitch.o.
5033 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5034 * passes.def (pass_rtl_unswitch): Likewise.
5035 * loop-init.c (gate_rtl_unswitch): Likewise.
5036 (rtl_unswitch): Likewise.
5037 (pass_data_rtl_unswitch): Likewise.
5038 (pass_rtl_unswitch): Likewise.
5039 (make_pass_rtl_unswitch): Likewise.
5040 * rtl.h (reversed_condition): Likewise.
5041 (compare_and_jump_seq): Likewise.
5042 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5044 * loop-unroll.c (compare_and_jump_seq): Likewise.
5046 2014-04-23 Richard Biener <rguenther@suse.de>
5048 PR tree-optimization/60903
5049 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5050 commented code block.
5051 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5052 loop flags to newly created BBs and edges.
5054 2014-04-23 Nick Clifton <nickc@redhat.com>
5056 * config/msp430/msp430.c (msp430_handle_option): Move function
5058 (msp430_option_override): Simplify mcu and mcpu option handling.
5059 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5060 support for -mhwmult command line option.
5061 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5062 -mhwmult command line option.
5063 (msp430_hwmult_enabled): Delete.
5064 (msp43o_output_labelref): Add support for -mhwmult command line option.
5065 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5066 (umulsidi3): Likewise.
5067 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5068 (mcpu, mlarge, msmall): Likewise.
5069 (mhwmult): New option.
5070 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5072 (msp430_is_f5_mcu): Remove prototype.
5073 (msp430_use_f5_series_hwmult): Add prototype.
5074 * config/msp430/msp430-opts.h: New file.
5075 * common/config/msp430: New directory.
5076 * common/config/msp430/msp430-common.c: New file.
5077 * config.gcc (msp430): Remove target_has_targetm_common.
5078 * doc/invoke.texi: Document -mhwmult command line option.
5080 2014-04-23 Nick Clifton <nickc@redhat.com>
5082 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5083 default-manifest.o if it can be found in the search path.
5084 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5086 2014-04-23 Terry Guo <terry.guo@arm.com>
5088 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5090 2014-04-23 Richard Biener <rguenther@suse.de>
5093 * tree-inline.c (declare_return_variable): Use mark_addressable.
5095 2014-04-23 Richard Biener <rguenther@suse.de>
5098 * loop-init.c (loop_optimizer_init): Make sure to apply
5099 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5101 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5104 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5106 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5107 if flag_sanitize_undefined_trap_on_error.
5108 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5109 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5110 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5111 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5112 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5113 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5114 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5115 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5116 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5117 * ubsan.c (ubsan_instrument_unreachable): Return
5118 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5119 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5120 if flag_sanitize_undefined_trap_on_error and
5121 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5122 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5123 instrument_bool_enum_load): Emit __builtin_trap () if
5124 flag_sanitize_undefined_trap_on_error and
5125 __builtin_handle_*_abort () if !flag_sanitize_recover.
5126 * doc/invoke.texi (-fsanitize-recover,
5127 -fsanitize-undefined-trap-on-error): Document.
5129 2014-04-22 Christian Bruel <christian.bruel@st.com>
5131 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5132 Force immediates to SImode.
5134 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5136 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5138 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5139 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5140 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5141 (nios2_fpu_insn): Add entry for round.
5142 (N2FPU_NO_ERRNO_P): Define.
5143 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5145 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5147 2014-04-22 Richard Henderson <rth@redhat.com>
5149 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5150 (add<GPI>3_compare0): Remove leading * from name.
5151 (add<GPI>3_carryin): Likewise.
5152 (sub<GPI>3_compare0): Likewise.
5153 (sub<GPI>3_carryin): Likewise.
5154 (<su_optab>mulditi3): New expander.
5155 (multi3): New expander.
5156 (madd<GPI>): Remove leading * from name.
5158 2014-04-22 Martin Jambor <mjambor@suse.cz>
5160 * cgraphclones.c (cgraph_function_versioning): Copy
5161 ipa_transforms_to_apply instead of asserting it is empty.
5163 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5166 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5167 on count_exp to get mode.
5169 2014-04-22 Andrew Pinski <apinski@cavium.com>
5171 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5172 Handle TLS for ILP32.
5173 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5174 (tlsie_small_<mode>): this and handle PTR.
5175 (tlsie_small_sidi): New pattern.
5176 (tlsle_small): Change to an expand to handle ILP32.
5177 (tlsle_small_<mode>): New pattern.
5178 (tlsdesc_small): Rename to ...
5179 (tlsdesc_small_<mode>): this and handle PTR.
5181 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5183 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5185 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5187 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5188 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5189 (aarch64_types_signed_poly_qualifiers): Likewise.
5190 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5191 (aarch64_types_poly_signed_qualifiers): Likewise.
5192 (TYPES_REINTERP_SS): Type macro added.
5193 (TYPES_REINTERP_SU): Likewise.
5194 (TYPES_REINTERP_SP): Likewise.
5195 (TYPES_REINTERP_US): Likewise.
5196 (TYPES_REINTERP_PS): Likewise.
5197 (aarch64_fold_builtin): New expression folding added.
5198 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5199 Declarations removed.
5200 (REINTERP_SS): Declarations added.
5201 (REINTERP_US): Likewise.
5202 (REINTERP_PS): Likewise.
5203 (REINTERP_SU): Likewise.
5204 (REINTERP_SP): Likewise.
5205 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5206 (vreinterpretq_p8_f64): Likewise.
5207 (vreinterpret_p16_f64): Likewise.
5208 (vreinterpretq_p16_f64): Likewise.
5209 (vreinterpret_f32_f64): Likewise.
5210 (vreinterpretq_f32_f64): Likewise.
5211 (vreinterpret_f64_f32): Likewise.
5212 (vreinterpret_f64_p8): Likewise.
5213 (vreinterpret_f64_p16): Likewise.
5214 (vreinterpret_f64_s8): Likewise.
5215 (vreinterpret_f64_s16): Likewise.
5216 (vreinterpret_f64_s32): Likewise.
5217 (vreinterpret_f64_s64): Likewise.
5218 (vreinterpret_f64_u8): Likewise.
5219 (vreinterpret_f64_u16): Likewise.
5220 (vreinterpret_f64_u32): Likewise.
5221 (vreinterpret_f64_u64): Likewise.
5222 (vreinterpretq_f64_f32): Likewise.
5223 (vreinterpretq_f64_p8): Likewise.
5224 (vreinterpretq_f64_p16): Likewise.
5225 (vreinterpretq_f64_s8): Likewise.
5226 (vreinterpretq_f64_s16): Likewise.
5227 (vreinterpretq_f64_s32): Likewise.
5228 (vreinterpretq_f64_s64): Likewise.
5229 (vreinterpretq_f64_u8): Likewise.
5230 (vreinterpretq_f64_u16): Likewise.
5231 (vreinterpretq_f64_u32): Likewise.
5232 (vreinterpretq_f64_u64): Likewise.
5233 (vreinterpret_s64_f64): Likewise.
5234 (vreinterpretq_s64_f64): Likewise.
5235 (vreinterpret_u64_f64): Likewise.
5236 (vreinterpretq_u64_f64): Likewise.
5237 (vreinterpret_s8_f64): Likewise.
5238 (vreinterpretq_s8_f64): Likewise.
5239 (vreinterpret_s16_f64): Likewise.
5240 (vreinterpretq_s16_f64): Likewise.
5241 (vreinterpret_s32_f64): Likewise.
5242 (vreinterpretq_s32_f64): Likewise.
5243 (vreinterpret_u8_f64): Likewise.
5244 (vreinterpretq_u8_f64): Likewise.
5245 (vreinterpret_u16_f64): Likewise.
5246 (vreinterpretq_u16_f64): Likewise.
5247 (vreinterpret_u32_f64): Likewise.
5248 (vreinterpretq_u32_f64): Likewise.
5250 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5252 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5253 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5254 (vreinterpret_p8_s8): Likewise.
5255 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5256 (vreinterpret_p8_s16): Likewise.
5257 (vreinterpret_p8_s32): Likewise.
5258 (vreinterpret_p8_s64): Likewise.
5259 (vreinterpret_p8_f32): Likewise.
5260 (vreinterpret_p8_u8): Likewise.
5261 (vreinterpret_p8_u16): Likewise.
5262 (vreinterpret_p8_u32): Likewise.
5263 (vreinterpret_p8_u64): Likewise.
5264 (vreinterpret_p8_p16): Likewise.
5265 (vreinterpretq_p8_s8): Likewise.
5266 (vreinterpretq_p8_s16): Likewise.
5267 (vreinterpretq_p8_s32): Likewise.
5268 (vreinterpretq_p8_s64): Likewise.
5269 (vreinterpretq_p8_f32): Likewise.
5270 (vreinterpretq_p8_u8): Likewise.
5271 (vreinterpretq_p8_u16): Likewise.
5272 (vreinterpretq_p8_u32): Likewise.
5273 (vreinterpretq_p8_u64): Likewise.
5274 (vreinterpretq_p8_p16): Likewise.
5275 (vreinterpret_p16_s8): Likewise.
5276 (vreinterpret_p16_s16): Likewise.
5277 (vreinterpret_p16_s32): Likewise.
5278 (vreinterpret_p16_s64): Likewise.
5279 (vreinterpret_p16_f32): Likewise.
5280 (vreinterpret_p16_u8): Likewise.
5281 (vreinterpret_p16_u16): Likewise.
5282 (vreinterpret_p16_u32): Likewise.
5283 (vreinterpret_p16_u64): Likewise.
5284 (vreinterpret_p16_p8): Likewise.
5285 (vreinterpretq_p16_s8): Likewise.
5286 (vreinterpretq_p16_s16): Likewise.
5287 (vreinterpretq_p16_s32): Likewise.
5288 (vreinterpretq_p16_s64): Likewise.
5289 (vreinterpretq_p16_f32): Likewise.
5290 (vreinterpretq_p16_u8): Likewise.
5291 (vreinterpretq_p16_u16): Likewise.
5292 (vreinterpretq_p16_u32): Likewise.
5293 (vreinterpretq_p16_u64): Likewise.
5294 (vreinterpretq_p16_p8): Likewise.
5295 (vreinterpret_f32_s8): Likewise.
5296 (vreinterpret_f32_s16): Likewise.
5297 (vreinterpret_f32_s32): Likewise.
5298 (vreinterpret_f32_s64): Likewise.
5299 (vreinterpret_f32_u8): Likewise.
5300 (vreinterpret_f32_u16): Likewise.
5301 (vreinterpret_f32_u32): Likewise.
5302 (vreinterpret_f32_u64): Likewise.
5303 (vreinterpret_f32_p8): Likewise.
5304 (vreinterpret_f32_p16): Likewise.
5305 (vreinterpretq_f32_s8): Likewise.
5306 (vreinterpretq_f32_s16): Likewise.
5307 (vreinterpretq_f32_s32): Likewise.
5308 (vreinterpretq_f32_s64): Likewise.
5309 (vreinterpretq_f32_u8): Likewise.
5310 (vreinterpretq_f32_u16): Likewise.
5311 (vreinterpretq_f32_u32): Likewise.
5312 (vreinterpretq_f32_u64): Likewise.
5313 (vreinterpretq_f32_p8): Likewise.
5314 (vreinterpretq_f32_p16): Likewise.
5315 (vreinterpret_s64_s8): Likewise.
5316 (vreinterpret_s64_s16): Likewise.
5317 (vreinterpret_s64_s32): Likewise.
5318 (vreinterpret_s64_f32): Likewise.
5319 (vreinterpret_s64_u8): Likewise.
5320 (vreinterpret_s64_u16): Likewise.
5321 (vreinterpret_s64_u32): Likewise.
5322 (vreinterpret_s64_u64): Likewise.
5323 (vreinterpret_s64_p8): Likewise.
5324 (vreinterpret_s64_p16): Likewise.
5325 (vreinterpretq_s64_s8): Likewise.
5326 (vreinterpretq_s64_s16): Likewise.
5327 (vreinterpretq_s64_s32): Likewise.
5328 (vreinterpretq_s64_f32): Likewise.
5329 (vreinterpretq_s64_u8): Likewise.
5330 (vreinterpretq_s64_u16): Likewise.
5331 (vreinterpretq_s64_u32): Likewise.
5332 (vreinterpretq_s64_u64): Likewise.
5333 (vreinterpretq_s64_p8): Likewise.
5334 (vreinterpretq_s64_p16): Likewise.
5335 (vreinterpret_u64_s8): Likewise.
5336 (vreinterpret_u64_s16): Likewise.
5337 (vreinterpret_u64_s32): Likewise.
5338 (vreinterpret_u64_s64): Likewise.
5339 (vreinterpret_u64_f32): Likewise.
5340 (vreinterpret_u64_u8): Likewise.
5341 (vreinterpret_u64_u16): Likewise.
5342 (vreinterpret_u64_u32): Likewise.
5343 (vreinterpret_u64_p8): Likewise.
5344 (vreinterpret_u64_p16): Likewise.
5345 (vreinterpretq_u64_s8): Likewise.
5346 (vreinterpretq_u64_s16): Likewise.
5347 (vreinterpretq_u64_s32): Likewise.
5348 (vreinterpretq_u64_s64): Likewise.
5349 (vreinterpretq_u64_f32): Likewise.
5350 (vreinterpretq_u64_u8): Likewise.
5351 (vreinterpretq_u64_u16): Likewise.
5352 (vreinterpretq_u64_u32): Likewise.
5353 (vreinterpretq_u64_p8): Likewise.
5354 (vreinterpretq_u64_p16): Likewise.
5355 (vreinterpret_s8_s16): Likewise.
5356 (vreinterpret_s8_s32): Likewise.
5357 (vreinterpret_s8_s64): Likewise.
5358 (vreinterpret_s8_f32): Likewise.
5359 (vreinterpret_s8_u8): Likewise.
5360 (vreinterpret_s8_u16): Likewise.
5361 (vreinterpret_s8_u32): Likewise.
5362 (vreinterpret_s8_u64): Likewise.
5363 (vreinterpret_s8_p8): Likewise.
5364 (vreinterpret_s8_p16): Likewise.
5365 (vreinterpretq_s8_s16): Likewise.
5366 (vreinterpretq_s8_s32): Likewise.
5367 (vreinterpretq_s8_s64): Likewise.
5368 (vreinterpretq_s8_f32): Likewise.
5369 (vreinterpretq_s8_u8): Likewise.
5370 (vreinterpretq_s8_u16): Likewise.
5371 (vreinterpretq_s8_u32): Likewise.
5372 (vreinterpretq_s8_u64): Likewise.
5373 (vreinterpretq_s8_p8): Likewise.
5374 (vreinterpretq_s8_p16): Likewise.
5375 (vreinterpret_s16_s8): Likewise.
5376 (vreinterpret_s16_s32): Likewise.
5377 (vreinterpret_s16_s64): Likewise.
5378 (vreinterpret_s16_f32): Likewise.
5379 (vreinterpret_s16_u8): Likewise.
5380 (vreinterpret_s16_u16): Likewise.
5381 (vreinterpret_s16_u32): Likewise.
5382 (vreinterpret_s16_u64): Likewise.
5383 (vreinterpret_s16_p8): Likewise.
5384 (vreinterpret_s16_p16): Likewise.
5385 (vreinterpretq_s16_s8): Likewise.
5386 (vreinterpretq_s16_s32): Likewise.
5387 (vreinterpretq_s16_s64): Likewise.
5388 (vreinterpretq_s16_f32): Likewise.
5389 (vreinterpretq_s16_u8): Likewise.
5390 (vreinterpretq_s16_u16): Likewise.
5391 (vreinterpretq_s16_u32): Likewise.
5392 (vreinterpretq_s16_u64): Likewise.
5393 (vreinterpretq_s16_p8): Likewise.
5394 (vreinterpretq_s16_p16): Likewise.
5395 (vreinterpret_s32_s8): Likewise.
5396 (vreinterpret_s32_s16): Likewise.
5397 (vreinterpret_s32_s64): Likewise.
5398 (vreinterpret_s32_f32): Likewise.
5399 (vreinterpret_s32_u8): Likewise.
5400 (vreinterpret_s32_u16): Likewise.
5401 (vreinterpret_s32_u32): Likewise.
5402 (vreinterpret_s32_u64): Likewise.
5403 (vreinterpret_s32_p8): Likewise.
5404 (vreinterpret_s32_p16): Likewise.
5405 (vreinterpretq_s32_s8): Likewise.
5406 (vreinterpretq_s32_s16): Likewise.
5407 (vreinterpretq_s32_s64): Likewise.
5408 (vreinterpretq_s32_f32): Likewise.
5409 (vreinterpretq_s32_u8): Likewise.
5410 (vreinterpretq_s32_u16): Likewise.
5411 (vreinterpretq_s32_u32): Likewise.
5412 (vreinterpretq_s32_u64): Likewise.
5413 (vreinterpretq_s32_p8): Likewise.
5414 (vreinterpretq_s32_p16): Likewise.
5415 (vreinterpret_u8_s8): Likewise.
5416 (vreinterpret_u8_s16): Likewise.
5417 (vreinterpret_u8_s32): Likewise.
5418 (vreinterpret_u8_s64): Likewise.
5419 (vreinterpret_u8_f32): Likewise.
5420 (vreinterpret_u8_u16): Likewise.
5421 (vreinterpret_u8_u32): Likewise.
5422 (vreinterpret_u8_u64): Likewise.
5423 (vreinterpret_u8_p8): Likewise.
5424 (vreinterpret_u8_p16): Likewise.
5425 (vreinterpretq_u8_s8): Likewise.
5426 (vreinterpretq_u8_s16): Likewise.
5427 (vreinterpretq_u8_s32): Likewise.
5428 (vreinterpretq_u8_s64): Likewise.
5429 (vreinterpretq_u8_f32): Likewise.
5430 (vreinterpretq_u8_u16): Likewise.
5431 (vreinterpretq_u8_u32): Likewise.
5432 (vreinterpretq_u8_u64): Likewise.
5433 (vreinterpretq_u8_p8): Likewise.
5434 (vreinterpretq_u8_p16): Likewise.
5435 (vreinterpret_u16_s8): Likewise.
5436 (vreinterpret_u16_s16): Likewise.
5437 (vreinterpret_u16_s32): Likewise.
5438 (vreinterpret_u16_s64): Likewise.
5439 (vreinterpret_u16_f32): Likewise.
5440 (vreinterpret_u16_u8): Likewise.
5441 (vreinterpret_u16_u32): Likewise.
5442 (vreinterpret_u16_u64): Likewise.
5443 (vreinterpret_u16_p8): Likewise.
5444 (vreinterpret_u16_p16): Likewise.
5445 (vreinterpretq_u16_s8): Likewise.
5446 (vreinterpretq_u16_s16): Likewise.
5447 (vreinterpretq_u16_s32): Likewise.
5448 (vreinterpretq_u16_s64): Likewise.
5449 (vreinterpretq_u16_f32): Likewise.
5450 (vreinterpretq_u16_u8): Likewise.
5451 (vreinterpretq_u16_u32): Likewise.
5452 (vreinterpretq_u16_u64): Likewise.
5453 (vreinterpretq_u16_p8): Likewise.
5454 (vreinterpretq_u16_p16): Likewise.
5455 (vreinterpret_u32_s8): Likewise.
5456 (vreinterpret_u32_s16): Likewise.
5457 (vreinterpret_u32_s32): Likewise.
5458 (vreinterpret_u32_s64): Likewise.
5459 (vreinterpret_u32_f32): Likewise.
5460 (vreinterpret_u32_u8): Likewise.
5461 (vreinterpret_u32_u16): Likewise.
5462 (vreinterpret_u32_u64): Likewise.
5463 (vreinterpret_u32_p8): Likewise.
5464 (vreinterpret_u32_p16): Likewise.
5465 (vreinterpretq_u32_s8): Likewise.
5466 (vreinterpretq_u32_s16): Likewise.
5467 (vreinterpretq_u32_s32): Likewise.
5468 (vreinterpretq_u32_s64): Likewise.
5469 (vreinterpretq_u32_f32): Likewise.
5470 (vreinterpretq_u32_u8): Likewise.
5471 (vreinterpretq_u32_u16): Likewise.
5472 (vreinterpretq_u32_u64): Likewise.
5473 (vreinterpretq_u32_p8): Likewise.
5474 (vreinterpretq_u32_p16): Likewise.
5476 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5478 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5480 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5482 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5483 (vqnegd_s64): Likewise.
5484 (vqabs_s64): Likewise.
5485 (vqabsd_s64): Likewise.
5487 2014-04-22 Richard Henderson <rth@redhat.com>
5489 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5490 computation to the top of the loop.
5492 2014-04-22 Renlin <renlin.li@arm.com>
5493 Jiong Wang <jiong.wang@arm.com>
5495 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5496 * config/aarch64/aarch64.c (aarch64_layout_frame)
5497 (aarch64_initial_elimination_offset): Likewise.
5499 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5501 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5504 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5506 * machmode.h (bitwise_mode_for_mode): Declare.
5507 * stor-layout.h (bitwise_type_for_mode): Likewise.
5508 * stor-layout.c (bitwise_mode_for_mode): New function.
5509 (bitwise_type_for_mode): Likewise.
5510 * builtins.c (fold_builtin_memory_op): Use it instead of
5511 int_mode_for_mode and build_nonstandard_integer_type.
5513 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5515 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5516 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5517 (*-*-solaris2*): Simplify.
5518 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5519 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5520 *-*-solaris2.9* handling.
5522 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5524 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5525 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5527 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5528 * configure: Regenerate.
5530 * config/i386/sol2-9.h: Remove.
5532 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5533 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5534 Remove Solaris 9 references.
5536 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5538 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5539 (floatuns<GPI:mode><GPF:mode>2): Remove.
5540 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5541 and floatuns conversions.
5542 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5543 and floatuns conversions.
5544 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5545 (w1,w2): New mode attributes for inequal width conversions.
5547 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5549 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5550 the output asm format.
5552 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5554 * config/aarch64/aarch64-simd.md
5555 (aarch64_cm<optab>di): Always split.
5556 (*aarch64_cm<optab>di): New.
5557 (aarch64_cmtstdi): Always split.
5558 (*aarch64_cmtstdi): New.
5560 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5562 PR tree-optimization/60823
5563 * omp-low.c (ipa_simd_modify_function_body): Go through
5564 all SSA_NAMEs and for those refering to vector arguments
5565 which are going to be replaced adjust SSA_NAME_VAR and,
5566 if it is a default definition, change it into a non-default
5567 definition assigned at the beginning of function from new_decl.
5568 (ipa_simd_modify_stmt_ops): Rewritten.
5569 * tree-dfa.c (set_ssa_default_def): When removing default def,
5570 check for NULL loc instead of NULL *loc.
5572 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5574 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5575 restrictions on core registers for DImode values in Thumb2.
5577 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5579 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5580 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5582 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5584 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5585 (*iordi_notzesidi_di): Likewise.
5586 (*iordi_notsesidi_di): Likewise.
5588 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5590 * config/arm/arm-protos.h (tune_params): New struct members.
5591 * config/arm/arm.c: Initialise tune_params per processor.
5592 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5593 for speed, based on new tune_params.
5595 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5597 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5598 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5599 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5600 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5601 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5602 (vrnda_f64): Likewise.
5603 (vrndi_f64): Likewise.
5604 (vrndm_f64): Likewise.
5605 (vrndn_f64): Likewise.
5606 (vrndp_f64): Likewise.
5607 (vrndx_f64): Likewise.
5609 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5611 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5612 GET_MODE_SIZE argument is enum machine_mode.
5614 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5617 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5618 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5620 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5623 * asan.c (asan_emit_stack_protection): Force the base to align to
5624 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5625 appropriate bits if STRICT_ALIGNMENT.
5626 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5628 (expand_used_vars): Leave a space in the stack frame for alignment
5629 if STRICT_ALIGNMENT.
5631 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5633 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5635 (gimple_store_p): Likewise.
5636 (gimple_assign_load_p): Likewise.
5637 (gimple_assign_cast_p): Likewise.
5638 (gimple_clobber_p): Likewise.
5640 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5641 rather than a gimple.
5642 (gimple_assign_cast_p): Likewise.
5644 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5647 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5648 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5650 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5651 more debug information for E500 if -mdebug=reg.
5653 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5656 * config/i386/i386.c (ix86_expand_builtin)
5657 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5658 register for target RTX.
5659 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5661 2014-04-18 Cong Hou <congh@google.com>
5663 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5664 the widen-mult pattern by handling two operands with different sizes,
5665 and operands whose size is smaller than half of the result type.
5667 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5669 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5670 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5671 (do_estimate_edge_time): Compute it.
5672 * ipa-inline.c (want_inline_small_function_p): Bypass
5673 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5675 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5677 * ipa-inline.c (spec_rem): New static variable.
5678 (dump_overall_stats): New function.
5679 (dump_inline_stats): New function.
5681 2014-04-18 Richard Henderson <rth@redhat.com>
5683 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5684 to GET_MODE_SIZE, not a reg_class_t.
5686 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5688 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5689 (vsx_xxmrglw_<mode>): Likewise.
5691 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5694 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5695 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5696 (rs6000_init_hard_regno_mode_ok): Likewise.
5698 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5700 * ipa-inline.c (inline_small_functions): Account only non-cold
5702 * doc/invoke.texi (inline-unit-growth): Update documentation.
5704 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5706 * config/rs6000/rs6000.md (addti3, subti3): New.
5708 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5711 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5712 comment. Check optimize_insn_for_size_p instead of
5713 optimize_insn_for_speed_p.
5715 2014-04-17 Martin Jambor <mjambor@suse.cz>
5717 * gimple-iterator.c (gsi_start_edge): New function.
5718 * gimple-iterator.h (gsi_start_edge): Declare.
5719 * tree-sra.c (single_non_eh_succ): New function.
5720 (disqualify_ops_if_throwing_stmt): Renamed to
5721 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5722 having one non-EH successor BB.
5723 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5724 generate loads into replacements.
5725 (sra_modify_assign): Likewise and and also use the simple path for
5727 (sra_modify_function_body): Commit statements on edges.
5729 2014-04-17 Richard Biener <rguenther@suse.de>
5732 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5733 comparison results and add clarifying comment.
5735 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5737 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5738 (blank_mode): Initialize it.
5739 (emit_mode_size_inline, emit_mode_nunits_inline,
5740 emit_mode_inner_inline): New functions.
5741 (emit_insn_modes_h): Call them and surround their output with
5742 #if GCC_VERSION >= 4001 ... #endif.
5743 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5744 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5745 mode_* arrays if the argument is __builtin_constant_p.
5746 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5747 is enum machine_mode.
5749 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5751 * passes.c (opt_pass::execute): Adjust.
5752 (pass_manager::execute_pass_mode_switching): Likewise.
5753 (early_local_passes::execute): Likewise.
5754 (execute_one_pass): Pass cfun to the pass's execute method.
5755 * tree-pass.h (opt_pass::execute): Add function * argument.
5756 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5757 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5758 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5759 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5760 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5761 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5762 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5763 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5764 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5765 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5766 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5767 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5768 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5769 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5770 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5771 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5772 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5773 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5774 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5775 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5776 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5777 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5778 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5779 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5780 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5781 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5782 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5783 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5784 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5787 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5789 * passes.c (opt_pass::gate): Take function * argument.
5790 (gate_all_early_local_passes): Merge into
5791 (early_local_passes::gate): this.
5792 (gate_all_early_optimizations): Merge into
5793 (all_early_optimizations::gate): this.
5794 (gate_all_optimizations): Mege into
5795 (all_optimizations::gate): this.
5796 (gate_all_optimizations_g): Merge into
5797 (all_optimizations_g::gate): this.
5798 (gate_rest_of_compilation): Mege into
5799 (rest_of_compilation::gate): this.
5800 (gate_postreload): Merge into
5801 (postreload::gate): this.
5802 (dump_one_pass): Pass cfun to the pass's gate method.
5803 (execute_ipa_summary_passes): Likewise.
5804 (execute_one_pass): Likewise.
5805 (ipa_write_summaries_2): Likewise.
5806 (ipa_write_optimization_summaries_1): Likewise.
5807 (ipa_read_summaries_1): Likewise.
5808 (ipa_read_optimization_summaries_1): Likewise.
5809 (execute_ipa_stmt_fixups): Likewise.
5810 * tree-pass.h (opt_pass::gate): Add function * argument.
5811 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5812 combine-stack-adj.c, combine.c, compare-elim.c,
5813 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5814 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5815 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5816 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5817 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5818 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5819 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5820 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5821 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5822 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5823 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5824 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5825 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5826 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5827 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5828 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5829 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5830 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5831 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5832 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5833 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5834 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5835 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5836 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5837 var-tracking.c, vtable-verify.c, web.c: Adjust.
5839 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5841 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5842 * configure: Regenerate.
5844 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5846 * passes.c (dump_one_pass): don't check pass->has_gate.
5847 (execute_ipa_summary_passes): Likewise.
5848 (execute_one_pass): Likewise.
5849 (ipa_write_summaries_2): Likewise.
5850 (ipa_write_optimization_summaries_1): Likewise.
5851 (ipa_read_optimization_summaries_1): Likewise.
5852 (execute_ipa_stmt_fixups): Likewise.
5853 * tree-pass.h (pass_data::has_gate): Remove.
5854 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5855 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5856 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5857 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5858 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5859 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5860 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5861 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5862 gimple-low.c, gimple-ssa-isolate-paths.c,
5863 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5864 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5865 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5866 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5867 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5868 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5869 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5870 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5871 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5872 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5873 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5874 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5875 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5876 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5877 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5878 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5879 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5880 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5881 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5882 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5883 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5884 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5887 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5889 * pass_manager.h (pass_manager::register_dump_files_1): Remove
5891 * passes.c (pass_manager::register_dump_files_1): Merge into
5892 (pass_manager::register_dump_files): this, and remove its handling of
5893 properties since the pass always has the properties anyway.
5894 (pass_manager::pass_manager): Adjust.
5896 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5898 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
5899 * passes.c (pass_manager::register_dump_files_1): Remove dead code
5900 dealing with properties.
5901 (pass_manager::register_dump_files): Adjust.
5903 2014-03-20 Mark Wielaard <mjw@redhat.com>
5905 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
5906 then represent the bound as normal constant value.
5908 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5911 Forward port from 4.8 branch
5912 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
5914 * config/i386/bmiintrin.h (_blsi_u32): New.
5918 (_blsmsk_u32): Ditto.
5919 (_blsmsk_u64): Ditto.
5920 (_tzcnt_u32): Ditto.
5921 (_tzcnt_u64): Ditto.
5923 2014-04-17 Kito Cheng <kito@0xlab.org>
5925 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
5927 2014-04-17 Richard Biener <rguenther@suse.de>
5930 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
5931 boolean results for comparisons.
5933 2014-04-17 Richard Biener <rguenther@suse.de>
5935 PR tree-optimization/60836
5936 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
5937 initial PHI args to be gimple values.
5939 2014-04-17 Richard Biener <rguenther@suse.de>
5941 PR tree-optimization/60841
5942 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
5943 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
5944 of stmts to SLP build.
5945 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
5946 (vect_analyze_slp): Likewise.
5947 (vect_analyze_slp_instance): Likewise.
5948 (vect_build_slp_tree): Limit overall SLP tree growth.
5949 * tree-vectorizer.h (vect_analyze_data_refs,
5950 vect_analyze_slp): Adjust prototypes.
5952 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5954 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
5957 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5959 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
5960 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
5961 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
5962 for TARGET_SLOW_PSHUFB
5964 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
5966 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
5967 * config/i386/i386.c (intel_cost): Ditto.
5969 2014-04-17 Joey Ye <joey.ye@arm.com>
5971 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
5973 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5975 * opts.c (common_handle_option): Disable -fipa-reference coorectly
5978 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
5980 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
5981 (type_all_derivations_known_p): New predicate.
5982 (type_all_ctors_visible_p): New predicate.
5983 (type_possibly_instantiated_p): New predicate.
5984 (get_odr_type): Compute all_derivations_known.
5985 (dump_odr_type): Dump the flag.
5986 (maybe_record_type): Cleanup.
5987 (record_target_from_binfo): Add bases_to_consider array;
5988 record bases for types w/o instances and skip CXX destructor.
5989 (possible_polymorphic_call_targets_1): Add bases_to_consider
5990 and consider_construction parameters; check if type may have instance.
5991 (get_polymorphic_call_info): Set maybe_in_construction to true
5992 when we know nothing.
5993 (record_targets_from_bases): Skip CXX destructors; they are
5994 never called for types in construction.
5995 (possible_polymorphic_call_targets): Do not record target when
5996 type may not have instance.
5998 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6001 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6002 external aliases alive, too.
6004 2014-04-16 Andrew Pinski <apinski@cavium.com>
6006 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6009 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6011 * final.c (compute_alignments): Do not apply loop alignment to a block
6012 falling through to the exit.
6014 2014-04-16 Catherine Moore <clm@codesourcery.com>
6016 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6017 Adjust constraints for microMIPS store patterns.
6019 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6021 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6023 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6025 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6026 (append_use): Run at -O0.
6027 (append_vdef): Likewise.
6028 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6029 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6031 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6033 PR tree-optimization/60844
6034 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6035 (propagate_op_to_single_use, remove_visited_stmt_chain,
6036 linearize_expr, repropagate_negates, reassociate_bb): Use it
6037 instead of gsi_remove.
6039 2014-04-16 Martin Jambor <mjambor@suse.cz>
6041 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6042 ipa_transforms_to_apply.
6043 (cgraph_function_versioning): Assert that old_node has empty
6044 ipa_transforms_to_apply.
6045 * trans-mem.c (ipa_tm_create_version): Likewise.
6046 * tree-inline.c (tree_function_versioning): Do not duplicate
6047 ipa_transforms_to_apply.
6049 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6052 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6054 Pass necessary as flags on 64-bit Solaris/x86.
6055 Use lowercase relocs for x86_64-*-*.
6056 * configure: Regenerate.
6058 2014-04-15 Jan Hubicka <jh@suse.cz>
6060 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6061 (maybe_record_node, likely_target_p): Use it.
6063 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6066 Revert following patch
6068 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6071 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6072 software floating point or no floating point registers, do not
6073 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6074 in GPRs that occurs after we tested for GPRs that would never be
6077 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6078 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6079 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6080 specifically allow DDmode, since that does not use the SPE SIMD
6083 2014-03-21 Mark Wielaard <mjw@redhat.com>
6085 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6086 as unsigned or int depending on type and value used.
6088 2014-04-15 Richard Biener <rguenther@suse.de>
6090 PR rtl-optimization/56965
6091 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6092 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6094 * alias.c (true_dependence_1): Do not call
6095 nonoverlapping_component_refs_p.
6096 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6097 nonoverlapping_component_refs_p.
6098 (indirect_refs_may_alias_p): Likewise.
6100 2014-04-15 Teresa Johnson <tejohnson@google.com>
6102 * cfg.c (dump_bb_info): Fix flags check.
6103 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6105 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6107 PR rtl-optimization/60663
6108 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6111 2014-04-15 Richard Biener <rguenther@suse.de>
6113 * lto-streamer.h (LTO_major_version): Bump to 4.
6115 2014-04-15 Richard Biener <rguenther@suse.de>
6117 * common.opt (lto_partition_model): New enum.
6118 (flto-partition=): Merge separate options with a single with argument,
6119 add -flto-partition=one support.
6120 * flag-types.h (enum lto_partition_model): Declare.
6121 * opts.c (finish_options): Remove duplicate -flto-partition=
6123 * lto-wrapper.c (run_gcc): Adjust.
6125 2014-04-15 Richard Biener <rguenther@suse.de>
6127 * alias.c (ncr_compar): New function.
6128 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6130 2014-04-15 Richard Biener <rguenther@suse.de>
6132 * alias.c (record_component_aliases): Do not walk BINFOs.
6134 2014-04-15 Richard Biener <rguenther@suse.de>
6136 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6137 Add struct function argument and adjust.
6138 (find_func_aliases_for_call): Likewise.
6139 (find_func_aliases): Likewise.
6140 (find_func_clobbers): Likewise.
6141 (intra_create_variable_infos): Likewise.
6142 (compute_points_to_sets): Likewise.
6143 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6145 2014-04-15 Richard Biener <rguenther@suse.de>
6147 * tree.c (iterative_hash_expr): Use enum tree_code_class
6148 to store TREE_CODE_CLASS.
6149 (tree_block): Likewise.
6150 (tree_set_block): Likewise.
6151 * tree.h (fold_build_pointer_plus_loc): Use
6152 convert_to_ptrofftype_loc.
6154 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6157 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6160 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6162 * cfgloop.h (struct loop): Move force_vectorize down.
6163 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6164 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6165 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6166 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6167 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6168 * tree-core.h (enum annot_expr_kind): Add new kind values.
6169 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6170 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6172 * tree.def (ANNOTATE_EXPR): Tweak comment.
6174 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6176 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6179 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6181 * tree.h (TYPE_IDENTIFIER): Declare.
6182 * tree.c (subrange_type_for_debug_p): Use it.
6183 * godump.c (go_format_type): Likewise.
6184 * dwarf2out.c (is_cxx_auto, modified_type_die,
6185 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6186 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6188 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6191 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6193 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6195 * config/i386/i386.c (examine_argument): Return bool. Return true if
6196 parameter should be passed in memory.
6197 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6198 (construct_container): Update calls to examine_argument.
6199 (function_arg_advance_64): Ditto.
6200 (return_in_memory_32): Merge with ix86_return_in_memory.
6201 (return_in_memory_64): Ditto.
6202 (return_in_memory_ms_64): Ditto.
6204 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6206 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6207 * coverage.c (coverage_compute_profile_id): Handle externally visible
6210 2014-04-14 Martin Jambor <mjambor@suse.cz>
6212 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6213 DECL_DISREGARD_INLINE_LIMITS functions.
6215 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6218 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6220 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6223 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6224 optimize_insn_for_speed_p instead of
6225 optimize_function_for_speed_p.
6227 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6229 * doc/invoke.texi (free): Document AArch64.
6231 2014-04-14 Richard Biener <rguenther@suse.de>
6233 PR tree-optimization/60042
6234 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6235 (insert_into_preds_of_block): Do not prevent PHI insertion
6236 for REFERENCE exprs here ...
6237 (eliminate_dom_walker::before_dom_children): ... but prevent
6238 their use here under similar conditions when applied to the
6239 IL after PRE optimizations.
6241 2014-04-14 Richard Biener <rguenther@suse.de>
6243 * passes.def: Move early points-to after early SRA.
6245 2014-04-14 Richard Biener <rguenther@suse.de>
6247 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6248 check for which sign-changes we allow when forwarding
6249 a converted value into a switch.
6251 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6253 * stor-layout.c (place_field): Finalize non-constant offset for the
6256 2014-04-14 Richard Biener <rguenther@suse.de>
6258 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6260 (expand_switch_using_bit_tests_p): Likewise.
6261 (process_switch): Compute and pass on speed_p based on the
6263 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6264 optimize_bb_for_speed_p.
6266 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6268 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6269 * function.h (struct function): Rename has_force_vect_loops into
6270 has_force_vectorize_loops.
6271 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6272 (input_struct_function_base): Likewise.
6273 * lto-streamer-out.c (output_cfg): Likewise.
6274 (output_struct_function_base): Likewise.
6275 * omp-low.c (expand_omp_simd): Likewise.
6276 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6277 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6278 (version_loop_for_if_conversion): Likewise.
6279 (tree_if_conversion): Likewise.
6280 (main_tree_if_conversion): Likewise.
6281 (gate_tree_if_conversion): Likewise.
6282 * tree-inline.c (copy_loops): Likewise.
6283 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6284 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6285 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6286 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6287 * tree-vectorizer.c (vectorize_loops): Likewise.
6288 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6290 2014-04-14 Richard Biener <rguenther@suse.de>
6293 * lto-streamer-out.c (wrap_refs): New function.
6294 (lto_output): Wrap symbol references in global initializes in
6295 type-preserving MEM_REFs.
6297 2014-04-14 Christian Bruel <christian.bruel@st.com>
6299 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6301 2014-04-14 Christian Bruel <christian.bruel@st.com>
6303 * config/sh/sh.md (setmemqi): New expand pattern.
6304 * config/sh/sh.h (CLEAR_RATIO): Define.
6305 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6306 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6308 2014-04-14 Richard Biener <rguenther@suse.de>
6311 * fold-const.c (negate_expr_p): Don't negate directional rounding
6313 (fold_negate_expr): Likewise.
6315 2014-04-14 Richard Biener <rguenther@suse.de>
6317 PR tree-optimization/59817
6318 PR tree-optimization/60453
6319 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6320 recursion to catch all CHRECs in the scalar evolution and restrict
6321 the predicate for the remains appropriately.
6323 2014-04-12 Catherine Moore <clm@codesourcery.com>
6325 * config/mips/constraints.md: Add new register constraint "kb".
6326 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6327 (*movhi_internal): Likewise.
6328 (*movqi_internal): Likewise.
6329 * config/mips/mips.h (M16_STORE_REGS): New register class.
6330 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6331 (REG_CLASS_CONTENTS): Likewise.
6332 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6334 2014-04-11 Tobias Burnus <burnus@net-b.de>
6337 * doc/invoke.texi (-Wformat-signedness): Document it.
6338 (Wformat=2): Mention that this enables -Wformat-signedness.
6340 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6342 * common/config/epiphany/epiphany-common.c
6343 (epiphany_option_optimization_table): Enable section anchors by
6344 default at -O1 or higher.
6345 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6346 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6347 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6348 carries no extra cost.
6349 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6350 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6351 * config/epiphany/predicates.md (memclob_operand): New predicate.
6352 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6353 Use memclob_operand predicate and X constraint for operand 3.
6355 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6357 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6358 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6361 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6363 PR rtl-optimization/60651
6364 * mode-switching.c (optimize_mode_switching): Make sure to emit
6365 sets of a lower numbered entity before sets of a higher numbered
6366 entity to a mode of the same or lower priority.
6367 When creating a seginfo for a basic block that starts with a code
6368 label, move the insertion point past the code label.
6369 (new_seginfo): Document and enforce requirement that
6370 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6371 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6372 * doc/tm.texi: Regenerate.
6374 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6377 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6379 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6381 * BASE-VER: Set to 4.10.0.
6383 2014-04-11 Tobias Burnus <burnus@net-b.de>
6386 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6387 * doc/gcc.texi (Service): Update description in the @menu
6388 * doc/invoke.texi (Option Summary): Remove misplaced and
6391 2014-04-11 Steve Ellcey <sellcey@mips.com>
6392 Jakub Jelinek <jakub@redhat.com>
6395 * expr.c (convert_move): Use emit_store_flag_force instead of
6396 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6399 2014-04-11 Richard Biener <rguenther@suse.de>
6402 * varasm.c (assemble_alias): Avoid endless error reporting
6403 recursion by setting TREE_ASM_WRITTEN.
6405 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6407 * config/s390/s390.md: Add a splitter for NOT rtx.
6409 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6411 PR rtl-optimization/60663
6412 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6414 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6415 Jakub Jelinek <jakub@redhat.com>
6418 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6419 flag from decl_node to node.
6421 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6424 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6425 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6426 ameliorating the cases where it can be.
6428 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6431 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6433 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6434 (loadsync_<mode>): Change mode.
6435 (load_quadpti, store_quadpti): New.
6436 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6437 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6438 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6440 2014-04-09 Cong Hou <congh@google.com>
6443 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6446 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6448 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6449 instead of vnor to exploit possible fusion opportunity in the
6451 (altivec_expand_vec_perm_const_le): Likewise.
6453 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6455 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6456 (loadsync_<mode>): Change mode.
6457 (load_quadpti, store_quadpti): New.
6458 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6459 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6461 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6464 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6465 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6466 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6468 2014-04-08 Richard Biener <rguenther@suse.de>
6471 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6472 a 64bit widest int print double-int similar to on HWI64 hosts.
6474 2014-04-08 Richard Biener <rguenther@suse.de>
6476 PR tree-optimization/60785
6477 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6478 default defs properly.
6480 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6482 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6483 (Weffc++): Likewise.
6485 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6487 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6488 set completep to false rather than true.
6490 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6493 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6494 ARM_TARGET2_DWARF_FORMAT.
6496 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6499 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6500 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6503 2014-04-07 Richard Biener <rguenther@suse.de>
6505 PR tree-optimization/60766
6506 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6507 (may_eliminate_iv): Convert cand_value_at result to desired type.
6509 2014-04-07 Jason Merrill <jason@redhat.com>
6512 * common.opt (-fno-gnu-unique): Add.
6513 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6515 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6517 * haifa-sched.c: Fix outdated function reference and minor
6518 grammar errors in introductory comment.
6520 2014-04-07 Richard Biener <rguenther@suse.de>
6523 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6525 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6527 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6530 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6531 size accounting for thunks.
6532 (pa_asm_output_mi_thunk): Use final_start_function() and
6533 final_end_function() to output function start and end directives.
6535 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6537 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6538 device specific ISA/ feature information. Remove short_sp and
6539 errata_skip ds. Add avr_device_specific_features enum to have device
6541 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6542 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6543 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6544 updated device specific info.
6545 * config/avr/avr-mcus.def: Merge device specific details to
6546 dev_attribute field.
6547 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6549 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6550 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6551 assembler if RMW isa supported by current device.
6552 * config/avr/genmultilib.awk: Update as device info structure changed.
6553 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6555 2014-04-04 Cong Hou <congh@google.com>
6557 PR tree-optimization/60656
6558 * tree-vect-stmts.c (supportable_widening_operation):
6559 Fix a bug that elements in a vector with vect_used_by_reduction
6560 property are incorrectly reordered when the operation on it is not
6561 consistant with the one in reduction operation.
6563 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6565 PR rtl-optimization/60155
6566 * gcse.c (record_set_data): New function.
6567 (single_set_gcse): New function.
6568 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6569 (hoist_code): Likewise.
6570 (get_pressure_class_and_nregs): Likewise.
6572 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6574 * explow.c (probe_stack_range): Emit a final optimization blockage.
6576 2014-04-04 Anthony Green <green@moxielogic.com>
6578 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6581 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6584 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6585 flags are set only during streaming.
6586 * ipa.c (process_references, walk_polymorphic_call_targets,
6587 symtab_remove_unreachable_nodes): Drop bodies of always inline
6588 after early inlining.
6589 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6591 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6592 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6595 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6598 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6601 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6603 (cortex_a53_fdivd): Likewise.
6605 2014-04-04 Martin Jambor <mjambor@suse.cz>
6608 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6610 * cgraph.c (clone_of_p): Also return true if thunks match.
6611 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6612 cgraph_function_or_thunk_node and an obsolete comment.
6613 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6615 (build_function_decl_skip_args): Likewise.
6616 (set_new_clone_decl_and_node_flags): New function.
6617 (duplicate_thunk_for_node): Likewise.
6618 (redirect_edge_duplicating_thunks): Likewise.
6619 (cgraph_clone_node): New parameter args_to_skip, pass it to
6620 redirect_edge_duplicating_thunks which is called instead of
6621 cgraph_redirect_edge_callee.
6622 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6623 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6625 2014-04-04 Jeff Law <law@redhat.com>
6628 * config/arm/predicates.md (const_int_I_operand): New predicate.
6629 (const_int_M_operand): Similarly.
6630 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6632 (insv_t2, extv_reg, extzv_t2): Likewise.
6633 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6634 (pop_multiple_with_writeback_and_return): Likewise.
6635 (vfp_pop_multiple_with_writeback): Likewise
6637 2014-04-04 Richard Biener <rguenther@suse.de>
6640 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6641 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6643 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6644 * gimplify.c (gimple_add_tmp_var_fn): New function.
6645 * gimple-expr.h (create_tmp_reg_fn): Declare.
6646 * gimple-expr.c (create_tmp_reg_fn): New function.
6647 * gimple-low.c (record_vars_into): Don't change cfun.
6648 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6649 code generation without cfun.
6651 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6654 * Makefile.in (install-driver): Fix shell scripting.
6656 2014-04-03 Cong Hou <congh@google.com>
6658 PR tree-optimization/60505
6659 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6660 threshold of number of iterations below which no vectorization
6662 * tree-vect-loop.c (new_loop_vec_info):
6663 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6664 * tree-vect-loop.c (vect_analyze_loop_operations):
6665 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6666 * tree-vect-loop.c (vect_transform_loop):
6667 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6668 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6669 of iterations of the loop and see if we should build the epilogue.
6671 2014-04-03 Richard Biener <rguenther@suse.de>
6673 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6674 (streamer_tree_cache_create): Adjust.
6675 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6676 to allow optional nodes array.
6677 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6678 (streamer_tree_cache_append): Likewise.
6679 (streamer_tree_cache_create): Create nodes array optionally
6680 as specified by parameter.
6681 * lto-streamer-out.c (create_output_block): Avoid maintaining
6682 the node array in the writer cache.
6683 (DFS_write_tree): Remove assertion.
6684 (produce_asm_for_decls): Free the out decl state hash table early.
6685 * lto-streamer-in.c (lto_data_in_create): Adjust for
6686 streamer_tree_cache_create prototype change.
6688 2014-04-03 Richard Biener <rguenther@suse.de>
6690 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6691 set TREE_CHAIN to NULL_TREE.
6693 2014-04-03 Richard Biener <rguenther@suse.de>
6695 PR tree-optimization/60740
6696 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6697 over all GIMPLE_COND operands.
6699 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6701 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6702 (Weffc++): Remove Scott's numbering, merge lists and reference
6705 2014-04-03 Nick Clifton <nickc@redhat.com>
6707 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6710 2014-04-03 Martin Jambor <mjambor@suse.cz>
6712 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6713 mention gcc_unreachable before failing.
6714 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6717 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6720 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6721 inconsistent code and instead mark the context inconsistent.
6722 (possible_polymorphic_call_targets): For inconsistent contexts
6723 return empty complete list.
6725 2014-04-02 Anthony Green <green@moxielogic.com>
6727 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6728 (extendqisi2, extendhisi2): Define.
6729 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6730 (WCHAR_TYPE): Change to unsigned int.
6732 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6734 PR tree-optimization/60733
6735 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6736 insertion point for PHI candidates to be the end of the feeding
6737 block for the PHI argument.
6739 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6741 PR rtl-optimization/60650
6742 * lra-constraints.c (process_alt_operands): Decrease reject for
6743 earlyclobber matching.
6745 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6747 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6749 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6751 * config/spu/spu.c (pad_bb): Do not crash when the last
6752 insn is CODE_FOR_blockage.
6754 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6756 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6757 lies outside the target mode.
6759 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6762 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6763 software floating point or no floating point registers, do not
6764 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6765 in GPRs that occurs after we tested for GPRs that would never be
6768 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6769 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6770 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6771 specifically allow DDmode, since that does not use the SPE SIMD
6774 2014-04-02 Richard Biener <rguenther@suse.de>
6777 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6778 MODE_INTs. Properly use negv_optab.
6779 (expand_abs): Likewise.
6781 2014-04-02 Richard Biener <rguenther@suse.de>
6784 * Makefile.in (install-driver): Guard extra installs with special
6787 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6789 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6792 2014-04-01 Richard Henderson <rth@redhat.com>
6795 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6796 alternative enabled before register allocation.
6798 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6800 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6801 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6803 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6804 (nios2_got_address): Update nios2_large_got_address call site.
6805 (nios2_delegitimize_address): New function.
6806 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6807 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6808 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6810 2014-04-01 Martin Husemann <martin@duskware.de>
6812 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6815 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6817 PR rtl-optimization/60604
6818 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6819 check from register_operand.
6820 (register_operand): Redefine in terms of general_operand.
6821 (nonmemory_operand): Use register_operand for the non-constant cases.
6823 2014-04-01 Richard Biener <rguenther@suse.de>
6825 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6827 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6829 * doc/invoke.texi (mapp-regs): Clarify.
6831 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6833 * config/i386/avx512fintrin.h (__v32hi): Define type.
6834 (__v64qi): Likewise.
6835 (_mm512_set1_epi8): Define.
6836 (_mm512_set1_epi16): Define.
6837 (_mm512_set4_epi32): Define.
6838 (_mm512_set4_epi64): Define.
6839 (_mm512_set4_pd): Define.
6840 (_mm512_set4_ps): Define.
6841 (_mm512_setr4_epi64): Define.
6842 (_mm512_setr4_epi32): Define.
6843 (_mm512_setr4_pd): Define.
6844 (_mm512_setr4_ps): Define.
6845 (_mm512_setzero_epi32): Define.
6847 2014-03-31 Martin Jambor <mjambor@suse.cz>
6850 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6851 callsite_arguments_match_p. Updated all callers. Also check types of
6852 corresponding formal parameters and actual arguments.
6853 (not_all_callers_have_enough_arguments_p) Renamed to
6854 some_callers_have_mismatched_arguments_p.
6856 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6858 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6860 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6863 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6866 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6868 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6869 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6871 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6872 Use FMAMODE_NOVF512 mode iterator.
6873 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6874 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
6875 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
6876 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
6878 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
6879 Use VF_128_256 mode iterator.
6880 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
6883 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6885 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
6886 static chain if needed.
6888 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6891 * lra-constraints.c (index_part_to_reg): New.
6892 (process_address): Use it.
6894 2014-03-27 Jeff Law <law@redhat.com>
6895 Jakub Jelinek <jakub@redhat.com>
6898 * expr.c (do_tablejump): Use simplify_gen_binary rather than
6899 gen_rtx_{PLUS,MULT} to build up the address expression.
6901 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
6902 creating non-canonical RTL.
6904 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6907 * ipa-inline.c (want_inline_small_function_p): Short circuit large
6908 functions; reorganize to make cheap checks first.
6909 (inline_small_functions): Do not estimate growth when dumping;
6911 * ipa-inline.h (inline_summary): Add min_size.
6912 (growth_likely_positive): New function.
6913 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
6914 (set_cond_stmt_execution_predicate): Cleanup.
6915 (estimate_edge_size_and_time): Compute min_size.
6916 (estimate_calls_size_and_time): Likewise.
6917 (estimate_node_size_and_time): Likewise.
6918 (inline_update_overall_summary): Update min_size.
6919 (do_estimate_edge_time): Likewise.
6920 (do_estimate_edge_size): Update.
6921 (do_estimate_edge_hints): Update.
6922 (growth_likely_positive): New function.
6924 2014-03-28 Jakub Jelinek <jakub@redhat.com>
6927 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
6928 also if addr has VOIDmode.
6930 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6932 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
6933 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
6935 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
6936 instructions as well as AdvancedSIMD loads.
6938 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6940 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
6941 Use crypto_aese type.
6942 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
6943 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
6944 crypto_aese, crypto_aesmc. Move to types.md.
6945 * config/arm/types.md (crypto_aes): Split into crypto_aese,
6947 * config/arm/iterators.md (crypto_type): Likewise.
6949 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
6951 * cgraph.c: Include expr.h and tree-dfa.h.
6952 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
6955 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
6958 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
6959 regs from checking multi-reg pseudos.
6961 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6963 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
6965 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6967 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
6968 if it would clobber the stack pointer, even temporarily.
6970 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
6972 * mode-switching.c: Make small adjustments to the top comment.
6974 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6976 * config/rs6000/constraints.md (wD constraint): New constraint to
6977 match the constant integer to get the top DImode/DFmode out of a
6978 vector in a VSX register.
6980 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
6981 match the constant integer to get the top DImode/DFmode out of a
6982 vector in a VSX register.
6984 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
6987 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
6990 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
6991 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
6993 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
6994 Optimize vec_extract of 64-bit values, where the value being
6995 extracted is in the top word, where we can use scalar
6996 instructions. Add direct move and store support. Combine the big
6997 endian/little endian vector select load support into a single insn.
6998 (vsx_extract_<mode>_internal1): Likewise.
6999 (vsx_extract_<mode>_internal2): Likewise.
7000 (vsx_extract_<mode>_load): Likewise.
7001 (vsx_extract_<mode>_store): Likewise.
7002 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7003 combined into vsx_extract_<mode>_load.
7004 (vsx_extract_<mode>_one_le): Likewise.
7006 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7007 define the top 64-bit vector element.
7009 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7012 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7013 Document vec_vbpermq builtin.
7016 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7017 enable use of xxsldwi and xxpermdi builtin functions.
7018 (vec_xxpermdi): Likewise.
7020 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7021 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7023 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7025 PR rtl-optimization/60650
7026 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7028 (find_spills_for): New.
7029 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7030 Spill all pseudos on the second iteration.
7032 2014-03-27 Marek Polacek <polacek@redhat.com>
7035 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7038 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7040 * config/s390/s390.c (s390_can_use_return_insn): Check for
7041 call-saved FPRs on 31 bit.
7043 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7046 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7047 if they need regimplification, just drop them instead of
7048 calling gimple_regimplify_operands on them.
7050 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7053 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7054 (aarch64_frame_pointer_required): Adjust logic.
7055 (aarch64_can_eliminate): Adjust logic.
7056 (aarch64_override_options_after_change): Adjust logic.
7058 2014-03-27 Dehao Chen <dehao@google.com>
7060 * ipa-inline.c (early_inliner): Update node's inline info.
7062 2014-03-26 Dehao Chen <dehao@google.com>
7064 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7065 compiler inserted conditional jumps for NAN float check.
7067 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7069 * ubsan.h (ubsan_create_data): Change second argument's type
7070 to const location_t *.
7071 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7073 (ubsan_create_data): Change second argument to const location_t *PLOC.
7074 Create Loc field whenever PLOC is non-NULL.
7075 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7076 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7080 * real.c (real_to_integer2): Change type of low to UHWI.
7082 2014-03-26 Tobias Burnus <burnus@net-b.de>
7084 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7085 (CILK_SELF_SPECS): New define.
7086 (driver_self_specs): Use it.
7088 2014-03-26 Richard Biener <rguenther@suse.de>
7090 * tree-pretty-print.c (percent_K_format): Implement special
7091 case for LTO and its stripped down BLOCK tree.
7093 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7096 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7098 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7099 one range is range_int_cst_p, but not both, at least optimize
7100 addition/subtraction of 0 and multiplication by 0 or 1.
7101 * gimple-fold.c (gimple_fold_call): Fold
7102 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7103 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7104 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7106 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7108 PR rtl-optimization/60452
7109 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7110 <case REG>: Return 1 for invalid offsets from the frame pointer.
7112 2014-03-26 Marek Polacek <polacek@redhat.com>
7115 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7118 2014-03-26 Marek Polacek <polacek@redhat.com>
7121 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7124 2014-03-26 Marek Polacek <polacek@redhat.com>
7127 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7128 multiplication in unsigned type.
7130 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7132 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7134 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7136 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7138 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7141 * cif-code.def (UNREACHABLE) New code.
7142 * ipa-inline.c (inline_small_functions): Skip edges to
7143 __builtlin_unreachable.
7144 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7145 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7146 predicate to __bulitin_unreachable.
7147 (set_cond_stmt_execution_predicate): Fix issue when
7148 invert_tree_comparison returns ERROR_MARK.
7149 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7150 propagate to inline clones.
7151 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7153 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7154 * cgraphclones.c (cgraph_clone_node): If call destination is already
7155 ureachable, do not redirect it back.
7156 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7159 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7161 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7162 Do not modify inline clones.
7164 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7166 * config/i386/i386.md (general_sext_operand): New mode attr.
7167 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7168 don't generate (sign_extend (const_int)).
7169 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7170 operands[2]. Use We constraint instead of <i> and
7171 <general_sext_operand> predicate instead of <general_operand>.
7172 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7173 * config/i386/constraints.md (We): New constraint.
7174 * config/i386/predicates.md (x86_64_sext_operand,
7175 sext_operand): New predicates.
7177 2014-03-25 Martin Jambor <mjambor@suse.cz>
7180 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7181 inconsistent devirtualizations to __builtin_unreachable.
7183 2014-03-25 Marek Polacek <polacek@redhat.com>
7186 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7188 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7190 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7191 order of elements for big-endian.
7193 2014-03-25 Richard Biener <rguenther@suse.de>
7196 * gimplify-me.c (gimple_regimplify_operands): Update the
7199 2014-03-25 Martin Jambor <mjambor@suse.cz>
7202 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7203 (lto_output_varpool_node): Likewise.
7204 (input_overwrite_node): Likewise.
7205 (input_varpool_node): Likewise.
7207 2014-03-25 Richard Biener <rguenther@suse.de>
7209 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7210 (run_gcc): Likewise.
7212 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7214 * combine.c (simplify_compare_const): Add MODE argument.
7215 Handle mode_width 0 as very large mode_width.
7216 (try_combine, simplify_comparison): Adjust callers.
7218 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7219 type to avoid signed integer overflow.
7220 * explow.c (plus_constant): Likewise.
7222 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7224 * doc/generic.texi: Correct typos.
7226 2014-03-24 Tobias Burnus <burnus@net-b.de>
7228 * doc/invoke.texi (-flto): Expand section about
7229 using static libraries with LTO.
7231 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7233 PR rtl-optimization/60501
7234 * optabs.def (addptr3_optab): New optab.
7235 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7236 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7237 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7239 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7241 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7243 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7245 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7248 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7249 (_mm256_undefined_ps): Define.
7250 (_mm256_undefined_pd): Define.
7251 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7252 (_mm_undefined_pd): Define.
7253 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7254 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7255 (_mm512_undefined_ps): Define.
7256 (_mm512_undefined_pd): Define.
7257 Use _mm*_undefined_*.
7258 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7260 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7262 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7263 (lshr_simd): DI mode added.
7264 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7265 (aarch64_ushr_simddi): Likewise.
7266 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7267 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7268 (vshrd_n_u64): Likewise.
7270 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7272 * Makefile.in (s-macro_list): Depend on cc1.
7274 2014-03-23 Teresa Johnson <tejohnson@google.com>
7276 * ipa-utils.c (ipa_print_order): Use specified dump file.
7278 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7280 PR rtl-optimization/60601
7281 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7283 * gcc.c (eval_spec_function): Initialize save_growing_value.
7285 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7288 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7289 code == MINUS_EXPR, never swap op0 with op1.
7291 * toplev.c (init_local_tick): Avoid signed integer multiplication
7293 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7294 shift by first operand's bitsize.
7296 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7299 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7301 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7302 TARGET_ISA_64BIT_P(x).
7304 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7306 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7307 pattern for vector nor instead of subtract from splat(-1).
7308 (altivec_expand_vec_perm_const_le): Likewise.
7310 2014-03-21 Richard Henderson <rth@twiddle.net>
7313 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7314 related insns after epilogue_completed.
7316 2014-03-21 Martin Jambor <mjambor@suse.cz>
7319 * cgraph.h (symtab_node): New flag body_removed.
7320 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7321 when removing bodies.
7322 * symtab.c (dump_symtab_base): Dump body_removed flag.
7323 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7324 had their bodies removed.
7326 2014-03-21 Martin Jambor <mjambor@suse.cz>
7329 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7332 2014-03-21 Richard Biener <rguenther@suse.de>
7334 PR tree-optimization/60577
7335 * tree-core.h (struct tree_base): Document nothrow_flag use
7336 in VAR_DECL_NONALIASED.
7337 * tree.h (VAR_DECL_NONALIASED): New.
7338 (may_be_aliased): Adjust.
7339 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7341 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7343 * expr.c (expand_expr_real_1): Remove outdated comment.
7345 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7348 * ira.c (adjust_cleared_regs): Call copy_rtx on
7349 *reg_equiv[REGNO (loc)].src_p before passing it to
7350 simplify_replace_fn_rtx.
7353 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7354 into CONST, put pic register as first operand of PLUS. Use
7355 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7357 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7359 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7361 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7363 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7364 around for store forwarding issue in the FPU on the UT699.
7365 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7366 loads and operations if -mfix-ut699 is specified.
7367 (divtf3_hq): Tweak attribute.
7368 (sqrttf2_hq): Likewise.
7370 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7372 * calls.c (store_one_arg): Remove incorrect const qualification on the
7373 type of the temporary.
7374 * cfgexpand.c (expand_return): Likewise.
7375 * expr.c (expand_constructor): Likewise.
7376 (expand_expr_real_1): Likewise.
7378 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7380 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7383 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7386 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7388 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7390 * config/arm/aarch-common-protos.h
7391 (alu_cost_table): Fix spelling of "extend".
7392 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7394 2014-03-19 Richard Biener <rguenther@suse.de>
7397 * tree-core.h (tree_type_common): Re-order pointer members
7398 to reduce recursion depth during GC walks.
7400 2014-03-19 Marek Polacek <polacek@redhat.com>
7403 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7404 before accessing it.
7406 2014-03-19 Richard Biener <rguenther@suse.de>
7409 * lto-streamer-in.c (input_function): In WPA stage do not drop
7412 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7414 PR tree-optimization/60559
7415 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7416 with build_zero_cst assignment.
7418 2014-03-18 Kai Tietz <ktietz@redhat.com>
7420 PR rtl-optimization/56356
7421 * sdbout.c (sdbout_parms): Verify that parms'
7422 incoming argument is valid.
7423 (sdbout_reg_parms): Likewise.
7425 2014-03-18 Richard Henderson <rth@redhat.com>
7428 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7429 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7430 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7432 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7434 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7435 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7436 Italicize plugin event names in description. Explain that
7437 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7438 Remind that no GCC functions should be called after PLUGIN_FINISH.
7439 Explain what pragmas with expansion are.
7441 2014-03-18 Martin Liska <mliska@suse.cz>
7443 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7444 gimple call statement is update.
7445 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7446 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7448 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7451 * ubsan.c (ubsan_instrument_unreachable): Call
7452 initialize_sanitizer_builtins.
7453 (ubsan_pass): Likewise.
7456 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7457 varpool_finalize_decl instead of rest_of_decl_compilation.
7459 2014-03-18 Richard Biener <rguenther@suse.de>
7461 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7462 by using bitmap_and_compl instead of bitmap_and_compl_into.
7463 (df_rd_transfer_function): Likewise.
7465 2014-03-18 Richard Biener <rguenther@suse.de>
7467 * doc/lto.texi (fresolution): Fix typo.
7469 2014-03-18 Richard Biener <rguenther@suse.de>
7471 * doc/invoke.texi (flto): Update for changes in 4.9.
7473 2014-03-18 Richard Biener <rguenther@suse.de>
7475 * doc/loop.texi: Remove section on the removed lambda framework.
7476 Update loop docs with recent changes in preserving loop structure.
7478 2014-03-18 Richard Biener <rguenther@suse.de>
7480 * doc/lto.texi (-fresolution): Document.
7482 2014-03-18 Richard Biener <rguenther@suse.de>
7484 * doc/contrib.texi: Adjust my name.
7486 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7489 * internal-fn.c: Include diagnostic-core.h.
7490 (expand_BUILTIN_EXPECT): New function.
7491 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7492 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7493 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7494 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7496 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7497 Revert 3 argument __builtin_expect code.
7498 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7499 * gimple-fold.c (gimple_fold_call): Likewise.
7500 * tree.h (fold_builtin_expect): New prototype.
7501 * builtins.c (build_builtin_expect_predicate): Add predictor
7502 argument, if non-NULL, create 3 argument __builtin_expect.
7503 (fold_builtin_expect): No longer static. Add ARG2 argument,
7504 pass it through to build_builtin_expect_predicate.
7505 (fold_builtin_2): Adjust caller.
7506 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7507 * internal-fn.def (BUILTIN_EXPECT): New.
7509 2014-03-18 Tobias Burnus <burnus@net-b.de>
7512 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7513 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7514 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7516 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7519 * predict.c (combine_predictions_for_bb): Fix up formatting.
7520 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7521 fill what it points to if non-NULL.
7522 (tree_predict_by_opcode): Adjust caller, use the predictor.
7523 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7525 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7527 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7528 proper constant for the store mode.
7530 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7532 * symtab.c (change_decl_assembler_name): Fix transparent alias
7535 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7537 * config/aarch64/aarch64.c: Correct the comments about the
7538 aarch64 stack layout.
7540 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7542 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7543 check for GF_OMP_FOR_KIND_FOR.
7545 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7547 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7548 ymm and zmm register names.
7550 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7553 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7554 note creation for the 2010-08-31 changes.
7556 2014-03-17 Marek Polacek <polacek@redhat.com>
7559 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7560 as -fno-tree-loop-vectorize.
7561 (expand_omp_simd): Likewise.
7563 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7565 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7566 (eligible_for_call_delay): New prototype.
7567 * config/sparc/sparc.c (tls_call_delay): Rename into...
7568 (eligible_for_call_delay): ...this. Return false if the instruction
7569 cannot be put in the delay slot of a branch.
7570 (eligible_for_restore_insn): Simplify.
7571 (eligible_for_return_delay): Return false if the instruction cannot be
7572 put in the delay slot of a branch and simplify.
7573 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7574 put in the delay slot of a branch.
7575 * config/sparc/sparc.md (fix_ut699): New attribute.
7576 (tls_call_delay): Delete.
7577 (in_call_delay): Reimplement.
7578 (eligible_for_sibcall_delay): Rename into...
7579 (in_sibcall_delay): ...this.
7580 (eligible_for_return_delay): Rename into...
7581 (in_return_delay): ...this.
7582 (in_branch_delay): Reimplement.
7583 (in_uncond_branch_delay): Delete.
7584 (in_annul_branch_delay): Delete.
7586 2014-03-14 Richard Henderson <rth@redhat.com>
7589 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7590 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7591 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7592 (floathi splitters): Remove.
7593 (float<SWI48x>xf2): New pattern.
7594 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7595 code that tried to handle DImode for 32-bit, but which was excluded
7596 by the pattern's condition. Drop allocation of stack temporary.
7597 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7598 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7599 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7600 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7601 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7602 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7603 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7604 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7605 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7606 (*float<SWI48x><X87MODEF>2_i387): Remove.
7607 (all float _with_temp splitters): Remove.
7608 (*float<SWI48x><MODEF>2_i387): New pattern.
7609 (*float<SWI48><MODEF>2_sse): New pattern.
7610 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7611 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7613 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7614 Marek Polacek <polacek@redhat.com>
7617 * common.opt (dump_base_name_prefixed): New Variable.
7618 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7619 if x_dump_base_name_prefixed is already set, set it at the end.
7621 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7623 PR rtl-optimization/60508
7624 * lra-constraints.c (get_reload_reg): Add new parameter
7626 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7627 Pass the new parameter values.
7629 2014-03-14 Richard Biener <rguenther@suse.de>
7631 * common.opt: Revert unintented changes from r205065.
7634 2014-03-14 Richard Biener <rguenther@suse.de>
7637 * cfghooks.c (split_block): Properly adjust all loops the
7638 block was a latch of.
7640 2014-03-14 Martin Jambor <mjambor@suse.cz>
7643 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7646 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7649 * config/avr/avr.c (avr_set_current_function): Pass function name
7650 through default_strip_name_encoding before sanity checking instead
7651 of skipping the first char of the assembler name.
7653 2014-03-13 Richard Henderson <rth@redhat.com>
7656 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7657 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7658 * config/i386/i386-protos.h: Likewise.
7659 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7660 in the expander instead of a splitter.
7661 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7662 any possibility of requiring a memory.
7663 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7664 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7665 (fp branch splitters): Update for ix86_split_fp_branch.
7666 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7667 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7668 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7669 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7670 (*fop_<MODEF>_3_i387): Likewise.
7671 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7672 (splitters for the fop_* register patterns): Remove.
7673 (fscalexf4_i387): Rename from *fscalexf4_i387.
7674 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7676 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7678 PR tree-optimization/59779
7679 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7680 type for bitsize and maxsize instead of HOST_WIDE_INT.
7682 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7684 PR rtl-optimization/57320
7685 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7686 the CFG after thread_prologue_and_epilogue_insns.
7688 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7690 PR rtl-optimization/57189
7691 * lra-constraints.c (process_alt_operands): Disfavor spilling
7694 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7696 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7698 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7700 PR tree-optimization/59025
7702 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7703 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7705 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7708 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7709 calls of avr_out_plus_1.
7711 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7713 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7714 BB's single pred and update the father loop's latch info later.
7716 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7718 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7722 (VEC_base): Likewise.
7723 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7724 registers, we need to swap double words in little endian mode.
7726 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7727 to be a container mode for 128-bit integer operations added in ISA
7728 2.07. Unlike TImode and PTImode, the preferred register set is
7729 the Altivec/VMX registers for the 128-bit operations.
7731 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7733 (rs6000_split_128bit_ok_p): Likewise.
7735 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7736 macros for creating ISA 2.07 normal and overloaded builtin
7737 functions with 3 arguments.
7738 (BU_P8V_OVERLOAD_3): Likewise.
7739 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7740 for use as overloaded functions.
7741 (VPERM_1TI_UNS): Likewise.
7742 (VSEL_1TI): Likewise.
7743 (VSEL_1TI_UNS): Likewise.
7744 (ST_INTERNAL_1ti): Likewise.
7745 (LD_INTERNAL_1ti): Likewise.
7746 (XXSEL_1TI): Likewise.
7747 (XXSEL_1TI_UNS): Likewise.
7748 (VPERM_1TI): Likewise.
7749 (VPERM_1TI_UNS): Likewise.
7750 (XXPERMDI_1TI): Likewise.
7751 (SET_1TI): Likewise.
7752 (LXVD2X_V1TI): Likewise.
7753 (STXVD2X_V1TI): Likewise.
7754 (VEC_INIT_V1TI): Likewise.
7755 (VEC_SET_V1TI): Likewise.
7756 (VEC_EXT_V1TI): Likewise.
7757 (EQV_V1TI): Likewise.
7758 (NAND_V1TI): Likewise.
7759 (ORC_V1TI): Likewise.
7760 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7761 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7763 (VADDUQM): Likewise.
7764 (VSUBCUQ): Likewise.
7765 (VADDEUQM): Likewise.
7766 (VADDECUQ): Likewise.
7767 (VSUBEUQM): Likewise.
7768 (VSUBECUQ): Likewise.
7770 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7771 __int128_t and __uint128_t types.
7772 (__uint128_type): Likewise.
7773 (altivec_categorize_keyword): Add support for vector __int128_t,
7774 vector __uint128_t, vector __int128, and vector unsigned __int128
7775 as a container type for TImode operations that need to be done in
7776 VSX/Altivec registers.
7777 (rs6000_macro_to_expand): Likewise.
7778 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7779 to support 128-bit integer instructions vaddcuq, vadduqm,
7780 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7781 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7783 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7784 for V1TImode, and set up preferences to use VSX/Altivec registers.
7785 Setup VSX reload handlers.
7786 (rs6000_debug_reg_global): Likewise.
7787 (rs6000_init_hard_regno_mode_ok): Likewise.
7788 (rs6000_preferred_simd_mode): Likewise.
7789 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7790 (easy_altivec_constant): Likewise.
7791 (output_vec_const_move): Likewise.
7792 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7794 (rs6000_expand_vector_extract): Likewise.
7795 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7797 (rs6000_const_vec): Add support for V1TImode.
7798 (rs6000_emit_le_vsx_load): Swap double words when loading or
7799 storing TImode/V1TImode.
7800 (rs6000_emit_le_vsx_store): Likewise.
7801 (rs6000_emit_le_vsx_move): Likewise.
7802 (rs6000_emit_move): Add support for V1TImode.
7803 (altivec_expand_ld_builtin): Likewise.
7804 (altivec_expand_st_builtin): Likewise.
7805 (altivec_expand_vec_init_builtin): Likewise.
7806 (altivec_expand_builtin): Likewise.
7807 (rs6000_init_builtins): Add support for V1TImode type. Add
7808 support for ISA 2.07 128-bit integer builtins. Define type names
7809 for the VSX/Altivec vector types.
7810 (altivec_init_builtins): Add support for overloaded vector
7811 functions with V1TImode type.
7812 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7813 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7815 (rs6000_split_128bit_ok_p): Likewise.
7816 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7817 __int128_t and vector __uint128_t.
7819 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7820 and mode attributes.
7827 (VS_scalar): Likewise.
7828 (VS_double): Likewise.
7829 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7831 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7832 we support the ISA 2.07 128-bit integer arithmetic instructions.
7833 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7834 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7835 and TImode types for use with the builtin functions.
7836 (V1TI_type_node): Likewise.
7837 (unsigned_V1TI_type_node): Likewise.
7838 (intTI_type_internal_node): Likewise.
7839 (uintTI_type_internal_node): Likewise.
7841 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7842 128-bit builtin functions.
7843 (UNSPEC_VADDEUQM): Likewise.
7844 (UNSPEC_VADDECUQ): Likewise.
7845 (UNSPEC_VSUBCUQ): Likewise.
7846 (UNSPEC_VSUBEUQM): Likewise.
7847 (UNSPEC_VSUBECUQ): Likewise.
7848 (VM): Add V1TImode to vector mode iterators.
7850 (VI_unit): Likewise.
7851 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7852 (altivec_vaddcuq): Likewise.
7853 (altivec_vsubuqm): Likewise.
7854 (altivec_vsubcuq): Likewise.
7855 (altivec_vaddeuqm): Likewise.
7856 (altivec_vaddecuq): Likewise.
7857 (altivec_vsubeuqm): Likewise.
7858 (altivec_vsubecuq): Likewise.
7860 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7862 (BOOL_128): Likewise.
7863 (BOOL_REGS_OUTPUT): Likewise.
7864 (BOOL_REGS_OP1): Likewise.
7865 (BOOL_REGS_OP2): Likewise.
7866 (BOOL_REGS_UNARY): Likewise.
7867 (BOOL_REGS_AND_CR0): Likewise.
7869 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7870 128-bit integer builtin support.
7871 (vec_vadduqm): Likewise.
7872 (vec_vaddecuq): Likewise.
7873 (vec_vaddeuqm): Likewise.
7874 (vec_vsubecuq): Likewise.
7875 (vec_vsubeuqm): Likewise.
7876 (vec_vsubcuq): Likewise.
7877 (vec_vsubuqm): Likewise.
7879 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7880 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
7881 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
7882 128-bit integer add/subtract to ISA 2.07.
7884 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
7886 * config/arc/arc.c (arc_predicate_delay_insns):
7887 Fix third argument passed to conditionalize_nonjump.
7889 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
7891 * config/aarch64/aarch64-builtins.c
7892 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
7893 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
7894 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
7895 instead of __builtin_lfloor.
7896 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
7898 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7900 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
7901 (tree_ssa_ifcombine_bb_1): New function.
7902 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
7903 is an empty forwarder block to then_bb or vice versa and then_bb
7904 and else_bb are effectively swapped.
7906 2014-03-12 Christian Bruel <christian.bruel@st.com>
7909 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
7910 REG_CFA_DEF_CFA note.
7911 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
7912 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
7914 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
7916 PR tree-optimization/60454
7917 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
7919 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7921 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
7922 Do not define target_cpu_default2 to generic.
7923 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
7924 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
7925 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
7927 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7928 Marc Glisse <marc.glisse@inria.fr>
7930 PR tree-optimization/60502
7931 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
7932 instead of build_low_bits_mask.
7934 2014-03-12 Jakub Jelinek <jakub@redhat.com>
7937 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
7938 if there are multiple uses, but op doesn't live on E edge.
7939 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
7940 clobber stmts before __builtin_unreachable.
7942 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
7944 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
7945 hard_frame_pointer_rtx.
7946 * cse.c (cse_insn): Remove volatile check.
7947 * cselib.c (cselib_process_insn): Likewise.
7948 * dse.c (scan_insn): Likewise.
7950 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7952 * config/arc/arc.c (conditionalize_nonjump): New function,
7954 (arc_ifcvt): ... this.
7955 (arc_predicate_delay_insns): Use it.
7957 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
7959 * config/arc/predicates.md (extend_operand): During/after reload,
7960 allow const_int_operand.
7961 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
7962 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
7963 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
7965 (umulsi3_highpart_i): Likewise.
7967 2014-03-11 Richard Biener <rguenther@suse.de>
7969 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
7970 Add asserts to guard possible wrong-code bugs.
7972 2014-03-11 Richard Biener <rguenther@suse.de>
7974 PR tree-optimization/60429
7975 PR tree-optimization/60485
7976 * tree-ssa-structalias.c (set_union_with_increment): Properly
7977 take into account all fields that overlap the shifted vars.
7978 (do_sd_constraint): Likewise.
7979 (do_ds_constraint): Likewise.
7980 (get_constraint_for_ptr_offset): Likewise.
7982 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
7984 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
7985 (nios2_compute_frame_layout):
7986 Add calculation of cfun->machine->fp_save_offset.
7987 (nios2_expand_prologue): Correct setting of frame pointer register
7989 (nios2_expand_epilogue): Update recovery of stack pointer from
7990 frame pointer accordingly.
7991 (nios2_initial_elimination_offset): Update calculation of offset
7992 for eliminating to HARD_FRAME_POINTER_REGNUM.
7994 2014-03-10 Jakub Jelinek <jakub@redhat.com>
7997 * ipa.c (symtab_remove_unreachable_nodes): Don't call
7998 cgraph_get_create_node on VAR_DECLs.
8000 2014-03-10 Richard Biener <rguenther@suse.de>
8003 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8005 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8007 * config/vms/vms.opt (vms_float_format): New variable.
8009 2014-03-08 Tobias Burnus <burnus@net-b.de>
8011 * doc/invoke.texi (-fcilkplus): Update implementation status.
8013 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8014 Richard Biener <rguenther@suse.de>
8016 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8017 consistently accross all TUs.
8018 (run_gcc): Enable -fshort-double automatically at link at link-time
8019 and disallow override.
8021 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8024 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8025 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8026 if they can't be used.
8028 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8030 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8031 for Solaris 11/x86 ld.
8032 * configure: Regenerate.
8034 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8036 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8037 (LIB_TLS_SPEC): Save as ld_tls_libs.
8038 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8039 (HAVE_AS_IX86_TLSLDM): New test.
8040 * configure, config.in: Regenerate.
8041 * config/i386/i386.c (legitimize_tls_address): Fall back to
8042 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8043 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8044 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8045 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8047 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8049 * common.opt (fira-loop-pressure): Mark as optimization.
8051 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8053 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8054 an OpenMP mappable type.
8056 2014-03-06 Matthias Klose <doko@ubuntu.com>
8058 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8059 MULTILIB_OSDIRNAMES is not defined.
8061 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8062 Meador Inge <meadori@codesourcery.com>
8065 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8066 (arm_legitimize_address): Call legitimize_tls_address for any
8067 arm_tls_referenced_p expression, handle constant addend. Call it
8068 before testing for !TARGET_ARM.
8069 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8071 2014-03-06 Richard Biener <rguenther@suse.de>
8077 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8079 * tree-streamer.c (record_common_node): Assert we don't record
8080 nodes with type double.
8081 (preload_common_node): Skip type double, complex double and double
8082 pointer since it is now frontend dependent due to fshort-double option.
8084 2014-03-06 Richard Biener <rguenther@suse.de>
8086 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8087 or -fno-lto is specified and the linker has full plugin support.
8088 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8089 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8090 * lto-wrapper.c (merge_and_complain): Merge compile-time
8091 optimization levels.
8092 (run_gcc): And pass it through to the link options.
8094 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8098 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8100 * cselib.c (remove_useless_values): Skip to avoid quadratic
8101 behavior if the condition moved from...
8102 (cselib_process_insn): ... here holds.
8104 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8107 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8108 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8111 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8112 (TM_H): Add x86-tune.def.
8114 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8116 * config/aarch64/aarch64.c (generic_tunings):
8117 Use cortexa57_extra_costs.
8119 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8122 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8123 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8124 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8127 2014-03-04 Heiher <r@hev.cc>
8129 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8130 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8132 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8134 * config/i386/predicates.md (const2356_operand): Change to ...
8135 (const2367_operand): ... this.
8136 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8138 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8139 (*avx512pf_scatterpf<mode>sf): Ditto.
8140 (avx512pf_scatterpf<mode>df): Ditto.
8141 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8142 (*avx512pf_scatterpf<mode>df): Ditto.
8143 * config/i386/i386.c (ix86_expand_builtin): Update
8144 incorrect hint operand error message.
8146 2014-03-04 Richard Biener <rguenther@suse.de>
8148 * lto-section-in.c (lto_get_section_data): Fix const cast.
8150 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8152 * tree-streamer.c (record_common_node): Assert we don't record
8153 nodes with type double.
8154 (preload_common_node): Skip type double, complex double and double
8155 pointer since it is now frontend dependent due to fshort-double option.
8157 2014-03-04 Richard Biener <rguenther@suse.de>
8160 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8161 (lto_input_toplevel_asms): Likewise.
8162 * lto-section-in.c (lto_get_section_data): Instead do it here
8165 2014-03-04 Richard Biener <rguenther@suse.de>
8167 PR tree-optimization/60382
8168 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8169 dead PHIs a reduction.
8171 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8173 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8175 (_mm_prefetch): Move out of GCC target("sse") pragma.
8176 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8177 GCC target("prfchw") pragma.
8178 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8180 * config/i386/i386.c (ix86_option_override_internal): Enable
8181 -mprfchw with -mprefetchwt1.
8183 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8185 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8188 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8190 * opts.h (CL_PCH_IGNORE): Define.
8191 * targhooks.c (option_affects_pch_p):
8192 Return false for options that have CL_PCH_IGNORE set.
8193 * opt-functions.awk: Process PchIgnore.
8194 * doc/options.texi: Document PchIgnore.
8196 * config/arc/arc.opt (misize): Add PchIgnore property.
8198 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8200 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8201 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8202 constraint on constants to permit them being loaded into
8203 GENERAL_REGS or BASE_REGS.
8205 2014-03-03 Nick Clifton <nickc@redhat.com>
8207 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8208 anti-cacnonical alternatives.
8209 (negandhi3_real): New pattern.
8210 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8212 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8214 * config/avr/avr-mcus.def: Remove atxmega16x1.
8215 * config/avr/avr-tables.opt: Regenerate.
8216 * config/avr/t-multilib: Regenerate.
8217 * doc/avr-mmcu.texi: Regenerate.
8219 2014-03-03 Tobias Grosser <tobias@grosser.es>
8220 Mircea Namolaru <mircea.namolaru@inria.fr>
8222 PR tree-optimization/58028
8223 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8226 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8228 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8229 not handled by recognizers.
8231 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8234 * function.c (expand_function_end): Don't emit
8235 clobber_return_register sequence if clobber_after is a BARRIER.
8236 * cfgexpand.c (construct_exit_block): Append instructions before
8237 return_label to prev_bb.
8239 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8241 * config/rs6000/constraints.md: Document reserved use of "wc".
8243 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8246 * ipa.c (function_and_variable_visibility): When dissolving comdat
8247 group, also set all symbols to local.
8249 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8254 2013-12-14 Jan Hubicka <jh@suse.cz>
8256 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8258 2014-03-02 Jon Beniston <jon@beniston.com>
8264 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8265 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8266 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8267 (simple_return, *simple_return): New patterns
8268 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8269 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8271 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8273 * dwarf2out.c (gen_subprogram_die): Tidy.
8275 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8278 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8279 (*mov_t_msb_neg_negc): ... this new insn.
8281 2014-02-28 Jason Merrill <jason@redhat.com>
8284 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8287 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8290 * dwarf2out.c (decltype_auto_die): New static.
8291 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8292 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8293 (is_cxx_auto): Likewise.
8295 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8297 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8298 we are not using general regs only.
8300 2014-02-28 Richard Biener <rguenther@suse.de>
8303 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8304 previous fix and only allow to remove trivial pre-headers
8305 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8306 (remove_forwarder_block): Properly update the latch of a loop.
8308 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8311 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8312 (cselib_preserved_hash_table): New.
8313 (preserve_constants_and_equivs): Move preserved vals to it.
8314 (cselib_find_slot): Look it up first.
8315 (cselib_init): Initialize it.
8316 (cselib_finish): Release it.
8317 (dump_cselib_table): Dump it.
8319 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8322 * cselib.c (remove_useless_values): Skip to avoid quadratic
8323 behavior if the condition moved from...
8324 (cselib_process_insn): ... here holds.
8326 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8329 * var-tracking.c (vt_initialize): Apply the same condition to
8330 preserve the CFA base value.
8332 2014-02-28 Joey Ye <joey.ye@arm.com>
8335 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8336 if reload in progress or completed.
8338 2014-02-28 Tobias Burnus <burnus@net-b.de>
8341 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8344 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8346 * doc/tm.texi.in (Condition Code Status): Update documention for
8347 relative locations of cc0-setter and cc0-user.
8349 2014-02-27 Jeff Law <law@redhat.com>
8351 PR rtl-optimization/52714
8352 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8353 into two independent simple sets, if I3 is a jump, ensure the
8354 pattern we place into I3 is a (set (pc) ...).
8356 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8357 Jeff Law <law@redhat.com>
8359 PR rtl-optimization/49847
8360 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8361 are in different blocks.
8362 * doc/tm.texi (Condition Code Status): Update documention for
8363 relative locations of cc0-setter and cc0-user.
8365 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8368 * lra.c (lra_emit_add): Check SUBREG too.
8370 2014-02-27 Andreas Schwab <schwab@suse.de>
8372 * config/m68k/m68k.c (m68k_option_override): Disable
8373 -flive-range-shrinkage for classic m68k.
8374 (m68k_override_options_after_change): Likewise.
8376 2014-02-27 Marek Polacek <polacek@redhat.com>
8379 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8380 -Wmaybe-uninitialized.
8382 2014-02-27 Alan Modra <amodra@gmail.com>
8385 * reload1.c (emit_input_reload_insns): When reload_override_in,
8386 set old to rl->in_reg when rl->in_reg is a subreg.
8388 2014-02-26 Richard Biener <rguenther@suse.de>
8391 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8393 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8395 * common/config/i386/predicates.md (const1256_operand): Remove.
8396 (const2356_operand): New.
8397 (const_1_to_2_operand): Remove.
8398 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8399 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8400 (*avx512pf_gatherpf<mode>sf): Ditto.
8401 (avx512pf_gatherpf<mode>df): Ditto.
8402 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8403 (*avx512pf_gatherpf<mode>df): Ditto.
8404 (avx512pf_scatterpf<mode>sf): Ditto.
8405 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8406 (*avx512pf_scatterpf<mode>sf): Ditto.
8407 (avx512pf_scatterpf<mode>df): Ditto.
8408 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8409 (*avx512pf_scatterpf<mode>df): Ditto.
8410 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8412 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8414 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8415 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8416 (_mm512_mask_testn_epi64_mask): Move to ...
8417 * config/i386/avx512cdintrin.h: Here.
8418 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8419 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8420 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8421 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8422 TARGET_AVX512F from TARGET_AVX512CD.
8424 2014-02-26 Richard Biener <rguenther@suse.de>
8427 * ipa.c (walk_polymorphic_call_targets): Properly guard
8428 call to inline_update_overall_summary.
8430 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8433 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8434 and latches only if requested. Fix latch if it is removed.
8435 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8436 LOOPS_HAVE_PREHEADERS.
8438 2014-02-25 Andrew Pinski <apinski@cavium.com>
8440 * builtins.c (expand_builtin_thread_pointer): Create a new target
8441 when the target is NULL.
8443 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8445 PR rtl-optimization/60317
8446 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8447 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8448 * lra-assigns.c: Include params.h.
8449 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8450 other reload pseudos considerations.
8452 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8454 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8455 to use canonical form for nor<mode>3.
8457 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8460 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8463 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8465 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8466 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8467 (ix86_handle_option): Handle OPT_mprefetchwt1.
8468 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8469 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8471 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8472 OPTION_MASK_ISA_PREFETCHWT1.
8473 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8474 (PTA_PREFETCHWT1): New.
8475 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8476 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8477 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8478 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8479 (*prefetch_avx512pf_<mode>_: Change into ...
8480 (*prefetch_prefetchwt1_<mode>: This.
8481 * config/i386/i386.opt (mprefetchwt1): New.
8482 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8483 (_mm_prefetch): Handle intent to write.
8484 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8486 2014-02-25 Richard Biener <rguenther@suse.de>
8489 * emit-rtl.c (mem_attrs_htab): Remove.
8490 (mem_attrs_htab_hash): Likewise.
8491 (mem_attrs_htab_eq): Likewise.
8492 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8493 (init_emit_once): Do not allocate mem_attrs_htab.
8495 2014-02-25 Richard Biener <rguenther@suse.de>
8498 * lto-opts.c (lto_write_options): Output non-explicit conservative
8499 -fwrapv, -fno-trapv and -fno-strict-overflow.
8500 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8501 (run_gcc): And pass them through.
8503 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8505 * sel-sched.c (calculate_new_fences): New parameter ptime.
8506 Calculate it as a maximum over all fence cycles.
8507 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8508 Print the final schedule timing when sched_verbose.
8510 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8512 PR rtl-optimization/60292
8513 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8514 bit fot the fence instruction.
8516 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8518 * calls.h: Fix typo in comment.
8520 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8522 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8523 adjusting offsetable addresses.
8525 2014-02-24 Guozhi Wei <carrot@google.com>
8527 * sparseset.h (sparseset_pop): Fix the wrong index.
8529 2014-02-24 Walter Lee <walt@tilera.com>
8531 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8532 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8534 * common/config/tilegx/tilegx-common.c
8535 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8536 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8538 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8539 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8540 (tilegx_gimplify_va_arg_expr): Handle big endian.
8541 (tilegx_expand_unaligned_load): Ditto.
8542 (tilegx_expand_unaligned_store): Ditto.
8543 (TARGET_RETURN_IN_MSB): New.
8544 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8545 (TARGET_ENDIAN_DEFAULT): New.
8546 (TARGET_BIG_ENDIAN): Handle big endian.
8547 (BYTES_BIG_ENDIAN): Ditto.
8548 (WORDS_BIG_ENDIAN): Ditto.
8549 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8552 * config/tilegx/tilegx.md (extv): Handle big endian.
8554 (insn_st<n>): Ditto.
8555 (insn_st<n>_add<bitsuffix>): Ditto.
8556 (insn_stnt<n>): Ditto.
8557 (insn_stnt<n>_add<bitsuffix>):Ditto.
8558 (vec_interleave_highv8qi): Handle big endian.
8559 (vec_interleave_highv8qi_be): New.
8560 (vec_interleave_highv8qi_le): New.
8561 (insn_v1int_h): Handle big endian.
8562 (vec_interleave_lowv8qi): Handle big endian.
8563 (vec_interleave_lowv8qi_be): New.
8564 (vec_interleave_lowv8qi_le): New.
8565 (insn_v1int_l): Handle big endian.
8566 (vec_interleave_highv4hi): Handle big endian.
8567 (vec_interleave_highv4hi_be): New.
8568 (vec_interleave_highv4hi_le): New.
8569 (insn_v2int_h): Handle big endian.
8570 (vec_interleave_lowv4hi): Handle big endian.
8571 (vec_interleave_lowv4hi_be): New.
8572 (vec_interleave_lowv4hi_le): New.
8573 (insn_v2int_l): Handle big endian.
8574 (vec_interleave_highv2si): Handle big endian.
8575 (vec_interleave_highv2si_be): New.
8576 (vec_interleave_highv2si_le): New.
8577 (insn_v4int_h): Handle big endian.
8578 (vec_interleave_lowv2si): Handle big endian.
8579 (vec_interleave_lowv2si_be): New.
8580 (vec_interleave_lowv2si_le): New.
8581 (insn_v4int_l): Handle big endian.
8582 * config/tilegx/tilegx.opt (mbig-endian): New option.
8583 (mlittle-endian): New option.
8584 * doc/install.texi: Document tilegxbe-linux.
8585 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8587 2014-02-24 Martin Jambor <mjambor@suse.cz>
8590 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8591 there are no parameter descriptors.
8593 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8595 PR rtl-optimization/60268
8596 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8597 initialization to ...
8598 (sched_rgn_init): ... here.
8599 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8601 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8603 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8606 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8608 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8611 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8613 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8614 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8616 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8618 * config/microblaze/predicates.md: Add cmp_op predicate.
8619 * config/microblaze/microblaze.md: Add branch_compare instruction
8620 which uses cmp_op predicate and emits cmp insn before branch.
8621 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8622 to microblaze_expand_conditional_branch and consolidate logic.
8623 (microblaze_expand_conditional_branch): emit branch_compare
8624 insn instead of handling cmp op separate from branch insn.
8626 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8628 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8631 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8633 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8634 define_insn with define_expand and new define_insn
8635 *altivec_lve<VI_char>x_internal.
8636 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8637 and new define_insn *altivec_stve<VI_char>x_internal.
8638 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8640 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8642 (altivec_expand_stvex_be): New function.
8644 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8646 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8647 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8648 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8649 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8651 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8654 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8655 instead of emit_move_insn.
8657 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8659 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8661 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8662 gen_altivec_vsumsws.
8664 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8666 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8667 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8668 (altivec_lvxl_<mode>): New define_expand incorporating
8669 -maltivec=be semantics where needed.
8670 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8671 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8672 semantics where needed.
8673 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8674 (altivec_stvx_<mode>): New define_expand incorporating
8675 -maltivec=be semantics where needed.
8676 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8677 VM2 iterator instead of V4SI.
8678 (altivec_stvxl_<mode>): New define_expand incorporating
8679 -maltivec=be semantics where needed.
8680 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8681 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8682 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8683 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8684 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8685 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8686 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8687 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8688 ALTIVEC_BUILTIN_STVXL.
8689 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8690 (altivec_expand_stvx_be): Likewise.
8691 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8692 (altivec_expand_lvx_be): Likewise.
8693 (altivec_expand_stvx_be): Likewise.
8694 (altivec_expand_builtin): Add cases for
8695 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8696 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8697 (altivec_init_builtins): Add definitions for
8698 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8699 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8701 2014-02-21 Catherine Moore <clm@codesourcery.com>
8703 * doc/invoke.texi (mvirt, mno-virt): Document.
8704 * config/mips/mips.opt (mvirt): New option.
8705 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8707 2014-02-21 Richard Biener <rguenther@suse.de>
8709 PR tree-optimization/60276
8710 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8711 (STMT_VINFO_MIN_NEG_DIST): New macro.
8712 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8713 STMT_VINFO_MIN_NEG_DIST.
8714 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8715 made for negative dependence distances still hold.
8717 2014-02-21 Richard Biener <rguenther@suse.de>
8720 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8721 DECL_INITIAL for globals not in the current function context.
8723 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8725 PR tree-optimization/56490
8726 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8727 * tree-ssa-uninit.c: Include params.h.
8728 (compute_control_dep_chain): Add num_calls argument, return false
8729 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8730 num_calls to recursive call.
8731 (find_predicates): Change dep_chain into normal array,
8732 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8733 variable and adjust compute_control_dep_chain caller.
8734 (find_def_preds): Likewise.
8736 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8738 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8739 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8741 2014-02-21 Nick Clifton <nickc@redhat.com>
8743 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8744 (pushhi1): Likewise.
8745 (popqi1): Add mode to pre_dec.
8748 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8750 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8751 mode for mask of V8SFmode permutation.
8753 2014-02-20 Richard Henderson <rth@redhat.com>
8756 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8757 a new pseudo for OLDVAL.
8759 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8762 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8763 gen_reg_rtx if d->testing_p.
8764 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8765 if d->testing_p and we will certainly return true.
8766 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8769 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8771 * emit-rtl.c (gen_reg_rtx): Assert that
8772 crtl->emit.regno_pointer_align_length is non-zero.
8774 2014-02-20 Richard Henderson <rth@redhat.com>
8777 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8778 on failure the store back into EXPECT.
8780 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8781 Sandra Loosemore <sandra@codesourcery.com>
8783 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8784 * config/nios2/nios2.c (nios2_function_profiler): Add
8785 -fPIC (flag_pic == 2) support.
8786 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8787 (nios2_large_offset_p): New function.
8788 (nios2_unspec_reloc_p): Move up position, update to use
8789 nios2_large_offset_p.
8790 (nios2_unspec_address): Remove function.
8791 (nios2_unspec_offset): New function.
8792 (nios2_large_got_address): New function.
8793 (nios2_got_address): Add large offset support.
8794 (nios2_legitimize_tls_address): Update usage of removed and new
8796 (nios2_symbol_binds_local_p): New function.
8797 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8798 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8799 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8800 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8801 processing for (const (unspec ...)).
8802 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8804 2014-02-20 Richard Biener <rguenther@suse.de>
8806 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8807 doing the substitution.
8808 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8810 2014-02-20 Martin Jambor <mjambor@suse.cz>
8813 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8814 info when checking whether lattices are bottom.
8816 2014-02-20 Richard Biener <rguenther@suse.de>
8819 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8822 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8825 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8826 parameter specifying the scaling.
8827 (inline_call): Update.
8828 (want_inline_recursively): Guard division by zero.
8829 (recursive_inlining): Update.
8830 * ipa-inline.h (clone_inlined_nodes): Update.
8832 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8835 * config/i386/i386.c (classify_argument): Pass structures of size
8836 64 bytes or less in register.
8838 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8839 Kirill Yukhin <kirill.yukhin@intel.com>
8841 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8842 (_mm_rcp28_round_ss): Ditto.
8843 (_mm_rsqrt28_round_sd): Ditto.
8844 (_mm_rsqrt28_round_ss): Ditto.
8845 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8846 (_mm_rcp14_round_ss): Ditto.
8847 (_mm_rsqrt14_round_sd): Ditto.
8848 (_mm_rsqrt14_round_ss): Ditto.
8849 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8850 the first input operand, get rid of match_dup.
8851 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8853 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8855 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8856 operand as the first input operand, set type attribute.
8857 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8859 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8860 operand as the first input operand, set type attribute.
8862 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8864 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8867 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8870 * config/i386/i386.c (construct_container): Remove TFmode check
8871 for X86_64_INTEGER_CLASS.
8873 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
8876 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
8877 only when -Wpsabi is enabled.
8879 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
8882 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
8883 passing arrays in registers are the same as for structs, so remove the
8884 special case for them.
8886 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
8888 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
8889 destination type, extract only the valid bits if the source type is not
8890 integral and has a different mode.
8892 2014-02-19 Richard Biener <rguenther@suse.de>
8895 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
8898 2014-02-19 Richard Biener <rguenther@suse.de>
8901 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
8902 (ipa_modify_call_arguments): Emit an argument load explicitely and
8903 preserve virtual SSA form there and for the replacement call.
8904 Do not update SSA form nor free dominance info.
8906 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8908 * ipa.c (function_and_variable_visibility): Also clear WEAK
8909 flag when disolving COMDAT_GROUP.
8911 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8913 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
8914 * ipa-prop.c (ipa_set_jf_known_type): Return early when
8916 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
8917 do more sanity checks.
8918 (detect_type_change): Return true when giving up early.
8919 (compute_complex_assign_jump_func): Fix type parameter of
8920 ipa_set_ancestor_jf.
8921 (compute_complex_ancestor_jump_func): Likewise.
8922 (update_jump_functions_after_inlining): Fix updating of
8924 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
8926 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
8928 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
8929 inline clones when edge disappears.
8931 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8934 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
8935 Split 64-bit moves into 2 patterns. Do not allow the use of
8936 direct move for TDmode in little endian, since the decimal value
8937 has little endian bytes within a word, but the 64-bit pieces are
8938 ordered in a big endian fashion, and normal subreg's of TDmode are
8940 (mov<mode>_64bit_dm): Likewise.
8941 (movtd_64bit_nodm): Likewise.
8943 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8945 PR tree-optimization/60174
8946 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
8947 statement of an SSA_NAME that occurs in an abnormal PHI node.
8949 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8952 * final.c (SEEN_BB): Remove.
8953 (SEEN_NOTE, SEEN_EMITTED): Renumber.
8954 (final_scan_insn): Don't force_source_line on second
8955 NOTE_INSN_BASIC_BLOCK.
8957 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
8960 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
8961 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
8962 (type_natural_mode): Warn ABI change when %zmm register is not
8963 available for AVX512F vector value passing.
8965 2014-02-18 Kai Tietz <ktietz@redhat.com>
8968 * config/i386/i386.c (ix86_expand_prologue): Use value in
8969 rax register as displacement when restoring %r10 or %rax.
8970 Fix wrong offset when restoring both registers.
8972 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
8974 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
8975 assertion with conditional return.
8977 2014-02-18 Jakub Jelinek <jakub@redhat.com>
8978 Uros Bizjak <ubizjak@gmail.com>
8981 * config/i386/driver-i386.c (host_detect_local_cpu): If
8982 YMM state is not saved by the OS, also clear has_f16c. Move
8983 CPUID 0x80000001 handling before YMM state saving checking.
8985 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
8987 PR rtl-optimization/58960
8988 * haifa-sched.c (alloc_global_sched_pressure_data): New,
8989 factored out from ...
8990 (sched_init): ... here.
8991 (free_global_sched_pressure_data): New, factored out from ...
8992 (sched_finish): ... here.
8993 * sched-int.h (free_global_sched_pressure_data): Declare.
8994 * sched-rgn.c (nr_regions_initial): New static global.
8995 (haifa_find_rgns): Initialize it.
8996 (schedule_region): Disable sched-pressure for the newly
8999 2014-02-17 Richard Biener <rguenther@suse.de>
9001 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9002 release SSA defs of pattern stmts.
9004 2014-02-17 Richard Biener <rguenther@suse.de>
9006 * tree-inline.c (expand_call_inline): Release the virtual
9007 operand defined by the call we are about to inline.
9009 2014-02-17 Richard Biener <rguenther@suse.de>
9011 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9013 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9014 Ilya Tocar <ilya.tocar@intel.com>
9016 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9017 arguments order in builtin.
9018 (_mm512_permutexvar_epi64): Ditto.
9019 (_mm512_mask_permutexvar_epi64): Ditto
9020 (_mm512_maskz_permutexvar_epi32): Ditto
9021 (_mm512_permutexvar_epi32): Ditto
9022 (_mm512_mask_permutexvar_epi32): Ditto
9024 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9026 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9027 (p8_vmrgow): Likewise.
9029 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9031 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9034 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9037 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9038 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9039 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9040 using direct move instructions on ISA 2.07. Also adjust
9041 instruction length for 64-bit.
9042 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9043 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9045 2014-02-15 Alan Modra <amodra@gmail.com>
9049 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9050 find_replacement on parts of insn rtl that might be reloaded.
9052 2014-02-15 Richard Biener <rguenther@suse.de>
9054 PR tree-optimization/60183
9055 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9056 (tree_ssa_phiprop): Calculate and free post-dominators.
9058 2014-02-14 Jeff Law <law@redhat.com>
9060 PR rtl-optimization/60131
9061 * ree.c (get_extended_src_reg): New function.
9062 (combine_reaching_defs): Use it rather than assuming location of REG.
9063 (find_and_remove_re): Verify first operand of extension is
9064 a REG before adding the insns to the copy list.
9066 2014-02-14 Roland McGrath <mcgrathr@google.com>
9068 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9069 * configure: Regenerated.
9070 * config.in: Regenerated.
9071 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9072 instead of ASM_SHORT.
9074 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9075 Richard Earnshaw <rearnsha@arm.com>
9077 PR rtl-optimization/59535
9078 * lra-constraints.c (process_alt_operands): Encourage alternative
9079 when unassigned pseudo class is superset of the alternative class.
9080 (inherit_reload_reg): Don't inherit when optimizing for code size.
9081 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9082 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9083 modes not less than 4 for Thumb1.
9085 2014-02-14 Kyle McMartin <kyle@redhat.com>
9088 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9090 2014-02-14 Richard Biener <rguenther@suse.de>
9092 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9093 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9094 Do not strip INDIRECT_REFs.
9096 2014-02-14 Richard Biener <rguenther@suse.de>
9099 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9100 DECL_FUNCTION_SPECIFIC_TARGET.
9101 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9102 * tree-streamer-out.c (pack_ts_target_option): Remove.
9103 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9104 (write_ts_function_decl_tree_pointers): Do not stream
9105 DECL_FUNCTION_SPECIFIC_TARGET.
9106 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9107 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9108 (lto_input_ts_function_decl_tree_pointers): Do not stream
9109 DECL_FUNCTION_SPECIFIC_TARGET.
9111 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9113 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9114 (get_initial_def_for_induction, vectorizable_induction): Ignore
9115 debug stmts when looking for exit_phi.
9116 (vectorizable_live_operation): Fix up condition.
9118 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9120 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9121 nreverse() because it changes the content of original tree list.
9123 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9125 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9126 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9128 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9130 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9131 GNU coding standards.
9133 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9136 * dwarf2out.c (gen_subprogram_die): Don't call
9137 add_calling_convention_attribute if subr_die is old_die.
9139 2014-02-13 Sharad Singhai <singhai@google.com>
9141 * doc/optinfo.texi: Fix order of nodes.
9143 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9145 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9146 operands[2], not operands[3].
9148 2014-02-13 Richard Biener <rguenther@suse.de>
9151 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9152 mention the possibility of an in-tree build.
9153 (CLooG): Update recommended version to 0.18.1, mention the
9154 possibility of an in-tree build and clarify that the ISL
9155 bundled with CLooG does not work.
9157 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9160 * expr.c (compress_float_constant): If x is a hard register,
9161 extend into a pseudo and then move to x.
9163 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9165 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9166 caused by bad second argument to warning_at() with -mhotpatch and
9167 nested functions (e.g. with gfortran).
9169 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9171 * opts.c (option_name): Remove "enabled by default" rider.
9173 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9175 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9177 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9178 Uros Bizjak <ubizjak@gmail.com>
9181 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9182 * configure: Regenerated.
9184 2014-02-12 Richard Biener <rguenther@suse.de>
9186 * vec.c (vec_prefix::calculate_allocation): Move as
9187 inline variant to vec.h.
9188 (vec_prefix::calculate_allocation_1): New out-of-line version.
9189 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9190 (vec_prefix::m_has_auto_buf): Rename to ...
9191 (vec_prefix::m_using_auto_storage): ... this.
9192 (vec_prefix::calculate_allocation): Inline the easy cases
9193 and dispatch to calculate_allocation_1 which doesn't need the
9195 (va_heap::reserve): Use gcc_checking_assert.
9196 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9197 m_using_auto_storage.
9198 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9200 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9201 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9202 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9204 2014-02-12 Richard Biener <rguenther@suse.de>
9206 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9207 when we found a dependence.
9209 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9211 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9213 (maybe_fold_stmt): ... into this new function.
9214 * omp-low.c (lower_omp): Update comment.
9216 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9219 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9222 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9224 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9225 identifiers in comments.
9226 (cortexa53_extra_costs): Likewise.
9227 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9228 (cortexa7_extra_costs): Likewise.
9229 (cortexa12_extra_costs): Likewise.
9230 (cortexa15_extra_costs): Likewise.
9231 (v7m_extra_costs): Likewise.
9233 2014-02-12 Richard Biener <rguenther@suse.de>
9236 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9237 of posix_memalign being successful.
9238 (lower_stmt): Restrict lowering of posix_memalign to when
9239 -ftree-bit-ccp is enabled.
9241 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9243 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9245 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9247 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9249 PR rtl-optimization/60116
9250 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9251 other_insn once the combination has been validated.
9253 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9256 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9258 * ipa-devirt.c: Include demangle.h
9259 (odr_violation_reported): New static variable.
9260 (add_type_duplicate): Update odr_violations.
9261 (maybe_record_node): Add completep parameter; update it.
9262 (record_target_from_binfo): Add COMPLETEP parameter;
9263 update it as needed.
9264 (possible_polymorphic_call_targets_1): Likewise.
9265 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9266 rename FINAL to COMPLETE.
9267 (record_targets_from_bases): Sanity check we found the binfo;
9268 fix COMPLETEP updating.
9269 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9270 parameter, fix computing of COMPLETEP.
9271 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9272 at LTO time do demangling.
9273 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9274 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9276 (gimple_get_virt_method_for_binfo): Likewise.
9277 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9278 gimple_get_virt_method_for_vtable): Update prototypes.
9280 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9283 * genautomata.c (add_presence_absence): Fix typo with
9284 {final_}presence_list.
9286 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9289 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9290 for VSX/Altivec vectors that land in GPR registers.
9292 2014-02-11 Richard Henderson <rth@redhat.com>
9293 Jakub Jelinek <jakub@redhat.com>
9296 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9297 around drhs if type conversion to lacc->type is not useless.
9299 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9301 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9303 (cortex-a57.cortex-a53): Likewise.
9304 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9306 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9308 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9311 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9313 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9314 add_options_for_arm_vfp3.
9316 2014-02-11 Jeff Law <law@redhat.com>
9319 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9320 object with an undesirable mode.
9322 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9325 * config/i386/sol2-9.h: New file.
9326 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9327 *-*-solaris2.9*): Use it.
9329 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9331 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9332 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9334 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9336 * config/microblaze/microblaze.c: Extend mcpu version format
9338 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9340 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9342 2014-02-10 Richard Henderson <rth@redhat.com>
9345 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9346 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9347 ms-abi vs -mno-accumulate-outgoing-args.
9348 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9349 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9352 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9355 * cfgexpand.c (expand_asm_operands): Attach source location to
9356 ASM_INPUT rtx objects.
9357 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9359 2014-02-10 Nick Clifton <nickc@redhat.com>
9361 * config/mn10300/mn10300.c (popcount): New function.
9362 (mn10300_expand_prologue): Include saved registers in stack usage
9365 2014-02-10 Jeff Law <law@redhat.com>
9368 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9369 when changing the SET_DEST of a prior insn to avoid an input reload.
9371 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9373 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9374 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9375 -mcall-openbsd, or -mcall-linux.
9376 (CC1_ENDIAN_BIG_SPEC): Remove.
9377 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9378 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9379 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9380 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9381 and %cc1_endian_default.
9382 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9384 2014-02-10 Richard Biener <rguenther@suse.de>
9386 PR tree-optimization/60115
9387 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9388 MEM_REF handling. Properly verify that the accesses are not
9389 out of the objects bound.
9391 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9393 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9396 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9398 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9399 proper constants and fix formatting.
9400 (possible_polymorphic_call_targets): Fix formatting.
9402 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9403 Ilya Tocar <ilya.tocar@intel.com>
9405 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9406 (_mm512_loadu_epi32): Renamed into...
9407 (_mm512_loadu_si512): This.
9408 (_mm512_storeu_epi32): Renamed into...
9409 (_mm512_storeu_si512): This.
9410 (_mm512_maskz_ceil_ps): Removed.
9411 (_mm512_maskz_ceil_pd): Ditto.
9412 (_mm512_maskz_floor_ps): Ditto.
9413 (_mm512_maskz_floor_pd): Ditto.
9414 (_mm512_floor_round_ps): Ditto.
9415 (_mm512_floor_round_pd): Ditto.
9416 (_mm512_ceil_round_ps): Ditto.
9417 (_mm512_ceil_round_pd): Ditto.
9418 (_mm512_mask_floor_round_ps): Ditto.
9419 (_mm512_mask_floor_round_pd): Ditto.
9420 (_mm512_mask_ceil_round_ps): Ditto.
9421 (_mm512_mask_ceil_round_pd): Ditto.
9422 (_mm512_maskz_floor_round_ps): Ditto.
9423 (_mm512_maskz_floor_round_pd): Ditto.
9424 (_mm512_maskz_ceil_round_ps): Ditto.
9425 (_mm512_maskz_ceil_round_pd): Ditto.
9426 (_mm512_expand_pd): Ditto.
9427 (_mm512_expand_ps): Ditto.
9428 * config/i386/i386.c (ix86_builtins): Remove
9429 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9430 (bdesc_args): Ditto.
9431 * config/i386/predicates.md (const1256_operand): New.
9432 (const_1_to_2_operand): Ditto.
9433 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9434 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9435 (*avx512pf_gatherpf<mode>sf): Ditto.
9436 (avx512pf_gatherpf<mode>df): Ditto.
9437 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9438 (*avx512pf_gatherpf<mode>df): Ditto.
9439 (avx512pf_scatterpf<mode>sf): Ditto.
9440 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9441 (*avx512pf_scatterpf<mode>sf): Ditto.
9442 (avx512pf_scatterpf<mode>df): Ditto.
9443 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9444 (*avx512pf_scatterpf<mode>df): Ditto.
9445 (avx512f_expand<mode>): Removed.
9446 (<shift_insn><mode>3<mask_name>): Change predicate type.
9448 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9450 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9451 not at the end of datarefs vector use ordered_remove to avoid
9452 reordering datarefs vector.
9455 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9456 mark local addressable non-static vars as GOVD_PRIVATE
9457 instead of GOVD_LOCAL.
9458 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9459 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9463 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9464 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9465 assume_aligned or alloc_align attributes.
9466 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9467 arguments. Handle also assume_aligned and alloc_align attributes.
9468 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9469 calls to functions with assume_aligned or alloc_align attributes.
9470 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9472 2014-02-08 Terry Guo <terry.guo@arm.com>
9474 * doc/invoke.texi: Document ARM -march=armv7e-m.
9476 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9478 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9479 flag on __cilkrts_rethrow builtin.
9482 * ipa-cp.c (determine_versionability): Fail at -O0
9483 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9484 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9487 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9490 * tree-inline.c (copy_forbidden): Fail for
9491 __attribute__((optimize (0))) functions.
9493 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9495 * varpool.c: Include pointer-set.h.
9496 (varpool_remove_unreferenced_decls): Variables in other partitions
9497 will not be output; be however careful to not lose information
9500 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9502 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9503 lookup in the vtable constructor.
9505 2014-02-07 Jeff Law <law@redhat.com>
9508 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9509 define_insn_and_split.
9511 * ipa-inline.c (inline_small_functions): Fix typos.
9513 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9515 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9516 (s390_can_use_return_insn): Declare.
9517 * config/s390/s390.h (EPILOGUE_USES): Define.
9518 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9520 (s390_chunkify_start): Handle return JUMP_LABELs.
9521 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9522 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9523 (s390_can_use_return_insn): New functions.
9524 (s390_fix_long_loop_prediction): Handle conditional returns.
9525 (TARGET_SET_UP_BY_PROLOGUE): Define.
9526 * config/s390/s390.md (ANY_RETURN): New code iterator.
9527 (*creturn, *csimple_return, return, simple_return): New patterns.
9529 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9531 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9532 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9533 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9534 REG_CFA_RESTORE list when deciding not to restore a register.
9536 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9538 * config/s390/s390.c: Include tree-pass.h and context.h.
9539 (s390_early_mach): New function, split out from...
9540 (s390_emit_prologue): ...here.
9541 (pass_data_s390_early_mach): New pass structure.
9542 (pass_s390_early_mach): New class.
9543 (s390_option_override): Create and register early_mach pass.
9544 Move to end of file.
9546 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9548 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9549 to match for the exit block.
9551 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9553 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9554 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9555 Reject misaligned operands.
9557 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9559 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9561 2014-02-07 Richard Biener <rguenther@suse.de>
9564 * gimple-low.c (lower_builtin_posix_memalign): New function.
9565 (lower_stmt): Call it to lower posix_memalign in a way
9566 to make alignment info accessible.
9568 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9571 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9572 __builtin_setjmp_receiver.
9574 2014-02-07 Richard Biener <rguenther@suse.de>
9577 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9578 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9579 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9580 Handle BUILT_IN_POSIX_MEMALIGN.
9581 (find_func_clobbers): Likewise.
9582 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9583 (call_may_clobber_ref_p_1): Likewise.
9585 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9588 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9591 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9594 * lto-cgraph.c (lto_output_node): Use
9595 symtab_get_symbol_partitioning_class.
9596 (lto_output_varpool_node): likewise.
9597 (symtab_get_symbol_partitioning_class): Move here from
9599 * cgraph.h (symbol_partitioning_class): Likewise.
9600 (symtab_get_symbol_partitioning_class): Declare.
9602 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9604 * ggc.h (ggc_internal_cleared_alloc): New macro.
9605 * vec.h (vec_safe_copy): Handle memory stats.
9606 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9607 * target-globals.c (save_target_globals): Likewise.
9609 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9612 * expr.c (emit_move_resolve_push): Export; be bit more selective
9613 on when to clear alias set.
9614 * expr.h (emit_move_resolve_push): Declare.
9615 * function.h (struct function): Add tail_call_marked.
9616 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9617 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9618 * config/i386/i386.md (TImode move expander): De not call
9620 (FP push expanders): Preserve memory attributes.
9621 * config/i386/sse.md (push<mode>1): Remove.
9622 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9623 (ix86_expand_push): Remove.
9624 * config/i386/mmx.md (push<mode>1): Remove.
9626 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9628 PR rtl-optimization/60030
9629 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9630 lopart with paradoxical subreg before shifting it up by hprec.
9632 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9634 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9635 Remove extra newline at end of file.
9636 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9637 (arm_issue_rate): Handle cortexa57.
9638 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9639 (cortex-a57.cortex-a53): Likewise.
9641 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9644 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9645 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9647 (arm_expand_prologue): Adjust all callers.
9648 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9649 info, registers also at the lowest numbered registers side. Use
9650 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9654 * var-tracking.c (adjust_mems): Before adding a SET to
9655 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9657 2014-02-06 Alan Modra <amodra@gmail.com>
9660 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9661 change SDmode to DDmode when lra_in_progress.
9663 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9666 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9667 free_data_ref on the dr first, and before goto again also set dr
9668 to the next dr. For simd_lane_access, free old datarefs[i] before
9669 overwriting it. For get_vectype_for_scalar_type failure, don't
9670 free_data_ref if simd_lane_access.
9672 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9675 * tree.h (opts_for_fn): New inline function.
9676 (opt_for_fn): Define.
9677 * config/i386/i386.c (ix86_function_regparm): Use
9678 opt_for_fn (decl, optimize) instead of optimize.
9680 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9682 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9683 for SYMBOL_REF in large memory model.
9685 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9687 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9689 (cortex-a57): Likewise.
9690 (cortex-a57.cortex-a53): Likewise.
9692 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9693 Kugan Vivekanandarajah <kuganv@linaro.org>
9695 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9697 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9699 2014-02-06 Richard Biener <rguenther@suse.de>
9701 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9702 set_loop_copy and initialize_original_copy_tables.
9704 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9706 * config/aarch64/aarch64-simd.md
9707 (aarch64_ashr_simddi): Change QI to SI.
9709 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9710 Jakub Jelinek <jakub@redhat.com>
9713 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9716 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9718 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9719 CODE_FOR_altivec_vpku[hw]um to
9720 CODE_FOR_altivec_vpku[hw]um_direct.
9721 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9722 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9723 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9724 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9726 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9728 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9729 generation for -maltivec=be.
9730 (altivec_vsumsws): Simplify redundant test.
9732 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9734 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9735 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9736 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9737 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9738 gen_altivec_vpkuwum.
9739 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9741 (altivec_vpks<VI_char>ss): Likewise.
9742 (altivec_vpks<VI_char>us): Likewise.
9743 (altivec_vpku<VI_char>us): Likewise.
9744 (altivec_vpku<VI_char>um): Likewise.
9745 (altivec_vpku<VI_char>um_direct): New (copy of
9746 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9748 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9749 target is little endian and -maltivec=be is not specified.
9750 (*altivec_vupkhs<VU_char>_direct): New (copy of
9751 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9752 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9753 target is little endian and -maltivec=be is not specified.
9754 (*altivec_vupkls<VU_char>_direct): New (copy of
9755 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9756 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9757 little endian and -maltivec=be is not specified.
9758 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9759 little endian and -maltivec=be is not specified.
9761 2014-02-05 Richard Henderson <rth@redhat.com>
9764 * combine-stack-adj.c: Revert r206943.
9765 * sched-int.h (struct deps_desc): Add last_args_size.
9766 * sched-deps.c (init_deps): Initialize it.
9767 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9768 contain REG_ARGS_SIZE notes.
9770 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9772 * lto-cgraph.c (asm_nodes_output): Make global.
9773 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9774 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9775 (driver_handle_option): Handle OPT_fwpa.
9777 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9780 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9781 a comment typo and formatting issue. If odr_hash hasn't been
9782 created, return vNULL and set *completep to false.
9785 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9786 bb with no successors.
9788 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9791 * doc/invoke.texi (-march): Clarify documentation for ARM.
9795 2014-02-05 Richard Biener <rguenther@suse.de>
9797 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9798 when not vectorizing because of too many alias checks.
9799 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9800 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9802 2014-02-05 Nick Clifton <nickc@redhat.com>
9804 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9805 accept extended registers in any mode when compiling for the MN10300.
9807 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9809 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9810 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9811 sanitization attributes.
9812 (can_inline_edge_p): Likewise.
9813 (sanitize_attrs_match_for_inline_p): New function.
9815 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9817 * ipa-prop.c (detect_type_change): Shor circuit testing of
9818 type changes on THIS pointer.
9820 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9823 * config/pa/pa.c (legitimize_tls_address): Return original address
9824 if not passed a SYMBOL_REF rtx.
9825 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9827 (pa_emit_move_sequence): Simplify TLS source operands.
9828 (pa_legitimate_constant_p): Reject all TLS constants.
9829 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9830 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9832 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9834 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9835 groups when we know they are controlled by LTO.
9836 * varasm.c (default_binds_local_p_1): If object is in other partition,
9837 it will be resolved locally.
9839 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9841 * config/host-linux.c (linux_gt_pch_use_address): Don't
9842 use SSIZE_MAX because it is not always defined.
9844 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9847 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9848 threshold for pseudo splitting.
9849 (update_ebb_live_info): Process call argument hard registers and
9850 hard registers from insn definition too.
9851 (max_small_class_regs_num): New constant.
9852 (inherit_in_ebb): Update live hard regs through EBBs. Update
9853 reloads_num only for small register classes. Don't split for
9856 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9859 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9862 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9864 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9867 2014-02-04 Marek Polacek <polacek@redhat.com>
9869 * gdbinit.in (pel): Define.
9871 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9873 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
9876 2014-02-04 Richard Biener <rguenther@suse.de>
9879 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
9880 in function context local.
9881 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
9882 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
9883 similar to LTO_imported_decl_ref.
9885 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9887 PR tree-optimization/60002
9888 * cgraphclones.c (build_function_decl_skip_args): Clear
9891 PR tree-optimization/60023
9892 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
9893 false to gsi_replace.
9894 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
9895 has been in some EH region and vec_stmt could throw, add
9896 vec_stmt into the same EH region.
9897 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
9898 has no lhs, ignore it.
9899 * internal-fn.c (expand_MASK_LOAD): Likewise.
9902 * tree-inline.c (copy_forbidden): Fail for
9903 __attribute__((optimize (0))) functions.
9906 * omp-low.c (simd_clone_struct_copy): If from->inbranch
9907 is set, copy one less argument.
9908 (expand_simd_clones): Don't subtract clone_info->inbranch
9909 from simd_clone_struct_alloc argument.
9911 PR rtl-optimization/57915
9912 * recog.c (simplify_while_replacing): If all unary/binary/relational
9913 operation arguments are constant, attempt to simplify those.
9916 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
9917 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
9919 2014-02-04 Richard Biener <rguenther@suse.de>
9921 PR tree-optimization/60012
9922 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
9923 TBAA disambiguation to all DDRs.
9925 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9928 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
9929 (LINK_SPEC): Use it for -shared, -shared-libgcc.
9931 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9934 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
9936 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9938 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
9939 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
9941 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9944 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
9945 to figure out targets of polymorphic calls with known decl.
9946 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9947 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
9948 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
9949 (get_polymorphic_call_info): ... here.
9950 (get_polymorphic_call_info_from_invariant): New function.
9952 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9954 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
9955 lookup via vtable pointer; check for type consistency
9956 and turn inconsitent facts into UNREACHABLE.
9957 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9958 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
9959 type inconsistent querries; return UNREACHABLE instead.
9961 2014-02-03 Richard Henderson <rth@twiddle.net>
9964 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
9965 already processed this node.
9966 (normalize_one_pred_1): Pass along mark_set.
9967 (normalize_one_pred): Create and destroy a pointer_set_t.
9968 (normalize_one_pred_chain): Likewise.
9970 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
9972 PR gcov-profile/58602
9973 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
9975 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
9978 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
9979 -fno-devirtualize; try to devirtualize by the knowledge of
9980 virtual table pointer given by aggregate propagation.
9981 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
9982 (ipa_print_node_jump_functions): Dump also offset that
9983 is relevant for polymorphic calls.
9984 (determine_known_aggregate_parts): Add arg_type parameter; use it
9985 instead of determining the type from pointer type.
9986 (ipa_compute_jump_functions_for_edge): Update call of
9987 determine_known_aggregate_parts.
9988 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
9989 (gimple_get_virt_method_for_binfo): ... here; simplify using
9990 vtable_pointer_value_to_vtable.
9991 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
9992 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
9993 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
9994 (vtable_pointer_value_to_vtable): Break out from ...; handle also
9996 (vtable_pointer_value_to_binfo): ... here.
9997 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
9999 2014-02-03 Teresa Johnson <tejohnson@google.com>
10001 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10002 redef of outer loop index variable.
10004 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10008 * doc/extend.texi (Function Attributes): Typo.
10010 2014-02-03 Cong Hou <congh@google.com>
10012 PR tree-optimization/60000
10013 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10014 if the vectorized statement is a store. A store statement can only
10015 appear at the end of pattern statements.
10017 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10019 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10020 (ix86_option_override_internal): Default long double to 64-bit for
10021 32-bit Bionic and to 128-bit for 64-bit Bionic.
10023 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10024 TARGET_LONG_DOUBLE_128 is true.
10025 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10027 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10028 (mlong-double-64): Negate -mlong-double-128.
10029 (mlong-double-128): New option.
10031 * config/i386/i386-c.c (ix86_target_macros): Define
10032 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10034 * doc/invoke.texi: Document -mlong-double-128.
10036 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10038 PR rtl-optimization/60024
10039 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10041 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10043 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10045 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10047 PR rtl-optimization/57662
10048 * sel-sched.c (code_motion_path_driver): Do not mark already not
10049 existing blocks in the visiting bitmap.
10051 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10053 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10054 on the insn being emitted.
10056 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10057 Will Deacon <will.deacon@arm.com>
10059 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10061 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10063 * config/arm/arm-tables.opt: Regenerate.
10065 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10067 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10068 for vector types other than V16QImode.
10069 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10070 define_expand, and call altivec_expand_vec_perm_le when producing
10071 code with little endian element order.
10072 (*altivec_vperm_<mode>_internal): New insn having previous
10073 behavior of altivec_vperm_<mode>.
10074 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10075 altivec_expand_vec_perm_le when producing code with little endian
10077 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10078 behavior of altivec_vperm_<mode>_uns.
10080 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10082 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10083 (altivec_vsumsws): Add handling for -maltivec=be with a little
10085 (altivec_vsumsws_direct): New.
10086 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10087 gen_altivec_vsumsws.
10089 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10091 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10092 vtable_pointer_value_to_binfo): New functions.
10093 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10094 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10096 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10098 * config/nios2/nios2.md (load_got_register): Initialize GOT
10099 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10100 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10102 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10104 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10105 preserverd by passthrough, do not propagate the type.
10107 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10109 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10110 (mips_atomic_assign_expand_fenv): New function.
10111 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10113 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10115 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10116 (__builtin_mips_set_fcsr): Likewise.
10117 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10118 MIPS_USI_FTYPE_VOID.
10119 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10120 (mips16_expand_set_fcsr): Likewise.
10121 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10122 (mips16_set_fcsr_stub): Likewise.
10123 (mips16_get_fcsr_one_only_stub): New class.
10124 (mips16_set_fcsr_one_only_stub): Likewise.
10125 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10126 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10127 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10128 (hard_float): New availability predicate.
10129 (mips_builtins): Add get_fcsr and set_fcsr.
10130 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10131 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10132 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10133 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10134 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10137 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10139 * config/mips/mips.c (mips_one_only_stub): New class.
10140 (mips_need_mips16_rdhwr_p): Replace with...
10141 (mips16_rdhwr_stub): ...this new variable.
10142 (mips16_stub_call_address): New function.
10143 (mips16_rdhwr_one_only_stub): New class.
10144 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10145 (mips_output_mips16_rdhwr): Delete.
10146 (mips_finish_stub): New function.
10147 (mips_code_end): Use it to handle rdhwr stubs.
10149 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10152 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10153 when calculating size of integer atomic types.
10155 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10157 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10159 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10161 PR tree-optimization/60003
10162 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10163 * profile.c (branch_prob): Use gimple_call_builtin_p
10164 to check for BUILT_IN_SETJMP_RECEIVER.
10165 * tree-inline.c (copy_bb): Call notice_special_calls.
10167 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10170 * lra-constraints.c (process_alt_operands): Update reload_sum only
10173 2014-01-31 Richard Henderson <rth@redhat.com>
10175 PR middle-end/60004
10176 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10177 until after else_eh is processed.
10179 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10181 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10182 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10183 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10184 in smmintrin.h, remove them.
10185 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10186 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10187 * config/i386/i386.md (ROUND_SAE): Fix value.
10188 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10189 (const48_operand): New.
10190 * config/i386/subst.md (round), (round_expand): Use
10191 const_4_or_8_to_11_operand.
10192 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10194 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10196 * config/i386/constraints.md (Yk): Swap meaning with k.
10197 * config/i386/i386.md (movhi_internal): Change Yk to k.
10198 (movqi_internal): Ditto.
10199 (*k<logic><mode>): Ditto.
10202 (kandn<mode>): Ditto.
10203 (*<code>hi_1): Ditto.
10204 (*<code>qi_1): Ditto.
10205 (kxnor<mode>): Ditto.
10206 (kortestzhi): Ditto.
10207 (kortestchi): Ditto.
10209 (*one_cmplhi2_1): Ditto.
10210 (*one_cmplqi2_1): Ditto.
10211 * config/i386/sse.md (): Change k to Yk.
10212 (avx512f_load<mode>_mask): Ditto.
10213 (avx512f_blendm<mode>): Ditto.
10214 (avx512f_store<mode>_mask): Ditto.
10215 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10216 (avx512f_storedqu<mode>_mask): Ditto.
10217 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10219 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10220 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10221 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10222 (avx512f_maskcmp<mode>3): Ditto.
10223 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10224 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10225 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10226 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10227 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10228 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10229 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10230 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10231 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10232 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10233 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10234 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10235 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10236 (vec_extract_lo_<mode>_maskm): Ditto.
10237 (vec_extract_hi_<mode>_maskm): Ditto.
10238 (avx512f_vternlog<mode>_mask): Ditto.
10239 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10240 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10241 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10242 (avx512f_<code>v8div16qi2_mask): Ditto.
10243 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10244 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10245 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10246 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10247 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10248 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10249 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10250 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10251 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10252 (avx512cd_maskb_vec_dupv8di): Ditto.
10253 (avx512cd_maskw_vec_dupv16si): Ditto.
10254 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10255 (avx512f_vpermi2var<mode>3_mask): Ditto.
10256 (avx512f_vpermi2var<mode>3_mask): Ditto.
10257 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10258 (*avx512f_gathersi<mode>): Ditto.
10259 (*avx512f_gathersi<mode>_2): Ditto.
10260 (*avx512f_gatherdi<mode>): Ditto.
10261 (*avx512f_gatherdi<mode>_2): Ditto.
10262 (*avx512f_scattersi<mode>): Ditto.
10263 (*avx512f_scatterdi<mode>): Ditto.
10264 (avx512f_compress<mode>_mask): Ditto.
10265 (avx512f_compressstore<mode>_mask): Ditto.
10266 (avx512f_expand<mode>_mask): Ditto.
10267 * config/i386/subst.md (mask): Change k to Yk.
10268 (mask_scalar_merge): Ditto.
10271 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10273 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10275 2014-01-31 Richard Biener <rguenther@suse.de>
10277 PR middle-end/59990
10278 * builtins.c (fold_builtin_memory_op): Make sure to not
10279 use a floating-point mode or a boolean or enumeral type for
10280 the copy operation.
10282 2014-01-30 DJ Delorie <dj@redhat.com>
10284 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10285 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10286 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10287 whenever main() has an epilogue.
10289 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10291 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10292 unused variable "field".
10293 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10294 (vsx_mergeh_<mode>): Likewise.
10295 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10296 (altivec_vmrghh): Likewise.
10297 (altivec_vmrghw): Likewise.
10298 (altivec_vmrglb): Likewise.
10299 (altivec_vmrglh): Likewise.
10300 (altivec_vmrglw): Likewise.
10301 (altivec_vspltb): Add missing uses.
10302 (altivec_vsplth): Likewise.
10303 (altivec_vspltw): Likewise.
10304 (altivec_vspltsf): Likewise.
10306 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10309 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10310 frame related instructions.
10312 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10314 PR rtl-optimization/59959
10315 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10316 any reload of register whose subreg is invalid.
10318 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10320 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10321 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10322 Add missing return type - void.
10324 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10326 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10327 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10328 remove element index adjustment for endian (now handled in vsx.md
10330 (altivec_expand_vec_perm_const): Use
10331 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10332 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10333 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10334 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10335 define_expand and a new define_insn *altivec_vspltb_internal;
10336 adjust for -maltivec=be on a little endian target.
10337 (altivec_vspltb_direct): New.
10338 (altivec_vsplth): Divide into a define_expand and a new
10339 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10340 little endian target.
10341 (altivec_vsplth_direct): New.
10342 (altivec_vspltw): Divide into a define_expand and a new
10343 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10344 little endian target.
10345 (altivec_vspltw_direct): New.
10346 (altivec_vspltsf): Divide into a define_expand and a new
10347 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10348 a little endian target.
10350 2014-01-30 Richard Biener <rguenther@suse.de>
10352 PR tree-optimization/59993
10353 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10354 can propagate form the earlier stmt and avoid the transform
10355 when the intermediate result is needed.
10357 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10359 * README.Portability: Fix typo.
10361 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10363 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10364 comparison_operator with ordered_comparison_operator.
10366 2014-01-30 Nick Clifton <nickc@redhat.com>
10368 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10369 Rename to mn10300_store_multiple_regs.
10370 * config/mn10300/mn10300.c: Likewise.
10371 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10372 store_multiple_regs.
10373 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10374 Call mn10300_store_multiple_regs.
10376 2014-01-30 Nick Clifton <nickc@redhat.com>
10377 DJ Delorie <dj@redhat.com>
10379 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10380 %fp 2 to keep registers after it properly word-aligned.
10381 (rl78_alloc_physical_registers_umul): Handle the case where both
10382 input operands are the same.
10384 2014-01-30 Richard Biener <rguenther@suse.de>
10386 PR tree-optimization/59903
10387 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10390 2014-01-30 Jason Merrill <jason@redhat.com>
10393 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10396 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10398 2014-01-30 Richard Biener <rguenther@suse.de>
10400 PR tree-optimization/59951
10401 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10403 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10406 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10407 SFmode to DFmode case.
10409 2014-01-29 DJ Delorie <dj@redhat.com>
10411 * config/msp430/msp430.opt (-minrt): New.
10412 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10414 (ENDFILE_SPEC): Likewise.
10416 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10418 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10419 (estimate_function_body_sizes): Use it.
10421 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10424 * dwarf2out.c (is_cxx_auto): New.
10425 (is_base_type): Use it.
10426 (gen_type_die_with_usage): Likewise.
10428 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10430 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10431 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10432 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10433 -maltivec=be with LE targets.
10434 (vsx_mergeh_<mode>): Likewise.
10435 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10436 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10437 (altivec_vmrghb): Replace with define_expand and new
10438 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10439 (altivec_vmrghb_direct): New define_insn.
10440 (altivec_vmrghh): Replace with define_expand and new
10441 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10442 (altivec_vmrghh_direct): New define_insn.
10443 (altivec_vmrghw): Replace with define_expand and new
10444 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10445 (altivec_vmrghw_direct): New define_insn.
10446 (*altivec_vmrghsf): Adjust for endianness.
10447 (altivec_vmrglb): Replace with define_expand and new
10448 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10449 (altivec_vmrglb_direct): New define_insn.
10450 (altivec_vmrglh): Replace with define_expand and new
10451 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10452 (altivec_vmrglh_direct): New define_insn.
10453 (altivec_vmrglw): Replace with define_expand and new
10454 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10455 (altivec_vmrglw_direct): New define_insn.
10456 (*altivec_vmrglsf): Adjust for endianness.
10457 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10458 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10459 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10460 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10461 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10462 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10463 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10464 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10466 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10468 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10469 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10472 2014-01-29 Richard Biener <rguenther@suse.de>
10474 PR tree-optimization/58742
10475 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10476 associate_pointerplus_align.
10477 (associate_pointerplus_diff): New function.
10478 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10479 and associate_pointerplus_diff.
10481 2014-01-29 Richard Biener <rguenther@suse.de>
10483 * lto-streamer.h (LTO_major_version): Bump to 3.
10484 (LTO_minor_version): Reset to 0.
10486 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10488 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10489 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10490 (arm_file_start): Generate correct asm header for armv7ve.
10491 * config/arm/bpabi.h: Add multilib support for armv7ve.
10492 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10493 and cortex-a15 to armv7ve.
10494 * config/arm/t-aprofile: Add multilib support for armv7ve.
10495 * doc/invoke.texi: Document -march=armv7ve.
10497 2014-01-29 Richard Biener <rguenther@suse.de>
10499 PR tree-optimization/58742
10500 * tree-ssa-forwprop.c (associate_plusminus): Return true
10501 if we changed sth, defer EH cleanup to ...
10502 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10503 (simplify_mult): New function.
10505 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10507 PR middle-end/59917
10508 PR tree-optimization/59920
10509 * tree.c (build_common_builtin_nodes): Remove
10510 __builtin_setjmp_dispatcher initialization.
10511 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10512 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10513 instead of gsi_after_labels + manually skipping debug stmts.
10514 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10515 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10516 * tree-inline.c (copy_edges_for_bb): Remove
10517 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10518 argument. Ignore computed_goto_p stmts. Don't call
10519 make_abnormal_goto_edges. If a call might need abnormal edges
10520 for non-local gotos, see if it already has an edge to
10521 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10522 with true argument, don't do anything then, otherwise add
10523 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10524 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10526 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10527 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10528 (lower_stmt): Don't set data->calls_builtin_setjmp.
10529 (lower_builtin_setjmp): Adjust comment.
10530 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10531 * tree-cfg.c (found_computed_goto): Remove.
10532 (factor_computed_gotos): Remove.
10533 (make_goto_expr_edges): Return bool, true for computed gotos.
10534 Don't call make_abnormal_goto_edges.
10535 (build_gimple_cfg): Don't set found_computed_goto, don't call
10536 factor_computed_gotos.
10537 (computed_goto_p): No longer static.
10538 (make_blocks): Don't set found_computed_goto.
10539 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10540 (make_edges): If make_goto_expr_edges returns true, push bb
10541 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10542 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10543 vector. Record mapping between bbs and OpenMP regions if there
10544 are any, adjust make_gimple_omp_edges caller. Call
10545 handle_abnormal_edges.
10546 (make_abnormal_goto_edges): Remove.
10547 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10548 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10549 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10550 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10551 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10552 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10553 filling *region also set *region_idx to (*region)->entry->index.
10556 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10557 For REGs set ORIGINAL_REGNO.
10559 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10561 * doc/md.texi: Mention that a target shouldn't implement
10562 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10565 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10567 PR tree-optimization/59594
10568 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10569 a copy of the datarefs vector rather than the vector itself.
10571 2014-01-28 Jason Merrill <jason@redhat.com>
10574 * dwarf2out.c (auto_die): New static.
10575 (gen_type_die_with_usage): Handle C++1y 'auto'.
10576 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10579 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10582 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10583 (SPEC_X32): Likewise.
10584 (SPEC_64): Likewise.
10585 * config/i386/i386.c (ix86_option_override_internal): Turn off
10586 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10588 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10589 * config/i386/i386.h (TARGET_16BIT): New macro.
10590 (TARGET_16BIT_P): Likewise.
10591 * config/i386/i386.opt: Add m16.
10592 * doc/invoke.texi: Document -m16.
10594 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10596 PR preprocessor/59935
10597 * input.c (location_get_source_line): Bail out on when line number
10598 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10600 2014-01-28 Richard Biener <rguenther@suse.de>
10602 PR tree-optimization/58742
10603 * tree-ssa-forwprop.c (associate_plusminus): Handle
10604 pointer subtraction of the form (T)(P + A) - (T)P.
10606 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10608 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10611 2014-01-28 Richard Biener <rguenther@suse.de>
10614 2014-01-28 Richard Biener <rguenther@suse.de>
10616 PR rtl-optimization/45364
10617 PR rtl-optimization/59890
10618 * var-tracking.c (local_get_addr_clear_given_value): Handle
10619 already cleared slot.
10620 (val_reset): Handle not allocated local_get_addr_cache.
10621 (vt_find_locations): Use post-order on the inverted CFG.
10623 2014-01-28 Richard Biener <rguenther@suse.de>
10625 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10627 2014-01-28 Richard Biener <rguenther@suse.de>
10629 PR rtl-optimization/45364
10630 PR rtl-optimization/59890
10631 * var-tracking.c (local_get_addr_clear_given_value): Handle
10632 already cleared slot.
10633 (val_reset): Handle not allocated local_get_addr_cache.
10634 (vt_find_locations): Use post-order on the inverted CFG.
10636 2014-01-28 Alan Modra <amodra@gmail.com>
10638 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10639 * configure.ac <recursive call for build != host>: Define
10640 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10641 and LD_FOR_BUILD too.
10642 * configure: Regenerate.
10644 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10646 * config/i386/i386.c (get_builtin_code_for_version): Separate
10647 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10648 Broadwell from Haswell.
10650 2014-01-27 Steve Ellcey <sellcey@mips.com>
10652 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10653 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10654 * config/mips/mips.c (mips_option_override): Change setting
10656 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10657 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10658 Change from Mask to Var.
10660 2014-01-27 Jeff Law <law@redhat.com>
10662 * ipa-inline.c (inline_small_functions): Fix typo.
10664 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10666 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10667 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10668 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10669 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10670 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10671 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10672 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10673 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10674 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10675 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10676 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10677 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10678 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10679 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10680 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10681 (_mm512_storeu_epi64): Ditto.
10682 (_mm512_cmpge_epi32_mask): Ditto.
10683 (_mm512_cmpge_epu32_mask): Ditto.
10684 (_mm512_cmpge_epi64_mask): Ditto.
10685 (_mm512_cmpge_epu64_mask): Ditto.
10686 (_mm512_cmple_epi32_mask): Ditto.
10687 (_mm512_cmple_epu32_mask): Ditto.
10688 (_mm512_cmple_epi64_mask): Ditto.
10689 (_mm512_cmple_epu64_mask): Ditto.
10690 (_mm512_cmplt_epi32_mask): Ditto.
10691 (_mm512_cmplt_epu32_mask): Ditto.
10692 (_mm512_cmplt_epi64_mask): Ditto.
10693 (_mm512_cmplt_epu64_mask): Ditto.
10694 (_mm512_cmpneq_epi32_mask): Ditto.
10695 (_mm512_cmpneq_epu32_mask): Ditto.
10696 (_mm512_cmpneq_epi64_mask): Ditto.
10697 (_mm512_cmpneq_epu64_mask): Ditto.
10698 (_mm512_expand_pd): Ditto.
10699 (_mm512_expand_ps): Ditto.
10700 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10701 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10702 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10703 * config/i386/i386.c (ix86_builtins): Add
10704 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10705 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10706 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10707 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10708 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10709 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10710 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10711 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10712 IX86_BUILTIN_PMOVUSQW512_MEM.
10713 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10714 __builtin_ia32_pmovsqd512mem_mask,
10715 __builtin_ia32_pmovqd512mem_mask,
10716 __builtin_ia32_pmovusqw512mem_mask,
10717 __builtin_ia32_pmovsqw512mem_mask,
10718 __builtin_ia32_pmovqw512mem_mask,
10719 __builtin_ia32_pmovusdw512mem_mask,
10720 __builtin_ia32_pmovsdw512mem_mask,
10721 __builtin_ia32_pmovdw512mem_mask,
10722 __builtin_ia32_pmovqb512mem_mask,
10723 __builtin_ia32_pmovusqb512mem_mask,
10724 __builtin_ia32_pmovsqb512mem_mask,
10725 __builtin_ia32_pmovusdb512mem_mask,
10726 __builtin_ia32_pmovsdb512mem_mask,
10727 __builtin_ia32_pmovdb512mem_mask.
10728 (bdesc_args): Add __builtin_ia32_expanddf512,
10729 __builtin_ia32_expandsf512.
10730 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10731 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10732 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10733 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10734 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10735 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10736 (avx512f_<code>v8div16qi2_mask_store): This.
10737 (avx512f_expand<mode>): New.
10739 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10741 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10743 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10744 (_mm512_prefetch_i32scatter_pd): Ditto.
10745 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10746 (_mm512_prefetch_i64scatter_pd): Ditto.
10747 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10748 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10749 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10750 (_mm512_prefetch_i32scatter_ps): Ditto.
10751 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10752 (_mm512_prefetch_i64scatter_ps): Ditto.
10753 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10754 * config/i386/i386-builtin-types.def: Define
10755 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10756 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10757 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10758 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10759 IX86_BUILTIN_SCATTERPFQPD.
10760 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10761 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10762 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10763 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10764 __builtin_ia32_scatterpfqps.
10765 (ix86_expand_builtin): Expand new built-ins.
10766 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10767 fix memory access data type.
10768 (*avx512pf_gatherpf<mode>_mask): Ditto.
10769 (*avx512pf_gatherpf<mode>): Ditto.
10770 (avx512pf_scatterpf<mode>): Ditto.
10771 (*avx512pf_scatterpf<mode>_mask): Ditto.
10772 (*avx512pf_scatterpf<mode>): Ditto.
10773 (GATHER_SCATTER_SF_MEM_MODE): New.
10774 (avx512pf_gatherpf<mode>df): Ditto.
10775 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10776 (*avx512pf_scatterpf<mode>df): Ditto.
10778 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10781 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10782 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10785 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10787 * common/config/arm/arm-common.c
10788 (arm_rewrite_mcpu): Handle multiple names.
10790 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10792 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10794 * gimple-builder.h (create_gimple_tmp): Delete.
10796 2014-01-27 Christian Bruel <christian.bruel@st.com>
10798 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10801 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10803 * config/pa/pa.md (call): Generate indirect long calls to non-local
10804 functions when outputing 32-bit code.
10805 (call_value): Likewise except for special call to buggy powf function.
10807 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10808 portable runtime and PIC indirect calls.
10809 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10810 and PIC call sequences. Use ldo instead of blr to set return register
10811 in PIC call sequence.
10813 2014-01-25 Walter Lee <walt@tilera.com>
10815 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10816 avoid clobbering a live register.
10818 2014-01-25 Walter Lee <walt@tilera.com>
10820 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10821 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10822 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10823 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10825 2014-01-25 Walter Lee <walt@tilera.com>
10827 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10828 arguments on even registers.
10829 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10831 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10832 (BIGGEST_ALIGNMENT): Ditto.
10833 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10835 2014-01-25 Walter Lee <walt@tilera.com>
10837 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10838 insns before bundling.
10839 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10841 2014-01-25 Walter Lee <walt@tilera.com>
10843 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10844 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10845 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10847 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10849 * config/mips/constraints.md (kl): Delete.
10850 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10851 define expands, using...
10852 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10853 instructions for MIPS16.
10854 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10855 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10857 2014-01-25 Walter Lee <walt@tilera.com>
10859 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10863 2014-01-25 Walter Lee <walt@tilera.com>
10865 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10866 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10868 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10870 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10873 2014-01-25 Jakub Jelinek <jakub@redhat.com>
10875 * print-rtl.c (in_call_function_usage): New var.
10876 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
10877 EXPR_LIST mode as mode and not as reg note name.
10879 PR middle-end/59561
10880 * cfgloopmanip.c (copy_loop_info): If
10881 loop->warned_aggressive_loop_optimizations, make sure
10882 the flag is set in target loop too.
10884 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
10886 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
10888 * builtins.def: Likewise.
10889 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
10890 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
10891 * ira.c (ira_setup_eliminable_regset): Likewise.
10892 * omp-low.c (gate_expand_omp): Likewise.
10893 (execute_lower_omp): Likewise.
10894 (diagnose_sb_0): Likewise.
10895 (gate_diagnose_omp_blocks): Likewise.
10896 (simd_clone_clauses_extract): Likewise.
10899 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10901 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
10902 correction for little endian...
10903 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
10906 2014-01-24 Jeff Law <law@redhat.com>
10908 PR tree-optimization/59919
10909 * tree-vrp.c (find_assert_locations_1): Do not register asserts
10910 for non-returning calls.
10912 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
10914 * common/config/aarch64/aarch64-common.c
10915 (aarch64_rewrite_mcpu): Handle multiple names.
10916 * config/aarch64/aarch64.h
10917 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10919 2014-01-24 Dodji Seketeli <dodji@redhat.com>
10921 * input.c (add_file_to_cache_tab): Handle the case where fopen
10924 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
10927 * config/i386/i386.md (pushsf splitter): Get stack adjustment
10928 from push operand if code of push isn't PRE_DEC.
10930 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
10933 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
10934 -mquad-memory-atomic. Update -mquad-memory documentation to say
10935 it is only used for non-atomic loads/stores.
10937 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
10938 -mquad-memory or -mquad-memory-atomic switches.
10940 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
10941 -mquad-memory-atomic to ISA 2.07 support.
10943 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
10944 to separate support of normal quad word memory operations (ldq, stq)
10945 from the atomic quad word memory operations.
10947 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
10948 support to separate non-atomic quad word operations from atomic
10949 quad word operations. Disable non-atomic quad word operations in
10950 little endian mode so that we don't have to swap words after the
10951 load and before the store.
10952 (quad_load_store_p): Add comment about atomic quad word support.
10953 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
10954 options printed with -mdebug=reg.
10956 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
10957 -mquad-memory-atomic as the test for whether we have quad word
10958 atomic instructions.
10959 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
10960 or -mp8-vector are used, allow byte/half-word atomic operations.
10962 * config/rs6000/sync.md (load_lockedti): Insure that the address
10963 is a proper indexed or indirect address for the lqarx instruction.
10964 On little endian systems, swap the hi/lo registers after the lqarx
10966 (load_lockedpti): Use indexed_or_indirect_operand predicate to
10967 insure the address is valid for the lqarx instruction.
10968 (store_conditionalti): Insure that the address is a proper indexed
10969 or indirect address for the stqcrx. instruction. On little endian
10970 systems, swap the hi/lo registers before doing the stqcrx.
10972 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
10973 insure the address is valid for the stqcrx. instruction.
10975 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
10976 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
10977 type of quad memory support is available.
10979 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
10981 PR regression/59915
10982 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
10983 there is a danger of looping.
10985 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
10987 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
10988 force flag_ira_loop_pressure if set via command line.
10990 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
10992 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
10993 (ashr_simd): New builtin handling DI mode.
10994 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
10995 (aarch64_sshr_simddi): New match pattern.
10996 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
10997 (vshrd_n_s64): Likewise.
10998 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11000 2014-01-23 Nick Clifton <nickc@redhat.com>
11002 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11003 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11004 favour of mcu specific scripts.
11005 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11008 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11009 Alex Velenko <Alex.Velenko@arm.com>
11011 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11012 (vaddv_s16): Likewise.
11013 (vaddv_s32): Likewise.
11014 (vaddv_u8): Likewise.
11015 (vaddv_u16): Likewise.
11016 (vaddv_u32): Likewise.
11017 (vaddvq_s8): Likewise.
11018 (vaddvq_s16): Likewise.
11019 (vaddvq_s32): Likewise.
11020 (vaddvq_s64): Likewise.
11021 (vaddvq_u8): Likewise.
11022 (vaddvq_u16): Likewise.
11023 (vaddvq_u32): Likewise.
11024 (vaddvq_u64): Likewise.
11025 (vaddv_f32): Likewise.
11026 (vaddvq_f32): Likewise.
11027 (vaddvq_f64): Likewise.
11028 (vmaxv_f32): Likewise.
11029 (vmaxv_s8): Likewise.
11030 (vmaxv_s16): Likewise.
11031 (vmaxv_s32): Likewise.
11032 (vmaxv_u8): Likewise.
11033 (vmaxv_u16): Likewise.
11034 (vmaxv_u32): Likewise.
11035 (vmaxvq_f32): Likewise.
11036 (vmaxvq_f64): Likewise.
11037 (vmaxvq_s8): Likewise.
11038 (vmaxvq_s16): Likewise.
11039 (vmaxvq_s32): Likewise.
11040 (vmaxvq_u8): Likewise.
11041 (vmaxvq_u16): Likewise.
11042 (vmaxvq_u32): Likewise.
11043 (vmaxnmv_f32): Likewise.
11044 (vmaxnmvq_f32): Likewise.
11045 (vmaxnmvq_f64): Likewise.
11046 (vminv_f32): Likewise.
11047 (vminv_s8): Likewise.
11048 (vminv_s16): Likewise.
11049 (vminv_s32): Likewise.
11050 (vminv_u8): Likewise.
11051 (vminv_u16): Likewise.
11052 (vminv_u32): Likewise.
11053 (vminvq_f32): Likewise.
11054 (vminvq_f64): Likewise.
11055 (vminvq_s8): Likewise.
11056 (vminvq_s16): Likewise.
11057 (vminvq_s32): Likewise.
11058 (vminvq_u8): Likewise.
11059 (vminvq_u16): Likewise.
11060 (vminvq_u32): Likewise.
11061 (vminnmv_f32): Likewise.
11062 (vminnmvq_f32): Likewise.
11063 (vminnmvq_f64): Likewise.
11065 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11067 * config/aarch64/aarch64-simd.md
11068 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11069 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11070 (*aarch64_mul3_elt<mode>): Likewise.
11071 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11072 (*aarch64_mul3_elt_to_64v2df): Likewise.
11073 (*aarch64_mla_elt<mode>): Likewise.
11074 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11075 (*aarch64_mls_elt<mode>): Likewise.
11076 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11077 (*aarch64_fma4_elt<mode>): Likewise.
11078 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11079 (*aarch64_fma4_elt_to_64v2df): Likewise.
11080 (*aarch64_fnma4_elt<mode>): Likewise.
11081 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11082 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11083 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11084 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11085 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11086 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11087 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11088 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11089 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11091 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11093 * config/aarch64/aarch64-simd.md
11094 (aarch64_be_checked_get_lane<mode>): New define_expand.
11095 * config/aarch64/aarch64-simd-builtins.def
11096 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11097 New builtin definition.
11098 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11099 Use new safe be builtin.
11101 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11103 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11105 (aarch64_be_st1<mode>): Likewise.
11106 (aarch_ld1<VALL:mode>): Define_expand modified.
11107 (aarch_st1<VALL:mode>): Likewise.
11108 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11109 (UNSPEC_ST1): Likewise.
11111 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11113 * config/microblaze/microblaze.md: Add trap insn and attribute
11115 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11117 PR preprocessor/58580
11118 * input.h (location_get_source_line): Take an additional line_size
11120 (void diagnostics_file_cache_fini): Declare new function.
11121 * input.c (struct fcache): New type.
11122 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11123 New static constants.
11124 (diagnostic_file_cache_init, total_lines_num)
11125 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11126 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11127 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11128 (get_next_line, read_next_line, goto_next_line, read_line_num):
11129 New static function definitions.
11130 (diagnostic_file_cache_fini): New function.
11131 (location_get_source_line): Take an additional output line_len
11132 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11134 * diagnostic.c (diagnostic_finish): Call
11135 diagnostic_file_cache_fini.
11136 (adjust_line): Take an additional input parameter for the length
11137 of the line, rather than calculating it with strlen.
11138 (diagnostic_show_locus): Adjust the use of
11139 location_get_source_line and adjust_line with respect to their new
11140 signature. While displaying a line now, do not stop at the first
11141 null byte. Rather, display the zero byte as a space and keep
11142 going until we reach the size of the line.
11143 * Makefile.in: Add vec.o to OBJS-libcommon
11145 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11146 Ilya Tocar <ilya.tocar@intel.com>
11148 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11149 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11150 (__builtin_ia32_kmov16): Ditto.
11151 * config/i386/i386.md (UNSPEC_KMOV): New.
11154 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11156 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11157 (_mm512_storeu_si512): Ditto.
11159 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11162 * rtl.h (get_referenced_operands): Declare.
11163 * recog.c (get_referenced_operands): New function.
11164 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11165 operands have been referenced when recording LO_SUM references.
11167 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11169 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11171 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11173 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11174 Enable for generic and recent AMD targets.
11176 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11178 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11179 ARG_SIZE note when adjustment was eliminated.
11181 2014-01-22 Jeff Law <law@redhat.com>
11183 PR tree-optimization/59597
11184 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11185 in file. Accept new argument REGISTERING and use it to modify
11186 dump output appropriately.
11187 (register_jump_thread): Corresponding changes.
11188 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11189 thread paths involving joiner blocks. Add code to dump cancelled
11190 jump threading paths.
11192 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11194 PR rtl-optimization/59477
11195 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11196 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11198 2014-01-22 Tom Tromey <tromey@redhat.com>
11200 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11202 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11204 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11206 PR rtl-optimization/59896
11207 * lra-constraints.c (process_alt_operands): Check unused note for
11208 matched operands of insn with no output reloads.
11210 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11212 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11213 (mips_move_from_gpr_cost): Likewise.
11215 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11217 PR rtl-optimization/59858
11218 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11219 ira_class_hard_regs_num.
11220 (process_alt_operands): Increase reject for dying matched operand.
11222 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11225 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11226 smaller than size, perform several stores or loads and stores
11227 at dst + count - size to store or copy all of size bytes, rather
11228 than just last modesize bytes.
11230 2014-01-20 DJ Delorie <dj@redhat.com>
11232 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11233 that CLOBBERs are REGs before propogating their values.
11235 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11237 PR middle-end/59789
11238 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11239 (cgraph_inline_failed_type): New function.
11240 * cgraph.h (DEFCIFCODE): Add type.
11241 (cgraph_inline_failed_type_t): New enum.
11242 (cgraph_inline_failed_type): New prototype.
11243 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11244 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11245 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11246 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11247 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11248 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11249 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11250 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11251 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11252 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11253 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11254 OPTIMIZATION_MISMATCH.
11255 * tree-inline.c (expand_call_inline): Emit errors during
11256 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11258 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11261 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11262 mode attribute in insn output.
11264 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11266 * output.h (output_constant): Delete.
11267 * varasm.c (output_constant): Make private.
11269 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11271 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11273 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11275 PR middle-end/59860
11276 * tree.h (fold_builtin_strcat): New prototype.
11277 * builtins.c (fold_builtin_strcat): No longer static. Add len
11278 argument, if non-NULL, don't call c_strlen. Optimize
11279 directly into __builtin_memcpy instead of __builtin_strcpy.
11280 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11281 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11283 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11285 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11286 for SImode_address_operand operands, having only a REG argument.
11288 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11290 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11291 loader name using mbig-endian.
11292 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11294 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11296 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11297 (-mtune): Likewise.
11300 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11302 * config/aarch64/aarch64-protos.h
11303 (aarch64_cannot_change_mode_class_ptr): Declare.
11304 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11305 aarch64_cannot_change_mode_class_ptr): New.
11306 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11307 backend hook aarch64_cannot_change_mode_class.
11309 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11311 * common/config/aarch64/aarch64-common.c
11312 (aarch64_handle_option): Don't handle any option order logic here.
11313 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11314 selected_cpu, warn on architecture version mismatch.
11315 (aarch64_override_options): Fix parsing order for option strings.
11317 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11318 Iain Sandoe <iain@codesourcery.com>
11321 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11322 warning. Amend comment to reflect current functionality.
11324 2014-01-20 Richard Biener <rguenther@suse.de>
11326 PR middle-end/59860
11327 * builtins.c (fold_builtin_strcat): Remove case better handled
11328 by tree-ssa-strlen.c.
11330 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11332 * config/aarch64/aarch64.opt
11333 (mcpu, march, mtune): Make case-insensitive.
11335 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11338 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11339 if operands[1] is a REG or ZERO_EXTEND of a REG.
11341 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11343 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11345 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11347 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11348 long non-pic millicode calls.
11350 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11352 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11354 2014-01-19 Kito Cheng <kito@0xlab.org>
11356 * builtins.c (expand_movstr): Check movstr expand done or fail.
11358 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11359 H.J. Lu <hongjiu.lu@intel.com>
11362 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11363 to DImode for zero-extended addresses.
11365 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11367 PR rtl-optimization/57763
11368 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11369 on the new indirect jump_insn and increment LABEL_NUSES (label).
11371 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11375 * config.gcc (x86_archs): New variable.
11376 (x86_64_archs): Likewise.
11377 (x86_cpus): Likewise.
11378 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11379 --with-arch/--with-cpu= options.
11380 Support --with-arch=/--with-cpu={nehalem,westmere,
11381 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11383 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11385 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11386 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11388 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11390 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11392 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11395 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11396 clear cpp_get_options (parse_in)->warn_unused_macros for
11397 ix86_target_macros_internal with cpp_define.
11399 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11401 * jump.c (delete_related_insns): Keep (use (insn))s.
11402 * reorg.c (redundant_insn): Check for barriers too.
11404 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11406 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11408 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11410 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11411 call to $$dyncall when TARGET_LONG_CALLS is true.
11413 2014-01-17 Jeff Law <law@redhat.com>
11415 * ree.c (combine_set_extension): Temporarily disable test for
11416 changing number of hard registers.
11418 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11420 PR middle-end/58125
11421 * ipa-inline-analysis.c (inline_free_summary):
11422 Do not free summary of aliases.
11424 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11426 PR middle-end/59706
11427 * gimplify.c (gimplify_expr): Use create_tmp_var
11428 instead of create_tmp_var_raw. If cond doesn't have
11429 integral type, don't add the IFN_ANNOTATE builtin at all.
11431 2014-01-17 Martin Jambor <mjambor@suse.cz>
11434 * ipa-cp.c (prev_edge_clone): New variable.
11435 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11436 Also resize prev_edge_clone vector.
11437 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11438 (ipcp_edge_removal_hook): New function.
11439 (ipcp_driver): Register ipcp_edge_removal_hook.
11441 2014-01-17 Andrew Pinski <apinski@cavium.com>
11442 Steve Ellcey <sellcey@mips.com>
11445 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11448 2014-01-17 Jeff Law <law@redhat.com>
11450 PR middle-end/57904
11451 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11452 so that pass_ccp runs first.
11454 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11456 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11457 (ix86_adjust_cost): Use !TARGET_XXX.
11458 (do_reorder_for_imul): Likewise.
11459 (swap_top_of_ready_list): Likewise.
11460 (ix86_sched_reorder): Likewise.
11462 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11464 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11465 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11466 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11467 (intel_memset): New. Duplicate slm_memset.
11468 (intel_cost): New. Duplicate slm_cost.
11469 (m_INTEL): New macro.
11470 (processor_target_table): Add "intel".
11471 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11472 with PROCESSOR_INTEL for "intel".
11473 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11474 PROCESSOR_SILVERMONT.
11475 (ix86_issue_rate): Likewise.
11476 (ix86_adjust_cost): Likewise.
11477 (ia32_multipass_dfa_lookahead): Likewise.
11478 (swap_top_of_ready_list): Likewise.
11479 (ix86_sched_reorder): Likewise.
11480 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11481 instead of TARGET_OPT_AGU.
11482 * config/i386/i386.h (TARGET_INTEL): New.
11483 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11484 (processor_type): Add PROCESSOR_INTEL.
11485 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11486 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11488 2014-01-17 Marek Polacek <polacek@redhat.com>
11491 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11494 2014-01-17 Richard Biener <rguenther@suse.de>
11496 PR tree-optimization/46590
11497 * opts.c (default_options_table): Add entries for
11498 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11499 all enabled at -O1 but not for -Og.
11500 * common.opt (fbranch-count-reg): Remove Init(1).
11501 (fmove-loop-invariants): Likewise.
11502 (ftree-pta): Likewise.
11504 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11506 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11507 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11508 decls to at least the GCC 4.8 used alignments.
11511 * tree-nested.c (convert_nonlocal_reference_stmt,
11512 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11513 of GIMPLE_BIND stmts, adjust associated decls.
11515 2014-01-17 Richard Biener <rguenther@suse.de>
11517 PR tree-optimization/46590
11518 * vec.h (vec<>::bseach): New member function implementing
11519 binary search according to C89 bsearch.
11520 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11521 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11522 bitmap pointer again. Make accesses_in_loop a flat array.
11523 (mem_ref_obstack): New global.
11524 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11525 (mark_ref_stored): Likewise.
11526 (ref_indep_loop_p_2): Likewise.
11527 (set_ref_stored_in_loop): New helper function.
11528 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11529 (memref_free): Adjust.
11530 (record_mem_ref_loc): Simplify.
11531 (gather_mem_refs_stmt): Adjust.
11532 (sort_locs_in_loop_postorder_cmp): New function.
11533 (analyze_memory_references): Sort accesses_in_loop after
11534 loop postorder number.
11535 (find_ref_loc_in_loop_cmp): New function.
11536 (for_all_locs_in_loop): Find relevant cluster of locs in
11537 accesses_in_loop and iterate without recursion.
11538 (execute_sm): Avoid uninit warning.
11539 (struct ref_always_accessed): Simplify.
11540 (ref_always_accessed::operator ()): Likewise.
11541 (ref_always_accessed_p): Likewise.
11542 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11543 loop postorder numbers here.
11544 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11547 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11550 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11551 on decls for which assemble_alias has been called.
11553 2014-01-17 Nick Clifton <nickc@redhat.com>
11555 * config/msp430/msp430.opt: (mcpu): New option.
11556 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11557 (msp430_option_override): Parse target_cpu. If the MCU name
11558 matches a generic string, clear target_mcu.
11559 (msp430_attr): Allow numeric interrupt values up to 63.
11560 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11561 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11563 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11565 * config/msp430/msp430.md (popm): Use %J rather than %I.
11566 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11567 (addhi_cy_i): Use immediate_operand for operand 2.
11568 * doc/invoke.texi: Document -mcpu option.
11570 2014-01-17 Richard Biener <rguenther@suse.de>
11572 PR rtl-optimization/38518
11573 * df.h (df_analyze_loop): Declare.
11574 * df-core.c: Include cfgloop.h.
11575 (df_analyze_1): Split out main part of df_analyze.
11576 (df_analyze): Adjust.
11577 (loop_inverted_post_order_compute): New function.
11578 (loop_post_order_compute): Likewise.
11579 (df_analyze_loop): New function avoiding whole-function
11580 postorder computes.
11581 * loop-invariant.c (find_defs): Use df_analyze_loop.
11582 (find_invariants): Adjust.
11583 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11585 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11587 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11588 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11590 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11592 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11593 traversal when removing references.
11595 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11598 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11600 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11602 PR middle-end/56791
11603 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11604 pushing a reload for an autoinc when we had previously reloaded an
11605 inner part of the address.
11607 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11609 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11611 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11612 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11613 when not giving up or versioning for alias only because of
11615 (vect_analyze_data_ref_dependences): Set to true.
11616 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11618 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11619 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11622 PR middle-end/58344
11623 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11626 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11627 operand 0 predicate for gathers, use a new pseudo as subtarget.
11629 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11631 PR middle-end/59609
11632 * lra-constraints.c (process_alt_operands): Add printing debug info.
11633 Check absence of input/output reloads for matched operands too.
11635 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11637 PR rtl-optimization/59835
11638 * ira.c (ira_init_register_move_cost): Increase cost for
11641 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11643 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11645 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11648 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11649 non-register objects. Use gen_(high/low)part more consistently.
11652 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11655 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11656 endian support, remove tests for WORDS_BIG_ENDIAN.
11657 (p8_mfvsrd_3_<mode>): Likewise.
11658 (reload_gpr_from_vsx<mode>): Likewise.
11659 (reload_gpr_from_vsxsf): Likewise.
11660 (p8_mfvsrd_4_disf): Likewise.
11662 2014-01-16 Richard Biener <rguenther@suse.de>
11664 PR rtl-optimization/46590
11665 * lcm.c (compute_antinout_edge): Use postorder iteration.
11666 (compute_laterin): Use inverted postorder iteration.
11668 2014-01-16 Nick Clifton <nickc@redhat.com>
11670 PR middle-end/28865
11671 * varasm.c (output_constant): Return the number of bytes actually
11673 (output_constructor_array_range): Update the field size with the
11674 number of bytes emitted by output_constant.
11675 (output_constructor_regular_field): Likewise. Also do not
11676 complain if the total number of bytes emitted is now greater
11677 than the expected fieldpos.
11678 * output.h (output_constant): Update prototype and descriptive comment.
11680 2014-01-16 Marek Polacek <polacek@redhat.com>
11682 PR middle-end/59827
11683 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11684 it is error_mark_node.
11686 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11688 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11689 VALID_AVX256_REG_OR_OI_MODE.
11691 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11693 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11694 current procedure should be profiled.
11696 2014-01-15 Andrew Pinski <apinski@cavium.com>
11698 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11699 of moving from/to the STACK_REG register class.
11701 2014-01-15 Richard Henderson <rth@redhat.com>
11704 * reginfo.c (global_regs_decl): Globalize.
11705 * rtl.h (global_regs_decl): Declare.
11706 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11707 reserved via global_regs.
11709 2014-01-15 Teresa Johnson <tejohnson@google.com>
11711 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11713 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11715 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11716 and vmulosh rather than call gen_vec_widen_smult_*.
11717 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11718 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11719 (vec_widen_smult_even_v16qi): Likewise.
11720 (vec_widen_umult_even_v8hi): Likewise.
11721 (vec_widen_smult_even_v8hi): Likewise.
11722 (vec_widen_umult_odd_v16qi): Likewise.
11723 (vec_widen_smult_odd_v16qi): Likewise.
11724 (vec_widen_umult_odd_v8hi): Likewise.
11725 (vec_widen_smult_odd_v8hi): Likewise.
11726 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11727 vmuloub rather than call gen_vec_widen_umult_*.
11728 (vec_widen_umult_lo_v16qi): Likewise.
11729 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11730 vmulosb rather than call gen_vec_widen_smult_*.
11731 (vec_widen_smult_lo_v16qi): Likewise.
11732 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11733 rather than call gen_vec_widen_umult_*.
11734 (vec_widen_umult_lo_v8hi): Likewise.
11735 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11736 rather than call gen_vec_widen_smult_*.
11737 (vec_widen_smult_lo_v8hi): Likewise.
11739 2014-01-15 Jeff Law <law@redhat.com>
11741 PR tree-optimization/59747
11742 * ree.c (find_and_remove_re): Properly handle case where a second
11743 eliminated extension requires widening a copy created for elimination
11744 of a prior extension.
11745 (combine_set_extension): Ensure that the number of hard regs needed
11746 for a destination register does not change when we widen it.
11748 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11750 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11751 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11752 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11753 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11754 (avr-*-rtems*): Likewise.
11755 (bfin*-rtems*): Likewise.
11756 (moxie-*-rtems*): Likewise.
11757 (h8300-*-rtems*): Likewise.
11758 (i[34567]86-*-rtems*): Likewise.
11759 (lm32-*-rtems*): Likewise.
11760 (m32r-*-rtems*): Likewise.
11761 (m68k-*-rtems*): Likewise.
11762 (microblaze*-*-rtems*): Likewise.
11763 (mips*-*-rtems*): Likewise.
11764 (powerpc-*-rtems*): Likewise.
11765 (sh-*-rtems*): Likewise.
11766 (sparc-*-rtems*): Likewise.
11767 (sparc64-*-rtems*): Likewise.
11768 (v850-*-rtems*): Likewise.
11769 (m32c-*-rtems*): Likewise.
11771 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11773 PR rtl-optimization/59511
11774 * ira.c (ira_init_register_move_cost): Use memory costs for some
11775 cases of register move cost calculations.
11776 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11777 instead of BB frequency.
11778 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11779 * lra-assigns.c (find_hard_regno_for): Ditto.
11781 2014-01-15 Richard Biener <rguenther@suse.de>
11783 PR tree-optimization/59822
11784 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11785 (vectorizable_load): Use it to hoist defs of uses of invariant
11786 loads out of the loop.
11788 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11789 Kugan Vivekanandarajah <kuganv@linaro.org>
11792 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11795 2014-01-15 Richard Biener <rguenther@suse.de>
11797 PR rtl-optimization/59802
11798 * lcm.c (compute_available): Use inverted postorder to seed
11799 the initial worklist.
11801 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11804 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11805 ADDR_REGS for invalid symrefs in non-PIC code.
11807 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11810 * builtins.c (determine_block_size): Initialize *probable_max_size
11811 even if len_rtx is CONST_INT.
11813 2014-01-14 Andrew Pinski <apinski@cavium.com>
11815 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11816 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11817 (cortexa53_tunings): Likewise.
11818 (aarch64_sched_issue_rate): New function.
11819 (TARGET_SCHED_ISSUE_RATE): Define.
11821 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11823 * ira-costs.c (find_costs_and_classes): Add missed
11824 ira_init_register_move_cost_if_necessary.
11826 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11829 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11831 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11834 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11835 to indicate if type is used for function return value. Warn ABI
11836 change if the vector mode isn't available for function return value.
11837 (ix86_function_arg_advance): Pass false to type_natural_mode.
11838 (ix86_function_arg): Likewise.
11839 (ix86_gimplify_va_arg): Likewise.
11840 (function_arg_32): Don't warn ABI change.
11841 (ix86_function_value): Pass true to type_natural_mode.
11842 (ix86_return_in_memory): Likewise.
11843 (ix86_struct_value_rtx): Removed.
11844 (TARGET_STRUCT_VALUE_RTX): Likewise.
11846 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11848 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11849 converting a conditional jump into a conditional return.
11851 2014-01-14 Richard Biener <rguenther@suse.de>
11853 PR tree-optimization/58921
11854 PR tree-optimization/59006
11855 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11856 hoisting invariant stmts.
11857 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11858 invariant loads on the preheader edge if possible.
11860 2014-01-14 Joey Ye <joey.ye@arm.com>
11862 * doc/plugin.texi (Building GCC plugins): Update to C++.
11864 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11866 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11867 (_mm_rcp28_round_ss): Ditto.
11868 (_mm_rsqrt28_round_sd): Ditto.
11869 (_mm_rsqrt28_round_ss): Ditto.
11870 (_mm_rcp28_sd): Ditto.
11871 (_mm_rcp28_ss): Ditto.
11872 (_mm_rsqrt28_sd): Ditto.
11873 (_mm_rsqrt28_ss): Ditto.
11874 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
11875 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
11876 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
11877 (IX86_BUILTIN_RCP28SD): Ditto.
11878 (IX86_BUILTIN_RCP28SS): Ditto.
11879 (IX86_BUILTIN_RSQRT28SD): Ditto.
11880 (IX86_BUILTIN_RSQRT28SS): Ditto.
11881 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
11882 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
11883 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
11884 (ix86_expand_special_args_builtin): Expand new FTYPE.
11885 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
11886 (srcp14<mode>): Make insn unary.
11887 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
11888 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
11889 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
11890 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
11891 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
11892 Fix rounding: make it SAE only.
11893 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
11895 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
11897 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
11898 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
11899 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
11900 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
11901 (round_saeonly_mask_scalar_operand4): Ditto.
11902 (round_saeonly_mask_scalar_op3): Ditto.
11903 (round_saeonly_mask_scalar_op4): Ditto.
11905 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11907 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
11908 Implement -maltivec=be for vec_insert and vec_extract.
11910 2014-01-10 DJ Delorie <dj@redhat.com>
11912 * config/msp430/msp430.md (call_internal): Don't allow memory
11913 references with SP as the base register.
11914 (call_value_internal): Likewise.
11915 * config/msp430/constraints.md (Yc): New. For memory references
11916 that don't use SP as a base register.
11918 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
11919 "an integer without a # prefix"
11920 * config/msp430/msp430.md (epilogue_helper): Use it.
11922 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11925 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
11926 AVX512F gather builtins.
11927 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
11928 on gather decls with INTEGER_TYPE masktype.
11929 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
11930 directly into the builtin rather than hoisting it before loop.
11932 PR tree-optimization/59387
11933 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
11934 (scev_const_prop): If folded_casts and type has undefined overflow,
11935 use force_gimple_operand instead of force_gimple_operand_gsi and
11936 for each added stmt if it is assign with
11937 arith_code_with_undefined_signed_overflow, call
11938 rewrite_to_defined_overflow.
11939 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
11940 gimple-fold.h instead.
11941 (arith_code_with_undefined_signed_overflow,
11942 rewrite_to_defined_overflow): Moved to ...
11943 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
11944 rewrite_to_defined_overflow): ... here. No longer static.
11945 Include gimplify-me.h.
11946 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
11947 rewrite_to_defined_overflow): New prototypes.
11949 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11951 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
11953 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
11955 * builtins.c (get_object_alignment_2): Minor tweak.
11956 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
11958 2014-01-13 Christian Bruel <christian.bruel@st.com>
11960 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
11961 optimized non constant lengths.
11963 2014-01-13 Jakub Jelinek <jakub@redhat.com>
11966 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
11967 load as __atomic_load_N if possible.
11969 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11971 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
11973 (rs6000_expand_builtin): Adjust call.
11975 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
11978 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
11979 * config/rs6000/rs6000.c: Include target-globals.h.
11980 (rs6000_set_current_function): Instead of doing target_reinit
11981 unconditionally, use save_target_globals_default_opts and
11982 restore_target_globals.
11984 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
11986 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
11987 (rs6000_expand_builtin): Handle mffs and mtfsf.
11988 (rs6000_init_builtins): Define mffs and mtfsf.
11989 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
11990 (rs6000_mffs): New pattern.
11991 (rs6000_mtfsf): New pattern.
11993 2014-01-11 Bin Cheng <bin.cheng@arm.com>
11995 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
11996 Start narrowing with START. Apply candidate-use pair
11997 and check overall cost in narrowing.
11998 (iv_ca_prune): Pass new argument.
12000 2014-01-10 Jeff Law <law@redhat.com>
12002 PR middle-end/59743
12003 * ree.c (combine_reaching_defs): Ensure the defining statement
12004 occurs before the extension when optimizing extensions with
12005 different source and destination hard registers.
12007 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12010 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12011 vtables into the type inheritance graph.
12013 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12015 PR rtl-optimization/59754
12016 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12017 modes in the REGNO != REGNO case.
12019 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12021 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12023 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12025 PR tree-optimization/59745
12026 * tree-predcom.c (tree_predictive_commoning_loop): Call
12027 free_affine_expand_cache if giving up because components is NULL.
12029 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12030 GC in payload of target_globals struct instead of allocating them on
12031 the heap and the larger structs separately using GC.
12032 * target-globals.h (struct target_globals): Make regs, hard_regs,
12033 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12034 of GTY((skip)) and change type to void *.
12035 (reset_target_globals): Cast loads from those fields to corresponding
12038 2014-01-10 Steve Ellcey <sellcey@mips.com>
12041 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12042 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12043 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12045 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12048 * aarch64-modes.def (CC_Zmode): New flags mode.
12049 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12050 represents an equality.
12051 (aarch64_get_condition_code): Handle CC_Zmode.
12052 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12054 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12056 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12057 extraction in good case.
12059 2014-01-10 Richard Biener <rguenther@suse.de>
12061 PR tree-optimization/59374
12062 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12063 checking after SLP discovery. Mark stmts not participating
12064 in any SLP instance properly.
12066 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12068 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12069 when handling a SET rtx.
12071 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12073 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12074 (cortex-a57): Likewise.
12075 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12077 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12079 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12080 non-iwmmxt builtins.
12082 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12086 * ipa-devirt.c record_target_from_binfo): Take as argument
12087 stack of binfos and lookup matching one for virtual inheritance.
12088 (possible_polymorphic_call_targets_1): Update.
12090 2014-01-10 Huacai Chen <chenhc@lemote.com>
12092 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12093 kernel strings for Loongson-2E/2F/3A.
12095 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12097 PR middle-end/59670
12098 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12099 is_gimple_call before calling gimple_call_internal_p.
12101 2014-01-09 Steve Ellcey <sellcey@mips.com>
12103 * Makefile.in (TREE_FLOW_H): Remove.
12104 (TREE_SSA_H): Add file names from tree-flow.h.
12105 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12106 * tree.h: Remove tree-flow.h reference.
12107 * hash-table.h: Remove tree-flow.h reference.
12108 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12109 reference with tree-ssa-loop.h.
12111 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12113 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12114 default element-order behavior for -maltivec.
12115 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12116 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12117 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12118 when targeting big endian, at least for now.
12119 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12121 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12123 PR middle-end/47735
12124 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12125 var satisfies use_register_for_decl, just take into account type
12126 alignment, rather than decl alignment.
12128 PR tree-optimization/59622
12129 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12130 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12131 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12132 Don't devirtualize for inplace at all. For targets.length () == 1,
12133 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12135 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12137 * config/i386/i386.md (cpu): Remove the unused btver1.
12139 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12141 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12143 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12146 * tree-core.h (struct target_globals): New forward declaration.
12147 (struct tree_target_option): Add globals field.
12148 * tree.h (TREE_TARGET_GLOBALS): Define.
12149 (prepare_target_option_nodes_for_pch): New prototype.
12150 * target-globals.h (struct target_globals): Define even if
12151 !SWITCHABLE_TARGET.
12152 * tree.c (prepare_target_option_node_for_pch,
12153 prepare_target_option_nodes_for_pch): New functions.
12154 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12155 * config/i386/i386.c: Include target-globals.h.
12156 (ix86_set_current_function): Instead of doing target_reinit
12157 unconditionally, use save_target_globals_default_opts and
12158 restore_target_globals.
12160 2014-01-09 Richard Biener <rguenther@suse.de>
12162 PR tree-optimization/59715
12163 * tree-cfg.h (split_critical_edges): Declare.
12164 * tree-cfg.c (split_critical_edges): Export.
12165 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12167 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12169 * cfgexpand.c (expand_stack_vars): Optionally disable
12170 asan stack protection.
12171 (expand_used_vars): Likewise.
12172 (partition_stack_vars): Likewise.
12173 * asan.c (asan_emit_stack_protection): Optionally disable
12174 after return stack usage.
12175 (instrument_derefs): Optionally disable memory access instrumentation.
12176 (instrument_builtin_call): Likewise.
12177 (instrument_strlen_call): Likewise.
12178 (asan_protect_global): Optionally disable global variables protection.
12179 * doc/invoke.texi: Added doc for new options.
12180 * params.def: Added new options.
12181 * params.h: Likewise.
12183 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12185 PR rtl-optimization/59724
12186 * ifcvt.c (cond_exec_process_if_block): Don't call
12187 flow_find_head_matching_sequence with 0 longest_match.
12188 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12189 non-active insns if !stop_after.
12190 (try_head_merge_bb): Revert 2014-01-07 changes.
12192 2014-01-08 Jeff Law <law@redhat.com>
12194 * ree.c (get_sub_rtx): New function, extracted from...
12195 (merge_def_and_ext): Here.
12196 (combine_reaching_defs): Use get_sub_rtx.
12198 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12200 * cgraph.h (varpool_variable_node): Do not choke on null node.
12202 2014-01-08 Catherine Moore <clm@codesourcery.com>
12204 * config/mips/mips.md (simple_return): Attempt to use JRC
12206 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12208 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12210 PR rtl-optimization/59137
12211 * reorg.c (steal_delay_list_from_target): Call update_block for
12213 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12215 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12217 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12218 two duplicate entries.
12220 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12223 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12225 * config/mips/mips.c (mips_truncated_op_cost): New function.
12226 (mips_rtx_costs): Adjust test for BADDU.
12227 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12229 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12231 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12232 (*baddu_si): ...this new pattern.
12234 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12237 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12239 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12241 PR middle-end/57748
12242 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12244 (expand_expr, expand_normal): Adjust.
12245 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12246 inner_reference_p. Use inner_reference_p to expand inner references.
12247 (store_expr): Adjust.
12248 * cfgexpand.c (expand_call_stmt): Adjust.
12250 2014-01-08 Rong Xu <xur@google.com>
12252 * gcov-io.c (gcov_var): Move from gcov-io.h.
12253 (gcov_position): Ditto.
12254 (gcov_is_error): Ditto.
12255 (gcov_rewrite): Ditto.
12256 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12257 only part to libgcc/libgcov.h.
12259 2014-01-08 Marek Polacek <polacek@redhat.com>
12261 PR middle-end/59669
12262 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12264 2014-01-08 Marek Polacek <polacek@redhat.com>
12267 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12269 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12271 PR rtl-optimization/59649
12272 * stor-layout.c (get_mode_bounds): For BImode return
12273 0 and STORE_FLAG_VALUE.
12275 2014-01-08 Richard Biener <rguenther@suse.de>
12277 PR middle-end/59630
12278 * gimple.h (is_gimple_builtin_call): Remove.
12279 (gimple_builtin_call_types_compatible_p): New.
12280 (gimple_call_builtin_p): New overload.
12281 * gimple.c (is_gimple_builtin_call): Remove.
12282 (validate_call): Rename to ...
12283 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12284 check return types.
12285 (validate_type): New static function.
12286 (gimple_call_builtin_p): New overload and adjust.
12287 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12288 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12289 (gimple_fold_stmt_to_constant_1): Likewise.
12290 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12292 2014-01-08 Richard Biener <rguenther@suse.de>
12294 PR middle-end/59471
12295 * gimplify.c (gimplify_expr): Gimplify register-register type
12296 VIEW_CONVERT_EXPRs to separate stmts.
12298 2014-01-07 Jeff Law <law@redhat.com>
12300 PR middle-end/53623
12301 * ree.c (combine_set_extension): Handle case where source
12302 and destination registers in an extension insn are different.
12303 (combine_reaching_defs): Allow source and destination registers
12304 in extension to be different under limited circumstances.
12305 (add_removable_extension): Remove restriction that the
12306 source and destination registers in the extension are the same.
12307 (find_and_remove_re): Emit a copy from the extension's
12308 destination to its source after the defining insn if
12309 the source and destination registers are different.
12311 PR middle-end/59285
12312 * ifcvt.c (merge_if_block): If we are merging a block with more than
12313 one successor with a block with no successors, remove any BARRIER
12314 after the second block.
12316 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12318 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12320 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12323 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12324 for 14-bit register offsets when INT14_OK_STRICT is false.
12326 2014-01-07 Roland Stigge <stigge@antcom.de>
12327 Michael Meissner <meissner@linux.vnet.ibm.com>
12330 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12331 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12333 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12335 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12338 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12340 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12341 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12342 rtx is const0_rtx or not.
12344 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12347 * target-globals.c (save_target_globals): Remove this_fn_optab
12349 * toplev.c: Include optabs.h.
12350 (target_reinit): Temporarily restore the global options if another
12351 set of options are in force.
12353 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12355 PR rtl-optimization/58668
12356 * cfgcleanup.c (flow_find_cross_jump): Don't count
12357 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12358 to determine what is counted.
12359 (flow_find_head_matching_sequence): Use active_insn_p to determine
12361 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12363 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12364 determine what is counted.
12366 PR tree-optimization/59643
12367 * tree-predcom.c (split_data_refs_to_components): If one dr is
12368 read and one write, determine_offset fails and the write isn't
12369 in the bad component, just put the read into the bad component.
12371 2014-01-07 Mike Stump <mikestump@comcast.net>
12372 Jakub Jelinek <jakub@redhat.com>
12375 * tree-core.h (struct tree_optimization_option): Change optabs
12376 type from unsigned char * to void *.
12377 * optabs.c (init_tree_optimization_optabs): Adjust
12378 TREE_OPTIMIZATION_OPTABS initialization.
12380 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12383 * config/i386/i386.h (struct machine_function): Add
12384 no_drap_save_restore field.
12385 * config/i386/i386.c (ix86_save_reg): Use
12386 !cfun->machine->no_drap_save_restore instead of
12387 crtl->stack_realign_needed.
12388 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12389 this function clears frame_pointer_needed. Set
12390 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12391 and DRAP reg is needed.
12393 2014-01-06 Marek Polacek <polacek@redhat.com>
12396 * doc/implement-c.texi: Mention that other integer types are
12397 permitted as bit-field types in strictly conforming mode.
12399 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12401 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12402 is newly allocated.
12404 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12406 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12408 2014-01-06 Martin Jambor <mjambor@suse.cz>
12411 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12413 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12415 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12419 * var-tracking.c (add_stores): Preserve the value of the source even if
12420 we don't record the store.
12422 2014-01-06 Terry Guo <terry.guo@arm.com>
12424 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12426 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12429 * config/darwin.c (darwin_function_section): Adjust return values to
12430 correspond to optimisation changes made in r206070.
12432 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12434 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12435 from prefetch_block tune setting.
12436 (nocona_cost): Correct size of prefetch block to 64.
12438 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12440 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12441 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12442 used to save the static chain register in the computation of the offset
12443 from which the FP registers need to be restored.
12445 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12447 PR tree-optimization/59519
12448 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12449 ICE if get_current_def (current_new_name) is already non-NULL, as long
12450 as it is a phi result of some other phi in *new_exit_bb that has
12453 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12454 or vmovdqu* for misaligned_operand.
12455 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12456 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12457 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12458 aligned_mem for AVX512F masked aligned load and store builtins and for
12459 non-temporal moves.
12461 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12463 PR tree-optimization/59651
12464 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12465 Address range for negative step should be added by TYPE_SIZE_UNIT.
12467 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12469 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12470 overlapping registers also for registers other than the stack pointer.
12472 2014-01-03 Marek Polacek <polacek@redhat.com>
12475 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12478 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12481 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12484 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12485 (push splitter): Use <P:MODE_SIZE> instead of
12486 GET_MODE_SIZE (<P:MODE>mode).
12487 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12488 (mov -1, reg peephole2): Likewise.
12489 * config/i386/sse.md (*mov<mode>_internal,
12490 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12491 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12492 *<code><mode>3, *andnot<mode>3<mask_name>,
12493 <mask_codefor><code><mode>3<mask_name>): Likewise.
12494 * config/i386/subst.md (mask_mode512bit_condition,
12495 sd_mask_mode512bit_condition): Likewise.
12497 2014-01-02 Xinliang David Li <davidxl@google.com>
12499 PR tree-optimization/59303
12500 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12501 (dump_predicates): Better output format.
12502 (pred_equal_p): New function.
12503 (is_neq_relop_p): Ditto.
12504 (is_neq_zero_form_p): Ditto.
12505 (pred_expr_equal_p): Ditto.
12506 (pred_neg_p): Ditto.
12507 (simplify_pred): Ditto.
12508 (simplify_preds_2): Ditto.
12509 (simplify_preds_3): Ditto.
12510 (simplify_preds_4): Ditto.
12511 (simplify_preds): Ditto.
12512 (push_pred): Ditto.
12513 (push_to_worklist): Ditto.
12514 (get_pred_info_from_cmp): Ditto.
12515 (is_degenerated_phi): Ditto.
12516 (normalize_one_pred_1): Ditto.
12517 (normalize_one_pred): Ditto.
12518 (normalize_one_pred_chain): Ditto.
12519 (normalize_preds): Ditto.
12520 (normalize_cond_1): Remove function.
12521 (normalize_cond): Ditto.
12522 (is_gcond_subset_of): Ditto.
12523 (is_subset_of_any): Ditto.
12524 (is_or_set_subset_of): Ditto.
12525 (is_and_set_subset_of): Ditto.
12526 (is_norm_cond_subset_of): Ditto.
12527 (pred_chain_length_cmp): Ditto.
12528 (convert_control_dep_chain_into_preds): Type change.
12529 (find_predicates): Ditto.
12530 (find_def_preds): Ditto.
12531 (destroy_predicates_vecs): Ditto.
12532 (find_matching_predicates_in_rest_chains): Ditto.
12533 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12534 (is_pred_expr_subset): Ditto.
12535 (is_pred_chain_subset_of): Ditto.
12536 (is_included_in): Ditto.
12537 (is_superset_of): Ditto.
12539 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12541 Update copyright years.
12543 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12545 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12546 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12547 config/arc/arc.md, config/arc/arc.opt,
12548 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12549 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12550 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12551 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12552 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12553 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12554 vtable-verify.c, vtable-verify.h: Use the standard form for the
12557 2014-01-02 Tobias Burnus <burnus@net-b.de>
12559 * gcc.c (process_command): Update copyright notice dates.
12560 * gcov-dump.c: Ditto.
12562 * doc/cpp.texi: Bump @copying's copyright year.
12563 * doc/cppinternals.texi: Ditto.
12564 * doc/gcc.texi: Ditto.
12565 * doc/gccint.texi: Ditto.
12566 * doc/gcov.texi: Ditto.
12567 * doc/install.texi: Ditto.
12568 * doc/invoke.texi: Ditto.
12570 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12572 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12574 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12576 * config/i386/sse.md (*mov<mode>_internal): Guard
12577 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12579 PR rtl-optimization/59647
12580 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12581 new_rtx into UNSIGNED_FLOAT rtxes.
12583 Copyright (C) 2014 Free Software Foundation, Inc.
12585 Copying and distribution of this file, with or without modification,
12586 are permitted in any medium without royalty provided the copyright
12587 notice and this notice are preserved.