1 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
3 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
4 predicate to detect a negative quotient.
6 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
8 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
9 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
10 Add X - Y CMP 0 to X CMP Y transformation.
11 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
13 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
15 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
18 2014-05-27 Steve Ellcey <sellcey@mips.com>
20 * config/mips/mips.c: Add include of cgraph.h.
22 2014-05-27 Richard Biener <rguenther@suse.de>
24 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
26 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
29 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
30 * config/arm/arm-cores.def (License): Same.
31 * config/arm/arm-opts.h (License): Same.
32 * config/arm/aout.h (License): Same.
33 * config/arm/bpabi.h (License): Same.
34 * config/arm/elf.h (License): Same.
35 * config/arm/linux-elf.h (License): Same.
36 * config/arm/linux-gas.h (License): Same.
37 * config/arm/netbsd-elf.h (License): Same.
38 * config/arm/uclinux-eabi.h (License): Same.
39 * config/arm/uclinux-elf.h (License): Same.
40 * config/arm/vxworks.h (License): Same.
42 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
44 * config/arm/neon.md (neon_bswap<mode>): New pattern.
45 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
46 (arm_init_neon_builtins): Handle NEON_BSWAP.
47 Define required type nodes.
48 (arm_expand_neon_builtin): Handle NEON_BSWAP.
49 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
50 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
51 * config/arm/iterators.md (VDQHSD): New mode iterator.
53 2014-05-27 Richard Biener <rguenther@suse.de>
55 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
56 Try using literal operands when comparing value-ranges failed.
58 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
60 * ira.c (commutative_operand): Adjust for change to recog_data.
61 [Missing from previous commit.]
63 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
65 * system.h (TEST_BIT): New macro.
66 * recog.h (alternative_mask): New type.
67 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
68 (recog_data_d): Replace alternative_enabled_p array with
70 (target_recog): New structure.
71 (default_target_recog, this_target_recog): Declare.
72 (get_enabled_alternatives, recog_init): Likewise.
73 * recog.c (default_target_recog, this_target_recog): New variables.
74 (get_enabled_alternatives): New function.
75 (extract_insn): Use it.
76 (recog_init): New function.
77 (preprocess_constraints, constrain_operands): Adjust for change to
79 * postreload.c (reload_cse_simplify_operands): Likewise.
80 * reload.c (find_reloads): Likewise.
81 * ira-costs.c (record_reg_classes): Likewise.
82 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
83 all alternatives after a disabled one would be skipped.
84 (ira_implicitly_set_insn_hard_regs): Likewise.
85 * ira.c (ira_setup_alts): Adjust for change to recog_data.
86 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
87 with enabled_alternatives.
88 * lra.c (free_insn_recog_data): Update accordingly.
89 (lra_update_insn_recog_data): Likewise.
90 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
91 * lra-constraints.c (process_alt_operands): Likewise. Handle
92 only_alternative as part of the enabled mask.
93 * target-globals.h (this_target_recog): Declare.
94 (target_globals): Add a recog field.
95 (restore_target_globals): Restore this_target_recog.
96 * target-globals.c: Include recog.h.
97 (default_target_globals): Initialize recog field.
98 (save_target_globals): Likewise.
99 * reginfo.c (reinit_regs): Call recog_init.
100 * toplev.c (backend_init_target): Likewise.
102 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
104 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
105 rather than any named insn's code.
107 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
110 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
111 * config/arm/arm-cores.def (License): Same.
113 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
115 * tree.h (decl_comdat_group): Declare.
116 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
117 * tree.c (decl_comdat_group): Here.
119 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
121 PR rtl-optimization/61222
122 * combine.c (simplify_shift_const_1): When moving a PLUS outside
123 the shift, truncate the PLUS operand to the result mode.
125 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
128 * config/i386/i386.c (ix86_rtx_costs)
129 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
132 2014-05-26 Martin Jambor <mjambor@suse.cz>
134 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
137 2014-05-26 Richard Biener <rguenther@suse.de>
139 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
140 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
141 Provide specializations.
142 (wi::int_traits <HOST_WIDE_INT>,
143 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
145 2014-05-26 Alan Modra <amodra@gmail.com>
148 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
149 params and return a bool. Remove dead code. Update comment.
150 Assert we have a const_int source. Remove bogus code from
151 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
152 handling of constants > 2G and reg_equal note, from..
153 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
154 return value. Update comment. If we can, use a new pseudo
155 for intermediate calculations.
156 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
158 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
159 call to rs6000_emit_set_const in splitter.
160 (movdi_internal64+2, +3): Likewise.
162 2014-05-26 Richard Biener <rguenther@suse.de>
164 * system.h: Define __STDC_FORMAT_MACROS before
165 including inttypes.h.
166 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
167 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
168 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
169 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
170 HOST_WIDEST_INT_C): Remove.
171 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
172 if C99 inttypes.h is not available.
173 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
174 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
175 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
176 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
177 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
178 (struct output_info): Likewise.
179 (print_statistics): Adjust.
180 (dump_bitmap_statistics): Likewise.
181 * bt-load.c (migrate_btr_defs): Print with PRId64.
182 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
183 (MAX_SAFE_MULTIPLIER): Adjust.
184 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
185 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
186 dump_cgraph_node): Likewise.
187 * final.c (dump_basic_block_info): Likewise.
188 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
189 * gcov.c (format_gcov): Likewise.
190 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
192 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
193 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
194 (inline_small_functions, dump_overall_stats, dump_inline_stats):
195 Use PRId64 for dumping.
196 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
197 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
198 (add_allocno_hard_regs): Adjust.
199 * loop-doloop.c (doloop_modify): Print using PRId64.
200 * loop-iv.c (inverse): Compute in uint64_t.
201 (determine_max_iter, iv_number_of_iterations): Likewise.
202 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
204 * lto-streamer-out.c (write_symbol): Use uint64_t.
205 * mcf.c (CAP_INFINITY): Use int64_t maximum.
206 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
207 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
208 * modulo-sched.c (const_iteration_count): Use int64_t.
209 (sms_schedule): Dump using PRId64.
210 * predict.c (dump_prediction): Likewise.
211 * pretty-print.h (pp_widest_integer): Remove.
212 * profile.c (get_working_sets, is_edge_inconsistent,
213 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
214 * tree-pretty-print.c (pp_double_int): Remove case handling
215 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
216 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
218 (pass_optimize_bswap::execute): Remove restriction on hosts.
219 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
220 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
221 * tree.c (widest_int_cst_value): Remove.
222 * tree.h (widest_int_cst_value): Likewise.
223 * value-prof.c (dump_histogram_value): Print using PRId64.
224 * gengtype.c (main): Also inject int64_t.
225 * ggc-page.c (struct max_alignment): Use int64_t.
226 * alloc-pool.c (struct allocation_object_def): Likewise.
227 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
229 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
230 * doc/tm.texi: Regenerated.
231 * gengtype-lex.l (IWORD): Handle [u]int64_t.
232 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
233 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
234 mmix_output_register_setting): Use [u]int64_t in prototypes.
235 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
236 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
237 mmix_output_octa, mmix_output_shifted_value): Adjust.
238 (mmix_intval): Adjust. Remove unreachable case.
239 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
241 2014-05-26 Richard Biener <rguenther@suse.de>
243 * configure.ac: Drop __int64 type check. Insist that we
244 found uint64_t and int64_t.
245 * hwint.h (HOST_BITS_PER___INT64): Remove.
246 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
247 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
248 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
249 (HOST_WIDEST_FAST_INT): Remove __int64 case.
250 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
251 for dst_q_src_df_rms_cdt.
252 * configure: Regenerate.
253 * config.in: Likewise.
255 2014-05-26 Michael Tautschnig <mt@debian.org>
258 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
259 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
261 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
263 PR rtl-optimization/61278
264 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
266 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
268 PR rtl-optimization/61220
269 Part of PR rtl-optimization/61225
270 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
271 insn; skip split_edge for a block with only one successor.
273 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
275 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
278 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
280 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
281 (update_vtable_references): New function.
282 (function_and_variable_visibility): Rewrite also vtable initializers.
283 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
285 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
287 * ggc.h (ggc_grow): New function.
288 * ggc-none.c (ggc_grow): New function.
289 * ggc-page.c (ggc_grow): Likewise.
291 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
293 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
294 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
295 comdat_can_be_unshared_p, cgraph_externally_visible_p,
296 varpool_externally_visible_p, can_replace_by_local_alias,
297 update_visibility_by_resolution_info, function_and_variable_visibility,
298 pass_data_ipa_function_and_variable_visibility,
299 make_pass_ipa_function_and_variable_visibility,
300 whole_program_function_and_variable_visibility,
301 pass_data_ipa_whole_program_visibility,
302 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
303 * cgraph.h (cgraph_local_node_p): Declare.
304 * ipa-visibility.c: New file.
305 * Makefile.in (OBJS): Add ipa-visiblity.o
307 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
309 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
310 that var decl is available.
312 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
314 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
316 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
317 (find_decls_types_r): Do not walk COMDAT_GROUP.
318 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
319 * varasm.c (make_decl_one_only): Use set_comdat_group;
320 create node if needed.
321 * ipa-inline-transform.c (save_inline_function_body): Update
322 way we decl->symtab mapping.
323 * symtab.c (symtab_hash, hash_node, eq_node
324 symtab_insert_node_to_hashtable): Remove.
325 (symtab_register_node): Update.
326 (symtab_unregister_node): Update.
327 (symtab_get_node): Reimplement as inline function.
328 (symtab_add_to_same_comdat_group): Update.
329 (symtab_dissolve_same_comdat_group_list): Update.
330 (dump_symtab_base): Update.
331 (verify_symtab_base): Update.
332 (symtab_make_decl_local): Update.
333 (fixup_same_cpp_alias_visibility): Update.
334 (symtab_nonoverwritable_alias): Update.
335 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
336 * ipa.c (update_visibility_by_resolution_info): UPdate.
337 * bb-reorder.c: Include cgraph.h
338 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
340 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
341 * cgraph.c (cgraph_get_create_node): Update.
342 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
344 (symtab_get_node): Make inline.
345 (symtab_insert_node_to_hashtable): Remove.
346 (symtab_can_be_discarded): Update.
347 (decl_comdat_group): New function.
348 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
350 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
352 (read_comdat_group): New function.
353 (input_node, input_varpool_node): Use it.
354 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
356 * mips.c (mips_start_unique_function): Likewise.
357 (ix86_code_end): Likewise.
358 (rs6000_code_end): Likweise.
359 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
361 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
363 * gengtype-state.c (fatal_reading_state): Bring offline.
364 * optabs.c (widening_optab_handler): Bring offline.
365 * optabs.h (widening_optab_handler): Likewise.
366 * final.c (get_attr_length_1): Likewise.
368 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
370 * sched-int.h (sd_iterator_cond): Manually tail recurse.
372 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
374 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
375 (ppc440-compare): Include shift with dot.
376 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
377 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
379 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
381 (e6500_sfx2): Include it.
382 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
383 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
384 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
385 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
386 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
387 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
388 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
389 *lshiftrt_internal1le, *lshiftrt_internal1be,
390 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
391 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
392 *rotldi3_internal10le, *rotldi3_internal10be,
393 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
394 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
395 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
396 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
397 define_insns): Use type "shift" in the appropriate alternatives.
399 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
401 * config/rs6000/rs6000.md (type): Add "logical". Delete
403 (dot): Adjust comment.
404 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
405 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
406 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
407 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
408 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
409 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
410 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
411 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
413 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
414 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
415 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
416 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
417 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
418 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
419 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
420 * config/rs6000/8540.md (ppc8540_su): Adjust.
421 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
422 cell-cmp-microcoded): Adjust.
423 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
424 * config/rs6000/e500mc.md (e500mc_su): Adjust.
425 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
426 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
427 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
428 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
429 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
430 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
431 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
433 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
434 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
435 Adjust. Adjust comment.
436 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
437 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
439 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
441 * config/rs6000/rs6000.md (type): Add "add".
442 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
443 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
444 define_insns): Use it.
445 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
447 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
448 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
449 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
450 * config/rs6000/601.md (ppc601-integer): Adjust.
451 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
452 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
453 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
454 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
455 * config/rs6000/8540.md (ppc8540_su): Adjust.
456 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
457 cell-cmp-microcoded): Adjust.
458 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
459 * config/rs6000/e500mc.md (e500mc_su): Adjust.
460 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
461 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
462 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
463 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
464 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
465 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
466 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
468 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
469 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
470 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
471 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
473 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
475 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
476 "delayed_compare", "var_delayed_compare".
477 (var_shift): New attribute.
478 (cell_micro): Adjust.
479 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
480 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
481 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
482 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
483 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
484 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
485 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
486 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
487 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
488 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
489 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
490 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
491 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
492 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
493 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
494 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
495 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
496 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
497 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
498 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
499 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
500 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
501 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
502 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
503 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
505 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
506 * config/rs6000/440.md (ppc440-integer): Adjust.
507 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
508 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
509 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
510 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
511 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
512 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
513 * config/rs6000/8540.md (ppc8540_su): Adjust.
514 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
515 cell-cmp-microcoded): Adjust.
516 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
517 * config/rs6000/e500mc.md (e500mc_su): Adjust.
518 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
519 e500mc64_delayed): Adjust.
520 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
521 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
522 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
523 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
524 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
525 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
526 power6-delayed-compare, power6-var-delayed-compare): Adjust.
527 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
528 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
530 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
531 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
533 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
535 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
536 (bits): New mode_attr.
537 (idiv_ldiv): Delete mode_attr.
538 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
539 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
540 rs6000_adjust_priority, is_nonpipeline_insn,
541 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
543 * config/rs6000/40x.md (ppc403-idiv): Adjust.
544 * config/rs6000/440.md (ppc440-idiv): Adjust.
545 * config/rs6000/476.md (ppc476-idiv): Adjust.
546 * config/rs6000/601.md (ppc601-idiv): Adjust.
547 * config/rs6000/603.md (ppc603-idiv): Adjust.
548 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
549 ppc620-ldiv): Adjust.
550 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
551 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
552 * config/rs6000/8540.md (ppc8540_divide): Adjust.
553 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
554 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
555 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
556 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
557 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
558 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
559 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
560 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
561 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
562 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
563 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
564 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
565 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
566 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
567 * config/rs6000/titan.md (titan_fxu_div): Adjust.
569 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
571 * config/rs6000/rs6000.md (type): Delete "insert_word",
572 "insert_dword". Add "insert".
573 (size): Update comment.
574 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
575 insn_must_be_first_in_group): Adjust.
576 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
577 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
578 *insvsi_internal6, insvdi_internal): Adjust.
580 * config/rs6000/40x.md (ppc403-integer): Adjust.
581 * config/rs6000/440.md (ppc440-integer): Adjust.
582 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
583 * config/rs6000/601.md (ppc601-integer): Adjust.
584 * config/rs6000/603.md (ppc603-integer): Adjust.
585 * config/rs6000/6xx.md (ppc604-integer): Adjust.
586 * config/rs6000/7450.md (ppc7450-integer): Adjust.
587 * config/rs6000/7xx.md (ppc750-integer): Adjust.
588 * config/rs6000/8540.md (ppc8540_su): Adjust.
589 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
590 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
591 * config/rs6000/e500mc.md (e500mc_su): Adjust.
592 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
593 * config/rs6000/e5500.md (e5500_sfx): Adjust.
594 * config/rs6000/e6500.md (e6500_sfx): Adjust.
595 * config/rs6000/mpc.md (mpccore-integer): Adjust.
596 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
597 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
598 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
599 * config/rs6000/power7.md (power7-integer): Adjust.
600 * config/rs6000/power8.md (power8-1cyc): Adjust.
601 * config/rs6000/rs64.md (rs64a-integer): Adjust.
602 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
604 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
606 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
607 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
608 (size): New attribute.
609 (dot): New attribute.
610 (cell_micro): Adjust.
611 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
612 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
613 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
614 umuldi3_highpart): Adjust.
615 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
616 rs6000_adjust_priority, is_nonpipeline_insn,
617 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
619 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
620 ppc405-imul3): Adjust.
621 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
622 * config/rs6000/476.md (ppc476-imul): Adjust.
623 * config/rs6000/601.md (ppc601-imul): Adjust.
624 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
625 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
626 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
627 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
628 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
630 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
631 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
632 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
634 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
635 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
636 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
637 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
638 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
639 * config/rs6000/mpc.md (mpccore-imul): Adjust.
640 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
641 power4-lmul, power4-imul, power4-imul3): Adjust.
642 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
643 power5-lmul, power5-imul, power5-imul3): Adjust.
644 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
645 power6-lmul, power6-imul, power6-imul3): Adjust.
646 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
647 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
649 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
651 * config/rs6000/titan.md (titan_imul): Adjust.
653 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
655 * config/rs6000/rs6000.md (type): Add new value "halfmul".
656 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
657 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
658 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
659 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
660 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
661 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
662 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
663 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
664 * config/rs6000/titan.md: Delete nonsensical comment.
665 (titan_imul): Add type imul3.
666 (titan_mulhw): Remove type imul3; add type halfmul.
668 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
670 * config/rs6000/rs6000.md (type): Reorder, reformat.
672 2014-05-23 Martin Jambor <mjambor@suse.cz>
674 PR tree-optimization/53787
675 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
676 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
677 analysis_done, update all uses.
678 * ipa-prop.c: Include domwalk.h
679 (param_analysis_info): Removed.
680 (param_aa_status): New type.
681 (ipa_bb_info): Likewise.
682 (func_body_info): Likewise.
683 (ipa_get_bb_info): New function.
684 (aa_overwalked): Likewise.
685 (find_dominating_aa_status): Likewise.
686 (parm_bb_aa_status_for_bb): Likewise.
687 (parm_preserved_before_stmt_p): Changed to use new param AA info.
688 (load_from_unmodified_param): Accept func_body_info as a parameter
689 instead of parms_ainfo.
690 (parm_ref_data_preserved_p): Changed to use new param AA info.
691 (parm_ref_data_pass_through_p): Likewise.
692 (ipa_load_from_parm_agg_1): Likewise. Update callers.
693 (compute_complex_assign_jump_func): Changed to use new param AA info.
694 (compute_complex_ancestor_jump_func): Likewise.
695 (ipa_compute_jump_functions_for_edge): Likewise.
696 (ipa_compute_jump_functions): Removed.
697 (ipa_compute_jump_functions_for_bb): New function.
698 (ipa_analyze_indirect_call_uses): Likewise, moved variable
700 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
701 and info, moved variable declarations down.
702 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
704 (ipa_analyze_stmt_uses): Likewise.
705 (ipa_analyze_params_uses): Removed.
706 (ipa_analyze_params_uses_in_bb): New function.
707 (ipa_analyze_controlled_uses): Likewise.
708 (free_ipa_bb_info): Likewise.
709 (analysis_dom_walker): New class.
710 (ipa_analyze_node): Handle node-specific forbidden analysis,
711 initialize and free func_body_info, use dominator walker.
712 (ipcp_modif_dom_walker): New class.
713 (ipcp_transform_function): Create and free func_body_info, use
714 ipcp_modif_dom_walker, moved a lot of functionality there.
716 2014-05-23 Marek Polacek <polacek@redhat.com>
717 Jakub Jelinek <jakub@redhat.com>
719 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
720 * gcc.c (sanitize_spec_function): Likewise.
721 * convert.c (convert_to_integer): Include "ubsan.h". Add
722 floating-point to integer instrumentation.
723 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
724 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
726 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
727 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
728 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
729 * ubsan.c: Include "realmpfr.h" and "dfp.h".
730 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
731 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
732 float/double/long double.
733 (ubsan_instrument_float_cast): New function.
734 * ubsan.h (ubsan_instrument_float_cast): Declare.
736 2014-05-23 Jiong Wang <jiong.wang@arm.com>
738 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
740 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
741 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
742 Adjust for tailcalling through registers.
743 * config/aarch64/aarch64.h (enum reg_class): New caller save
745 (REG_CLASS_NAMES): Likewise.
746 (REG_CLASS_CONTENTS): Likewise.
747 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
748 Allow tailcalling without decls.
750 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
752 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
753 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
755 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
756 gsi, and variables v_* to v*.
758 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
760 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
762 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
764 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
765 * omp-low.c: Update accordingly.
767 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
768 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
769 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
770 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
771 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
772 GF_OMP_TARGET_KIND_UPDATE.
774 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
775 Explicitly enumerate the expected region types.
777 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
780 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
781 documentation; the old documentation didn't clearly state the
782 constraints on the contents of the pointed-to storage.
784 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
786 Fix bootstrap error on ia64
787 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
788 Return default value.
790 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
792 PR tree-optimization/54733
793 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
795 (find_bswap_or_nop_load): New.
796 (find_bswap_1): Renamed to ...
797 (find_bswap_or_nop_1): This. Also add support for memory source.
798 (find_bswap): Renamed to ...
799 (find_bswap_or_nop): This. Also add support for memory source and
800 detection of bitwise operations equivalent to load in host endianness.
801 (execute_optimize_bswap): Likewise. Also move its leading comment back
802 in place and split statement transformation into ...
803 (bswap_replace): This.
805 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
807 PR rtl-optimization/61215
808 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
809 simplify_gen_subreg until final substitution.
811 2014-05-23 Alan Modra <amodra@gmail.com>
814 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
815 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
816 Use "Y" constraint rather than "m".
818 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
820 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
822 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
823 New function declaration.
824 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
825 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
826 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
827 (aarch64_init_builtins) : Initialize builtins
828 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
829 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
830 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
831 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
832 and __builtins_aarch64_set_fpsr.
833 (aarch64_atomic_assign_expand_fenv): New function.
834 * config/aarch64/aarch64.md (set_fpcr): New pattern.
835 (get_fpcr) : Likewise.
836 (set_fpsr) : Likewise.
837 (get_fpsr) : Likewise.
838 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
839 and UNSPECV_SET_FPSR.
840 * doc/extend.texi (AARCH64 Built-in Functions) : Document
841 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
842 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
844 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
846 PR rtl-optimization/60969
847 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
848 constraints. Set up mem cost for NO_REGS case.
850 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
852 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
854 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
856 * config/darwin.c: Include "lto-section-names.h".
857 (LTO_SEGMENT_NAME): Don't define.
858 * config/i386/winnt.c: Include "lto-section-names.h".
859 * lto-streamer.c: Include "lto-section-names.h".
860 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
861 * lto-wrapper.c: Include "lto-section-names.h".
862 (LTO_SECTION_NAME_PREFIX): Don't define.
863 * lto-section-names.h: New file.
864 * cgraphunit.c: Include "lto-section-names.h".
866 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
868 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
870 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
873 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
875 2014-05-22 Nick Clifton <nickc@redhat.com>
877 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
879 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
881 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
882 -> (T)A transformation to integer types.
884 2014-05-22 Teresa Johnson <tejohnson@google.com>
886 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
887 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
888 (gcov_rewrite): Use gcov_nonruntime_assert.
890 (gcov_write_words): Ditto.
891 (gcov_write_length): Ditto.
892 (gcov_read_words): Use gcov_nonruntime_assert, and remove
893 gcc_assert from IN_LIBGCOV code.
894 (gcov_read_summary): Use gcov_error to flag profile corruption.
895 (gcov_sync): Use gcov_nonruntime_assert.
896 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
897 (gcov_histo_index): Use gcov_nonruntime_assert.
898 (static void gcov_histogram_merge): Ditto.
899 (compute_working_sets): Ditto.
900 * gcov-io.h (gcov_nonruntime_assert): Define.
901 (gcov_error): Define for !IN_LIBGCOV
903 2014-05-22 Richard Biener <rguenther@suse.de>
905 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
906 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
907 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
908 and deallocation site.
909 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
910 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
911 passing through the incoming points-to set.
912 (handle_lhs_call): Use flags argument instead of recomputing it.
913 (find_func_aliases_for_call): Call handle_lhs_call with proper
916 2014-05-22 Jakub Jelinek <jakub@redhat.com>
918 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
919 all padding bits in REAL_VALUE_TYPE are cleared.
921 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
923 Cleanup and improve multipass_dfa_lookahead_guard
924 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
925 (core2i7_first_cycle_multipass_begin,)
926 (core2i7_first_cycle_multipass_issue,)
927 (core2i7_first_cycle_multipass_backtrack): Update signature.
929 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
930 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
931 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
933 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
934 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
936 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
938 * doc/tm.texi: Regenerate.
940 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
941 * haifa-sched.c (ready_try): Make signed to allow negative values.
942 (rebug_ready_list_1): Update.
943 (choose_ready): Simplify.
944 (sched_extend_ready_list): Update.
946 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
948 Remove IA64 speculation tweaking flags
949 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
950 speculation tuning flags.
951 (msched-prefer-non-data-spec-insns,)
952 (msched-prefer-non-control-spec-insns): Obsolete options.
953 * haifa-sched.c (choose_ready): Remove handling of
954 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
955 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
956 and PREFER_NON_DATA_SPEC.
957 * sel-sched.c (process_spec_exprs): Remove handling of
958 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
960 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
962 Improve scheduling debug output
963 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
964 (advance_one_cycle): Update.
965 (schedule_insn, queue_to_ready): Add debug printouts.
966 (debug_ready_list_1): New static function.
967 (debug_ready_list): Update.
968 (max_issue): Add debug printouts.
969 (dump_insn_stream): New static function.
970 (schedule_block): Use it. Also better indent printouts.
972 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
974 Fix sched_insn debug counter
975 * haifa-sched.c (schedule_insn): Update.
976 (struct haifa_saved_data): Add nonscheduled_insns_begin.
977 (save_backtrack_point, restore_backtrack_point): Update.
978 (first_nonscheduled_insn): New static function.
979 (queue_to_ready, choose_ready): Use it.
980 (schedule_block): Init nonscheduled_insns_begin.
981 (sched_emit_insn): Update.
984 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
986 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
988 (aarch64_secondary_reload) : LikeWise.
989 (aarch64_class_max_nregs) : Remove CORE_REGS.
990 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
991 (REG_CLASS_NAMES) : Likewise.
992 (REG_CLASS_CONTENTS) : LikeWise.
993 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
995 2014-05-21 Guozhi Wei <carrot@google.com>
998 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1000 (vqdmulhq_n_s16): Likewise.
1002 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1004 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1006 2014-05-21 Marek Polacek <polacek@redhat.com>
1009 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1011 2014-05-21 Martin Jambor <mjambor@suse.cz>
1013 * doc/invoke.texi (Optimize Options): Document parameters
1014 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1015 ipa-cp-array-index-hint-bonus.
1017 2014-05-21 Mark Wielaard <mjw@redhat.com>
1020 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1021 version >= 3 or not strict DWARF.
1022 * langhooks.h (struct lang_hooks_for_types): Add
1023 enum_underlying_base_type.
1024 * langhooks.c (lhd_enum_underlying_base_type): New function.
1025 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1026 enum_underlying_base_type.
1027 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1028 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1029 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1031 2014-05-21 Richard Biener <rguenther@suse.de>
1033 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1035 2014-05-21 John Marino <gnugcc@marino.st>
1037 * config.gcc (*-*-dragonfly*): New target.
1038 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1039 * configure: Regenerate.
1040 * config/dragonfly-stdint.h: New.
1041 * config/dragonfly.h: New.
1042 * config/dragonfly.opt: New.
1043 * config/i386/dragonfly.h: New.
1044 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1046 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1048 * tree.def (VOID_CST): New.
1049 * tree-core.h (TI_VOID): New.
1050 * tree.h (void_node): New.
1051 * tree.c (tree_node_structure_for_code, tree_code_size)
1052 (iterative_hash_expr): Handle VOID_CST.
1053 (build_common_tree_nodes): Initialize void_node.
1055 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1057 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1059 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1061 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1064 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1065 flag_reorder_blocks_and_partition.
1066 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1068 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1071 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1073 (*addc_r_t): Add new insn_and_split.
1075 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1078 * omp-low.c (handle_simd_reference): New function.
1079 (lower_rec_input_clauses): Use it. Defer adding reference
1080 initialization even for reduction without placeholder if in simd,
1081 handle it properly later on.
1083 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1085 PR tree-optimization/60899
1086 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1087 assume all static symbols will have definition wile parsing and
1088 check the do have definition later in compilation; check that
1089 variable referring symbol will be output before concluding that
1090 reference is safe; be conservative for referring local statics;
1091 be more precise about when comdat is output in other partition.
1093 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1096 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1098 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1099 (ipa_inline): Loop inline_to_all_callers until no more aliases
1102 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1104 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1105 set writeonly flag only for vars actually written to.
1107 2014-05-20 Dehao Chen <dehao@google.com>
1109 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1110 and callee count to get clone count.
1111 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1114 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1116 PR rtl-optimization/61243
1117 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1119 2014-05-20 Xinliang David Li <davidxl@google.com>
1121 * cgraphunit.c (walk_polymorphic_call_targets): Add
1122 dbgcnt and fopt-info support.
1123 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1124 * ipa-devirt.c (ipa_devirt): Ditto.
1125 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1126 * ipa.c (walk_polymorphic_call_targets): Ditto.
1127 * gimple-fold.c (fold_gimple_assign): Ditto.
1128 (gimple_fold_call): Ditto.
1129 * dbgcnt.def: New counter.
1131 2014-05-20 DJ Delorie <dj@redhat.com>
1133 * config/msp430/msp430.md (split): Don't allow subregs when
1134 splitting SImode adds.
1135 (andneghi): Fix subtraction logic.
1136 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1138 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1140 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1142 * except.c (switch_to_exception_section, resolve_unique_section,
1143 get_named_text_section, default_function_rodata_section,
1144 align_variable, get_block_for_decl, default_section_type_flags):
1145 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1146 * symtab.c (symtab_add_to_same_comdat_group,
1147 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1148 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1150 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1151 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1152 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1153 (c6x_function_in_section_p): Likewise.
1154 * config/darwin.c (machopic_select_section): Likewise.
1155 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1156 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1157 * config/mep/mep.c (mep_select_section): LIkewise.
1158 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1160 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1162 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1163 EH region of calls to pure functions that can throw an exception.
1164 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1165 (copy_reference_ops_from_call): Also copy the EH region of the call if
1166 it can throw an exception.
1168 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1170 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1171 nested VEC_SELECTs that are inverses of each other.
1173 2014-05-20 Richard Biener <rguenther@suse.de>
1175 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1176 (extract_and_process_scc_for_name): not here.
1177 (cond_dom_walker::before_dom_children): Only process
1178 stmts that end the BB in interesting ways.
1179 (run_scc_vn): Mark param uses as visited.
1181 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1183 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1186 2014-05-20 Nick Clifton <nickc@redhat.com>
1188 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1189 (msp430_gimplify_va_arg_expr): New function.
1190 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1192 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1193 operand 0 in order to prevent confusion about the number of
1196 2014-05-20 Richard Biener <rguenther@suse.de>
1198 PR tree-optimization/61221
1199 * tree-ssa-pre.c (el_to_update): Remove.
1200 (eliminate_dom_walker::before_dom_children): Handle released
1201 VDEFs by value-numbering them to the associated VUSE. Update
1202 stmt immediately for substituted call address.
1203 (eliminate): Remove delayed stmt updating code.
1204 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1205 possibly late re-numbered vuses.
1206 (vn_reference_lookup_2): Adjust.
1207 (vn_reference_lookup_pieces): Likewise.
1208 (vn_reference_lookup): Likewise.
1210 2014-05-20 Richard Biener <rguenther@suse.de>
1212 * config.gcc: Remove need_64bit_hwint.
1213 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1214 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1216 * config.in: Regenerate.
1217 * configure: Likewise.
1219 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1221 * doc/extend.texi: Create Label Attributes section,
1222 move all label attributes into it and reference it.
1224 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1226 * arm.c (thumb1_reorg): When scanning backwards skip anything
1227 that's not a proper insn.
1229 2014-05-19 Richard Biener <rguenther@suse.de>
1231 PR tree-optimization/61221
1232 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1233 Do nothing for unreachable blocks.
1234 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1235 Improve unreachability detection.
1237 2014-05-19 Richard Biener <rguenther@suse.de>
1239 PR tree-optimization/61209
1240 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1242 2014-05-19 Nick Clifton <nickc@redhat.com>
1244 * except.c (init_eh): Fix computation of builtin setjmp buffer
1245 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1247 2014-05-19 Richard Biener <rguenther@suse.de>
1249 PR tree-optimization/61184
1250 * tree-vrp.c (is_negative_overflow_infinity): Use
1251 TREE_OVERFLOW_P and do that check first.
1252 (is_positive_overflow_infinity): Likewise.
1253 (is_overflow_infinity): Likewise.
1254 (vrp_operand_equal_p): Properly treat operands with
1255 differing overflow as not equal.
1257 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1259 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1260 shift simplification where it was intended.
1262 2014-05-19 Christian Bruel <christian.bruel@st.com>
1265 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1267 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1270 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1273 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1275 * reg-notes.def (CROSSING_JUMP): Likewise.
1276 * rtl.h (rtx_def): Update comment for jump flag.
1277 (CROSSING_JUMP_P): Define.
1278 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1279 of a REG_CROSSING_JUMP note.
1280 * cfghooks.c (tidy_fallthru_edges): Likewise.
1281 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1282 * emit-rtl.c (try_split): Likewise.
1283 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1284 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1285 * jump.c (redirect_jump_2): Likewise.
1286 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1287 (relax_delay_slots): Likewise.
1288 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1289 (bbit_di): Likewise.
1290 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1291 * config/sh/sh.md (jump_compact): Likewise.
1292 * bb-reorder.c (rotate_loop): Likewise.
1293 (pass_duplicate_computed_gotos::execute): Likewise.
1294 (add_reg_crossing_jump_notes): Rename to...
1295 (update_crossing_jump_flags): ...this.
1296 (pass_partition_blocks::execute): Update accordingly.
1298 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1300 * tree.h: Remove extraneous template <>.
1302 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1304 * ipa.c (symtab_remove_unreachable_nodes): Remove
1305 symbol from comdat group if its body was eliminated.
1306 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1307 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1308 (symtab_unregister_node): ... this one.
1309 (verify_symtab_base): More strict checking of comdats.
1310 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1312 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1314 * tree-pass.h (make_pass_ipa_comdats): New pass.
1315 * timevar.def (TV_IPA_COMDATS): New timevar.
1316 * passes.def (pass_ipa_comdats): Add.
1317 * Makefile.in (OBJS): Add ipa-comdats.o
1318 * ipa-comdats.c: New file.
1320 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1322 * ipa.c (update_visibility_by_resolution_info): New function.
1323 (function_and_variable_visibility): Use it.
1325 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1327 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1329 (FOR_EACH_DEFINED_SYMBOL): New macro.
1330 (varpool_first_static_initializer, varpool_next_static_initializer,
1331 varpool_first_defined_variable, varpool_next_defined_variable):
1333 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1335 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1337 * ggc-page.c (ggc_handle_finalizers): Add comment.
1339 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1341 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1342 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1343 (ggc_internal_cleared_alloc): Likewise.
1344 * ggc-page.c (finalizer): New class.
1345 (vec_finalizer): Likewise.
1346 (globals::finalizers): New member.
1347 (globals::vec_finalizers): Likewise.
1348 (ggc_internal_alloc): Record the finalizer if any for the block being
1350 (ggc_handle_finalizers): New function.
1351 (ggc_collect): Call ggc_handle_finalizers.
1352 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1354 (ggc_internal_cleared_alloc): Likewise.
1355 (finalize): New function.
1356 (need_finalization_p): Likewise.
1357 (ggc_alloc): Install the type's destructor as the finalizer if it
1359 (ggc_cleared_alloc): Likewise.
1360 (ggc_vec_alloc): Likewise.
1361 (ggc_cleared_vec_alloc): Likewise.
1363 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1365 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1367 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1369 * alias.c (record_alias_subset): Adjust.
1370 * bitmap.c (bitmap_element_allocate): Likewise.
1371 (bitmap_gc_alloc_stat): Likewise.
1372 * cfg.c (init_flow): Likewise.
1373 (alloc_block): Likewise.
1374 (unchecked_make_edge): Likewise.
1375 * cfgloop.c (alloc_loop): Likewise.
1376 (flow_loops_find): Likewise.
1377 (rescan_loop_exit): Likewise.
1378 * cfgrtl.c (init_rtl_bb_info): Likewise.
1379 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1380 (cgraph_allocate_node): Likewise.
1381 (cgraph_create_edge_1): Likewise.
1382 (cgraph_allocate_init_indirect_info): Likewise.
1383 * cgraphclones.c (cgraph_clone_edge): Likewise.
1384 * cgraphunit.c (add_asm_node): Likewise.
1385 (init_lowered_empty_function): Likewise.
1386 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1387 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1388 (alpha_use_linkage): Likewise.
1389 * config/arc/arc.c (arc_init_machine_status): Likewise.
1390 * config/arm/arm.c (arm_init_machine_status): Likewise.
1391 * config/avr/avr.c (avr_init_machine_status): Likewise.
1392 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1393 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1394 * config/cris/cris.c (cris_init_machine_status): Likewise.
1395 * config/darwin.c (machopic_indirection_name): Likewise.
1396 (darwin_build_constant_cfstring): Likewise.
1397 (darwin_enter_string_into_cfstring_table): Likewise.
1398 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1399 * config/frv/frv.c (frv_init_machine_status): Likewise.
1400 * config/i386/i386.c (get_dllimport_decl): Likewise.
1401 (ix86_init_machine_status): Likewise.
1402 (assign_386_stack_local): Likewise.
1403 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1404 (i386_pe_maybe_record_exported_symbol): Likewise.
1405 (i386_pe_record_stub): Likewise.
1406 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1407 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1408 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1409 (m32c_note_pragma_address): Likewise.
1410 * config/mep/mep.c (mep_init_machine_status): Likewise.
1411 (mep_note_pragma_flag): Likewise.
1412 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1413 (mips16_local_alias): Likewise.
1414 (mips_init_machine_status): Likewise.
1415 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1416 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1417 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1418 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1419 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1420 * config/pa/pa.c (pa_init_machine_status): Likewise.
1421 (pa_get_deferred_plabel): Likewise.
1422 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1423 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1424 (rs6000_init_machine_status): Likewise.
1425 (output_toc): Likewise.
1426 * config/s390/s390.c (s390_init_machine_status): Likewise.
1427 * config/score/score.c (score_output_external): Likewise.
1428 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1429 * config/spu/spu.c (spu_init_machine_status): Likewise.
1430 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1431 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1432 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1433 * coverage.c (coverage_end_function): Likewise.
1434 * dbxout.c (dbxout_init): Likewise.
1435 * doc/gty.texi: Don't mention variable_size attribute.
1436 * dwarf2cfi.c (new_cfi): Adjust.
1437 (new_cfi_row): Likewise.
1438 (copy_cfi_row): Likewise.
1439 (create_cie_data): Likewise.
1440 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1441 (new_loc_descr): Likewise.
1442 (find_AT_string_in_table): Likewise.
1443 (add_addr_table_entry): Likewise.
1444 (new_die): Likewise.
1445 (add_var_loc_to_decl): Likewise.
1446 (clone_die): Likewise.
1447 (clone_as_declaration): Likewise.
1448 (break_out_comdat_types): Likewise.
1449 (new_loc_list): Likewise.
1450 (add_loc_descr_to_each): Likewise.
1451 (add_location_or_const_value_attribute): Likewise.
1452 (add_linkage_name): Likewise.
1453 (lookup_filename): Likewise.
1454 (dwarf2out_var_location): Likewise.
1455 (new_line_info_table): Likewise.
1456 (dwarf2out_init): Likewise.
1457 (mem_loc_descriptor): Likewise.
1458 (loc_descriptor): Likewise.
1459 (add_const_value_attribute): Likewise.
1460 (tree_add_const_value_attribute): Likewise.
1461 (comp_dir_string): Likewise.
1462 (dwarf2out_vms_debug_main_pointer): Likewise.
1463 (string_cst_pool_decl): Likewise.
1464 * emit-rtl.c (set_mem_attrs): Likewise.
1465 (get_reg_attrs): Likewise.
1466 (start_sequence): Likewise.
1467 (init_emit): Likewise.
1468 (init_emit_regs): Likewise.
1469 * except.c (init_eh_for_function): Likewise.
1470 (gen_eh_region): Likewise.
1471 (gen_eh_region_catch): Likewise.
1472 (gen_eh_landing_pad): Likewise.
1473 (add_call_site): Likewise.
1474 * function.c (add_frame_space): Likewise.
1475 (insert_temp_slot_address): Likewise.
1476 (assign_stack_temp_for_type): Likewise.
1477 (get_hard_reg_initial_val): Likewise.
1478 (allocate_struct_function): Likewise.
1479 (prepare_function_start): Likewise.
1480 (types_used_by_var_decl_insert): Likewise.
1481 * gengtype.c (variable_size_p): Remove function.
1482 (enum alloc_quantity): Remove enum.
1483 (write_typed_alloc_def): Remove function.
1484 (write_typed_struct_alloc_def): Likewise.
1485 (write_typed_typedef_alloc_def): Likewise.
1486 (write_typed_alloc_defns): Likewise.
1488 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1489 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1490 * ggc.h (ggc_alloc): new function.
1491 (ggc_cleared_alloc): Likewise.
1492 (ggc_vec_alloc): Template on type of vector element, and remove
1493 element size argument.
1494 (ggc_cleared_vec_alloc): Likewise.
1495 * gimple.c (gimple_build_omp_for): Adjust.
1496 (gimple_copy): Likewise.
1497 * ipa-cp.c (get_replacement_map): Likewise.
1498 (find_aggregate_values_for_callers_subset): Likewise.
1499 (known_aggs_to_agg_replacement_list): Likewise.
1500 * ipa-devirt.c (get_odr_type): Likewise.
1501 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1502 (read_agg_replacement_chain): Likewise.
1503 * loop-iv.c (get_simple_loop_desc): Likewise.
1504 * lto-cgraph.c (input_node_opt_summary): Likewise.
1505 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1506 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1507 (input_eh_region): Likewise.
1508 (input_eh_lp): Likewise.
1509 (input_cfg): Likewise.
1510 * optabs.c (set_optab_libfunc): Likewise.
1511 (init_tree_optimization_optabs): Likewise.
1512 (set_conv_libfunc): Likewise.
1513 * passes.c (do_per_function_toporder): Likewise.
1514 * rtl.h: Don't use variable_size gty attribute.
1515 * sese.c (if_region_set_false_region): Adjust.
1516 * stringpool.c (gt_pch_save_stringpool): Likewise.
1517 * target-globals.c (save_target_globals): Likewise.
1518 * toplev.c (general_init): Likewise.
1519 * trans-mem.c (record_tm_replacement): Likewise.
1520 (split_bb_make_tm_edge): Likewise.
1521 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1522 * tree-data-ref.h (lambda_vector_new): Likewise.
1523 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1524 * tree-iterator.c (tsi_link_before): Likewise.
1525 (tsi_link_after): Likewise.
1526 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1527 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1528 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1529 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1530 * tree-ssa.c (init_tree_ssa): Adjust.
1531 * tree-ssanames.c (set_range_info): Likewise.
1532 (get_ptr_info): Likewise.
1533 (duplicate_ssa_name_ptr_info): Likewise.
1534 (duplicate_ssa_name_range_info): Likewise.
1535 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1536 (unpack_ts_fixed_cst_value_fields): Likewise.
1537 * tree.c (build_fixed): Likewise.
1538 (build_real): Likewise.
1539 (build_string): Likewise.
1540 (decl_priority_info): Likewise.
1541 (decl_debug_expr_insert): Likewise.
1542 (decl_value_expr_insert): Likewise.
1543 (decl_debug_args_insert): Likewise.
1544 (type_hash_add): Likewise.
1545 (build_omp_clause): Likewise.
1546 * ubsan.c (decl_for_type_insert): Likewise.
1547 * varasm.c (get_unnamed_section): Likewise.
1548 (get_noswitch_section): Likewise.
1549 (get_section): Likewise.
1550 (get_block_for_section): Likewise.
1551 (create_block_symbol): Likewise.
1552 (build_constant_desc): Likewise.
1553 (create_constant_pool): Likewise.
1554 (force_const_mem): Likewise.
1555 (record_tm_clone_pair): Likewise.
1556 * varpool.c (varpool_create_empty_node): Likewise.
1558 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1560 * dwarf2out.c (tree_add_const_value_attribute): Call
1561 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1562 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1563 instead of ggc_internal_<x>alloc_stat.
1564 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1565 (ggc_realloc): Likewise.
1566 * ggc-none.c (ggc_internal_alloc): Likewise.
1567 (ggc_internal_cleared_alloc): Likewise.
1568 * ggc-page.c: Likewise.
1569 * ggc.h (ggc_internal_alloc_stat): Likewise.
1570 (ggc_internal_alloc): Remove macro.
1571 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1572 (ggc_internal_cleared_alloc): Remove macro.
1573 (GGC_RESIZEVEC): Adjust.
1574 (ggc_resizevar): Remove macro.
1575 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1576 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1577 (ggc_internal_vec_cleared_alloc): Remove macro.
1578 (ggc_alloc_atomic_stat): Drop _stat suffix.
1579 (ggc_alloc_atomic): Remove macro.
1580 (ggc_alloc_cleared_atomic): Remove macro.
1581 (ggc_alloc_string_stat): Drop _stat suffix.
1582 (ggc_alloc_string): Remove macro.
1583 (ggc_alloc_rtx_def_stat): Adjust.
1584 (ggc_alloc_tree_node_stat): Likewise.
1585 (ggc_alloc_cleared_tree_node_stat): Likewise.
1586 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1587 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1588 * gimple.c (gimple_build_omp_for): Likewise.
1589 (gimple_copy): Likewise.
1590 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1591 * toplev.c (realloc_for_line_map): Adjust.
1592 * tree-data-ref.h (lambda_vector_new): Likewise.
1593 * tree-phinodes.c (allocate_phi_node): Likewise.
1594 * tree.c (grow_tree_vec_stat): Likewise.
1595 * vec.h (va_gc::reserve): Adjust.
1597 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1599 * config/microblaze/microblaze.c (break_handler): New Declaration.
1600 (microblaze_break_function_p,microblaze_is_break_handler): New.
1601 (compute_frame_size): Use microblaze_break_function_p.
1602 Add the test of break_handler.
1603 (microblaze_function_prologue) : Add the test of variable
1604 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1605 (microblaze_function_epilogue) : Add the test of break_handler.
1606 (microblaze_globalize_label) : Add the test of break_handler.
1607 Check the name by BREAK_HANDLER_NAME.
1609 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1611 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1612 microblaze_is_break_handler test.
1613 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1614 Use SYMBOL_REF_DECL.
1616 * config/microblaze/microblaze-protos.h
1617 (microblaze_break_function_p,microblaze_is_break_handler):
1620 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1621 new MicroBlaze break_handler functions.
1623 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1625 * doc/extend.texi (Size of an asm): Move node text according
1626 to its @menu entry position.
1628 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1630 PR tree-optimization/61140
1631 PR tree-optimization/61150
1632 PR tree-optimization/61197
1633 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1635 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1637 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1639 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1641 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1642 __SIZEOF_INT128__ is defined.
1644 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1646 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1647 (rs6000_delegitimize_address): Use it.
1649 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1651 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1652 inplace argument. Store the new address in the original MEM when true.
1653 * emit-rtl.c (change_address_1): Likewise.
1654 (adjust_address_1, adjust_automodify_address_1, offset_address):
1656 * rtl.h (plus_constant): Add an inplace argument.
1657 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1658 when true. Avoid generating (plus X (const_int 0)).
1659 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1660 in-place. Pass true to plus_constant.
1661 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1663 2014-05-16 Dehao Chen <dehao@google.com>
1665 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1667 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1670 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1672 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1674 2014-05-16 Dehao Chen <dehao@google.com>
1676 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1677 optimize_function_for_size_p.
1678 * regs.h (REG_FREQ_FROM_BB): Likewise.
1680 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1683 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1684 negt_reg_operand cases.
1685 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1687 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1689 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1691 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1694 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1697 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1699 * tree-cfg.c (dump_function_to_file): Dump the return type of
1700 functions, in a line to itself before the function body, mimicking
1701 the layout of a C function.
1703 2014-05-16 Dehao Chen <dehao@google.com>
1705 * cfghooks.c (make_forwarder_block): Use direct computation to
1706 get fall-through edge's count and frequency.
1708 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1710 * config/arc/arc.c (arc_init): Fix typo in error message.
1711 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1712 (split_stack_prologue_scratch_regno): Likewise.
1713 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1714 word from error message.
1716 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1718 * ira-costs.c: Fix typo in comment.
1720 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1722 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1724 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1726 * varpool.c (dump_varpool_node): Dump write-only flag.
1727 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1729 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1730 write-only variables.
1731 * ipa.c (process_references): New function.
1732 (set_readonly_bit): New function.
1733 (set_writeonly_bit): New function.
1734 (clear_addressable_bit): New function.
1735 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1736 fix handling of aliases.
1737 * cgraph.h (struct varpool_node): Add writeonly flag.
1739 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1741 PR rtl-optimization/60969
1742 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1743 Calculate costs for this case.
1745 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1747 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1748 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1750 2014-05-16 Richard Biener <rguenther@suse.de>
1752 PR tree-optimization/61194
1753 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1754 bool patterns ending in a COND_EXPR.
1756 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1758 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1760 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1762 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1763 where we were unable to cost an RTX.
1765 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1767 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1770 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1771 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1773 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1775 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1776 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1778 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1779 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1781 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1782 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1784 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1787 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1788 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1790 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1793 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1794 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1796 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1797 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1799 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1800 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1802 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1805 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1806 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1808 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1809 ZERO_EXTEND and SIGN_EXTEND better.
1811 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1812 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1814 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1817 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1818 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1820 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1821 costs when costing loads and stores to memory.
1823 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1824 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
1826 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
1829 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1831 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
1833 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1834 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1836 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
1838 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
1840 (aarch64_rtx_mult_cost): New.
1841 (aarch64_rtx_costs): Use it, refactor as appropriate.
1843 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1844 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1846 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
1847 emit instructions, return number of instructions which would
1849 (aarch64_add_constant): Update call to aarch64_build_constant.
1850 (aarch64_output_mi_thunk): Likewise.
1851 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
1854 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1856 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
1857 (TARGET_RTX_COSTS): Call it.
1859 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1861 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
1862 (cortexa57_vector_cost): Likewise.
1863 (cortexa57_tunings): Use them.
1865 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1867 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
1868 (cpu_addrcost_table): Use it.
1869 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
1870 (aarch64_address_cost): Rewrite using aarch64_classify_address,
1873 2014-05-16 Richard Biener <rguenther@suse.de>
1875 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
1876 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
1877 (visit_phi): Ignore edges marked as not executable.
1878 (class cond_dom_walker): New.
1879 (cond_dom_walker::before_dom_children): Value-number
1880 control statements and mark successor edges as not
1881 executable if possible.
1882 (run_scc_vn): First walk all control statements in
1883 dominator order, marking edges as not executable.
1884 * tree-inline.c (copy_edges_for_bb): Be not confused
1885 about random edge flags.
1887 2014-05-16 Richard Biener <rguenther@suse.de>
1889 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
1891 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
1894 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
1895 (__TM_simple_begin): Use it.
1896 (__TM_begin): Likewise.
1898 2014-05-15 Martin Jambor <mjambor@suse.cz>
1901 * ipa-prop.c (update_indirect_edges_after_inlining): Check
1902 type_preserved flag when the indirect edge is polymorphic.
1904 2014-05-15 Martin Jambor <mjambor@suse.cz>
1906 PR tree-optimization/61090
1907 * tree-sra.c (sra_modify_expr): Pass the current gsi to
1908 build_ref_for_model.
1910 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1912 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
1913 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
1915 2014-05-15 Jakub Jelinek <jakub@redhat.com>
1917 PR tree-optimization/61158
1918 * fold-const.c (fold_binary_loc): If X is zero-extended and
1919 shiftc >= prec, make sure zerobits is all ones instead of
1920 invoking undefined behavior.
1922 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1924 * regcprop.h: New file.
1925 * regcprop.c (skip_debug_insn_p): New decl.
1926 (replace_oldest_value_reg): Check skip_debug_insn_p.
1927 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
1928 * shrink-wrap.c: Include regcprop.h.
1929 (prepare_shrink_wrap): Call
1930 copyprop_hardreg_forward_bb_without_debug_insn.
1932 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1934 * shrink-wrap.h: Update comment.
1935 * shrink-wrap.c: Update comment.
1936 (next_block_for_reg): Rename to live_edge_for_reg.
1937 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
1938 (move_insn_for_shrink_wrap): Split live_edge.
1939 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
1941 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
1943 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
1945 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
1946 * config/sparc/sparc.md (fptype_ut699): New attribute.
1947 (in_branch_delay): Return false if -mfix-ut699 is specified and
1948 fptype_ut699 is set to single.
1949 (truncdfsf2): Add fptype_ut699 attribute.
1950 (fix_truncdfsi2): Likewise.
1951 (floatsisf2): Change fptype attribute.
1952 (fix_truncsfsi2): Likewise.
1953 (negtf2_notv9): Delete.
1954 (negtf2_v9): Likewise.
1955 (negtf2_hq): New instruction.
1956 (negtf2): New instruction and splitter.
1957 (negdf2_notv9): Rewrite.
1958 (abstf2_notv9): Delete.
1959 (abstf2_hq_v9): Likewise.
1960 (abstf2_v9): Likewise.
1961 (abstf2_hq): New instruction.
1962 (abstf2): New instruction and splitter.
1963 (absdf2_notv9): Rewrite.
1965 2014-05-14 Cary Coutant <ccoutant@google.com>
1968 * opts.c (common_handle_option): Don't special-case "-g".
1969 (set_debug_level): Default to at least level 2 with "-g".
1971 2014-05-14 DJ Delorie <dj@redhat.com>
1973 * config/msp430/msp430.c (msp430_builtin): Add
1974 MSP430_BUILTIN_DELAY_CYCLES.
1975 (msp430_init_builtins): Register void __delay_cycles(long long).
1976 (msp430_builtin_decl): Add it.
1977 (cg_magic_constant): New.
1978 (msp430_expand_delay_cycles): New.
1979 (msp430_expand_builtin): Call it.
1980 (msp430_print_operand_raw): Change integer printing from "int" to
1982 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
1983 (delay_cycles_start): New.
1984 (delay_cycles_end): New.
1985 (delay_cycles_32): New.
1986 (delay_cycles_32x): New.
1987 (delay_cycles_16): New.
1988 (delay_cycles_16x): New.
1989 (delay_cycles_2): New.
1990 (delay_cycles_1): New.
1991 * doc/extend.texi: Document __delay_cycles().
1993 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
1995 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
1996 length attribute computation.
1998 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2001 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2003 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2006 * config/sparc/sparc.md: Fix types of low and high in DI constant
2007 splitter. Use gen_int_mode in some other splitters.
2009 2014-05-14 Martin Jambor <mjambor@suse.cz>
2012 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2014 2014-05-14 James Norris <jnorris@codesourcery.com>
2016 * omp-low.c (expand_parallel_call): Remove shadow variable.
2017 (expand_omp_taskreg): Likewise.
2019 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2021 * common/config/i386/i386-common.c
2022 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2023 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2024 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2025 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2026 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2027 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2028 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2029 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2030 xsavecintrin.h, xsavesintrin.h.
2031 (x86_64-*-*): Ditto.
2032 * config/i386/clflushoptintrin.h: New.
2033 * config/i386/xsavecintrin.h: Ditto.
2034 * config/i386/xsavesintrin.h: Ditto.
2035 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2036 (bit_XSAVES): Ditto.
2037 (bit_XSAVES): Ditto.
2038 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2039 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2041 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2042 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2043 OPTION_MASK_ISA_XSAVES.
2044 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2046 (PTA_CLFLUSHOPT) Define.
2047 (PTA_XSAVEC): Ditto.
2048 (PTA_XSAVES): Ditto.
2049 (ix86_option_override_internal): Handle new options.
2050 (ix86_valid_target_attribute_inner_p): Ditto.
2051 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2052 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2053 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2054 (bdesc_special_args): Add __builtin_ia32_xsaves,
2055 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2056 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2057 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2058 (ix86_expand_builtin): Handle new builtins.
2059 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2060 (TARGET_CLFLUSHOPT_P): Ditto.
2061 (TARGET_XSAVEC): Ditto.
2062 (TARGET_XSAVEC_P): Ditto.
2063 (TARGET_XSAVES): Ditto.
2064 (TARGET_XSAVES_P): Ditto.
2065 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2066 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2067 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2069 (ANY_XRSTOR64): Ditto.
2071 (xrstor): Change into <xrstor>.
2072 (xrstor_rex64): Change into <xrstor>_rex64.
2073 (xrstor64): Change into <xrstor>64
2075 * config/i386/i386.opt (mclflushopt): New.
2078 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2080 * doc/invoke.texi: Document new options.
2082 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2084 PR rtl-optimization/60866
2085 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2086 Default it to -1. Pass it down to init_simplejump_data.
2087 (init_simplejump_data): New parameter old_seqno. Pass it down
2088 to get_seqno_for_a_jump.
2089 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2090 initializing new jump seqno as a last resort. Add comment.
2091 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2092 jump and pass it down to sel_init_new_insn.
2093 (sel_redirect_edge_and_branch_force): Likewise.
2095 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2097 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2098 shifted values to avoid build warning.
2100 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2102 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2103 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2104 (cfg_layout_merge_blocks): Likewise.
2105 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2107 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2109 PR rtl-optimization/60901
2110 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2111 bb predecessor belongs to the same scheduling region. Adjust comment.
2113 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2115 * doc/sourcebuild.texi: (dfp_hw): Document.
2116 (p8vector_hw): Likewise.
2117 (powerpc_eabi_ok): Likewise.
2118 (powerpc_elfv2): Likewise.
2119 (powerpc_htm_ok): Likewise.
2120 (ppc_recip_hw): Likewise.
2123 2014-05-13 Cary Coutant <ccoutant@google.com>
2125 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2127 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2129 * gengtype-parse.c (require3): Eliminate in favor of...
2131 (require_template_declaration): Update to support optional single *
2134 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2135 (create_user_defined_type): Handle a single level of explicit
2136 pointerness within template arguments.
2137 (struct write_types_data): Add field "kind".
2138 (filter_type_name): Handle "*" character.
2139 (write_user_func_for_structure_ptr): Require a write_types_data
2140 rather than just a prefix string, so that we can look up the kind
2141 of the wtd and use it as an index into wrote_user_func_for_ptr,
2142 ensuring that such functions are written at most once. Support
2143 subclasses by invoking the marking function of the ultimate base class.
2144 (write_user_func_for_structure_body): Require a write_types_data
2145 rather than just a prefix string, so that we can pass this to
2146 write_user_func_for_structure_ptr.
2147 (write_func_for_structure): Likewise.
2148 (ggc_wtd): Add initializer of new "kind" field.
2149 (pch_wtd): Likewise.
2151 * gengtype.h (enum write_types_kinds): New.
2152 (struct type): Add field wrote_user_func_for_ptr to the "s"
2155 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2157 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2158 instead of const_binop.
2159 (fold_binary_loc): Likewise.
2161 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2163 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2164 calculation to match get_ref_base_and_extent.
2166 2014-05-13 Catherine Moore <clm@codesourcery.com>
2167 Sandra Loosemore <sandra@codesourcery.com>
2169 * configure.ac: Fix assembly for explicit JALR relocation check.
2170 * configure: Regenerate.
2172 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2174 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2175 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2176 Remove associated type declarations and initialisations.
2177 (arm_expand_neon_builtin): Likewise.
2178 (neon_emit_pair_result_insn): Delete.
2179 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2180 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2181 (neon_vzip<mode>): Likewise.
2182 (neon_vuzp<mode>): Likewise.
2184 2014-05-13 Richard Biener <rguenther@suse.de>
2187 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2188 it needs revisiting whether the call still may be tail-called.
2190 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2192 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2193 * rtl.h (block_symbol): Reduce number of fields to 2.
2194 (rtx_def): Add u2.symbol_ref_flags.
2195 (SYMBOL_REF_FLAGS): Use it.
2196 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2197 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2198 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2199 Lower index of SYMBOL_REF_DATA.
2200 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2201 Print SYMBOL_REF_FLAGS at the same time.
2202 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2204 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2206 * rtl.def (VAR_LOCATION): Remove "i" field.
2207 * rtl.h (rtx_def): Add u2.var_location_status.
2208 (PAT_VAR_LOCATION_STATUS): Use it.
2209 (gen_rtx_VAR_LOCATION): Declare.
2210 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2211 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2212 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2214 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2216 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2217 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2219 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2221 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2222 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2223 * rtl.h (rtx_def): Add insn_uid to u2 field.
2224 (RTX_FLAG_CHECK8): Delete in favor of...
2225 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2226 (INSN_DELETED_P): Update accordingly.
2227 (INSN_UID): Use u2.insn_uid.
2228 (INSN_CHAIN_CODE_P): Define.
2229 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2230 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2231 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2232 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2233 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2234 indices accordingly.
2235 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2236 Update indices for insn-chain rtxes.
2237 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2238 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2239 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2240 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2241 * combine.c (try_combine): Likewise.
2242 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2244 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2246 * rtl.def (REG): Remove middle field.
2247 * rtl.h (rtx_def): Add orignal_regno to u2.
2248 (ORIGINAL_REGNO): Use it instead of field 1.
2249 (REG_ATTRS): Lower field index accordingly.
2250 * gengtype.c (adjust_field_rtx_def): Remove handling of
2251 ORIGINAL_REGNO. Move REG_ATTRS index down.
2252 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2253 code that prints the REGNO.
2255 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2257 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2260 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2262 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2264 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2266 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2267 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2269 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2271 * config/aarch64/aarch64-protos.h
2272 (aarch64_hard_regno_caller_save_mode): New prototype.
2273 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2275 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2277 2014-05-13 Christian Bruel <christian.bruel@st.com>
2279 * target.def (mode_switching): New hook vector.
2280 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2281 (mode_exit, modepriority_to_mode): Likewise.
2282 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2283 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2284 * target.h: Include tm.h and hard-reg-set.h.
2285 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2286 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2287 * doc/tm.texi Regenerate.
2288 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2289 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2290 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2291 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2292 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2293 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2294 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2295 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2296 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2297 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2298 (ix86_emit_mode_set): Hookify.
2299 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2301 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2302 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2303 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2304 (epiphany_mode_priority_to_mode): Remove declaration.
2305 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2306 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2307 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2309 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2310 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2311 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2313 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2316 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2317 is const0_rtx, return immediately. Don't test count == 0 when
2320 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2322 * Makefile.in: add shrink-wrap.o.
2323 * config/i386/i386.c: include "shrink-wrap.h"
2324 * function.c: Likewise.
2325 (requires_stack_frame_p, next_block_for_reg,
2326 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2327 dup_block_and_redirect): Move to shrink-wrap.c
2328 (thread_prologue_and_epilogue_insns): Extract three code segments
2329 as functions in shrink-wrap.c
2330 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2332 * shrink-wrap.c: New file.
2333 * shrink-wrap.h: New file.
2335 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2337 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2338 reference to Solaris.
2340 2014-05-12 Mike Stump <mikestump@comcast.net>
2343 * genattrtab.c (filename): Add.
2344 (convert_set_attr_alternative): Improve error message.
2345 (check_defs): Restore read_md_filename for error messages.
2346 (gen_insn): Save filename.
2348 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2350 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2351 -fno-local-ivars and -fivar-visibility.
2352 * c-family/c.opt: Make -Wshadow also implicitly enable
2355 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2357 * doc/tm.texi: Remove reference to deleted macro.
2358 * doc/tm.texi.in: Likewise.
2360 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2363 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2366 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2369 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2370 * config/arm/aout.h (License): Same.
2371 * config/arm/bpabi.h (License): Same.
2372 * config/arm/elf.h (License): Same.
2373 * config/arm/linux-elf.h (License): Same.
2374 * config/arm/linux-gas.h (License): Same.
2375 * config/arm/netbsd-elf.h (License): Same.
2376 * config/arm/uclinux-eabi.h (License): Same.
2377 * config/arm/uclinux-elf.h (License): Same.
2378 * config/arm/vxworks.h (License): Same.
2380 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2382 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2383 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2384 number of operands to 3.
2385 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2386 * tree-nested.c (convert_nonlocal_omp_clauses,
2387 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2388 * gimplify.c (gimplify_scan_omp_clauses): Handle
2389 OMP_CLAUSE_LINEAR_STMT.
2390 * omp-low.c (lower_rec_input_clauses): Fix typo.
2391 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2392 cast between Fortran boolean_type_node and C _Bool if
2395 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2397 PR tree-optimization/61136
2398 * wide-int.h (multiple_of_p): Define a version that doesn't return
2400 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2401 integer_zerop/const_binop pair.
2402 (multiple_of_p): Likewise, converting both operands to widest_int
2405 2014-05-09 Teresa Johnson <tejohnson@google.com>
2407 * cgraphunit.c (analyze_functions): Use correct dump file.
2409 2014-05-09 Florian Weimer <fweimer@redhat.com>
2411 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2413 (stack_protect_return_slot_p): New function.
2414 (expand_used_vars): Call stack_protect_decl_p and
2415 stack_protect_return_slot_p for -fstack-protector-strong.
2417 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2420 * fold-const.c (fold_binary_loc): Changed width of mask.
2422 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2424 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2425 unsigned int initializers for regno_in, regno_out.
2427 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2430 * config/avr/avr.md (cc): Add new attribute set_vzn.
2431 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2432 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2433 with INC, DEC or NEG.
2434 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2435 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2436 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2438 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2441 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2443 * wide-int.cc (UTItype): Define.
2444 (UDWtype): Define for appropriate W_TYPE_SIZE.
2446 2014-05-09 Richard Biener <rguenther@suse.de>
2448 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2449 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2450 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2451 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2452 ssa_propagate): Adjust.
2454 2014-05-08 Jeff Law <law@redhat.com>
2456 PR tree-optimization/61009
2457 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2458 tri-state rather than a boolean. When a block is too big to
2459 thread through, inform caller via negative return value.
2460 (thread_across_edge): If a block was too big for normal threading,
2461 then it's too big for a joiner too, so remove temporary equivalences
2462 and return immediately.
2464 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2465 Matthias Klose <doko@ubuntu.com>
2468 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2470 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2473 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2475 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2478 * config/alpha/alpha.c: Include gimple-iterator.h.
2479 (alpha_gimple_fold_builtin): New function. Move
2480 ALPHA_BUILTIN_UMULH folding from ...
2481 (alpha_fold_builtin): ... here.
2482 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2484 2014-05-08 Wei Mi <wmi@google.com>
2487 * config/i386/i386.c (ix86_compute_frame_layout): Update
2488 preferred_stack_boundary for call, expanded from tls descriptor.
2489 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2490 to depend on SP register.
2491 (*tls_local_dynamic_base_32_gnu): Ditto.
2492 (*tls_local_dynamic_32_once): Ditto.
2493 (tls_global_dynamic_64_<mode>): Set
2494 ix86_tls_descriptor_calls_expanded_in_cfun.
2495 (tls_local_dynamic_base_64_<mode>): Ditto.
2496 (tls_global_dynamic_32): Set
2497 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2498 to depend on SP register.
2499 (tls_local_dynamic_base_32): Ditto.
2501 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2503 * config/arm/arm_neon.h: Update comment.
2504 * config/arm/neon-docgen.ml: Delete.
2505 * config/arm/neon-gen.ml: Delete.
2506 * doc/arm-neon-intrinsics.texi: Update comment.
2508 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2510 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2512 (vand, vorr, veor, vorn, vbic): Remove.
2513 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2515 (neon_vsub_unspec): Likewise.
2516 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2518 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2520 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2521 (vadd_s16): Likewise.
2522 (vadd_s32): Likewise.
2523 (vadd_f32): Likewise.
2524 (vadd_u8): Likewise.
2525 (vadd_u16): Likewise.
2526 (vadd_u32): Likewise.
2527 (vadd_s64): Likewise.
2528 (vadd_u64): Likewise.
2529 (vaddq_s8): Likewise.
2530 (vaddq_s16): Likewise.
2531 (vaddq_s32): Likewise.
2532 (vaddq_s64): Likewise.
2533 (vaddq_f32): Likewise.
2534 (vaddq_u8): Likewise.
2535 (vaddq_u16): Likewise.
2536 (vaddq_u32): Likewise.
2537 (vaddq_u64): Likewise.
2538 (vmul_s8): Likewise.
2539 (vmul_s16): Likewise.
2540 (vmul_s32): Likewise.
2541 (vmul_f32): Likewise.
2542 (vmul_u8): Likewise.
2543 (vmul_u16): Likewise.
2544 (vmul_u32): Likewise.
2545 (vmul_p8): Likewise.
2546 (vmulq_s8): Likewise.
2547 (vmulq_s16): Likewise.
2548 (vmulq_s32): Likewise.
2549 (vmulq_f32): Likewise.
2550 (vmulq_u8): Likewise.
2551 (vmulq_u16): Likewise.
2552 (vmulq_u32): Likewise.
2553 (vsub_s8): Likewise.
2554 (vsub_s16): Likewise.
2555 (vsub_s32): Likewise.
2556 (vsub_f32): Likewise.
2557 (vsub_u8): Likewise.
2558 (vsub_u16): Likewise.
2559 (vsub_u32): Likewise.
2560 (vsub_s64): Likewise.
2561 (vsub_u64): Likewise.
2562 (vsubq_s8): Likewise.
2563 (vsubq_s16): Likewise.
2564 (vsubq_s32): Likewise.
2565 (vsubq_s64): Likewise.
2566 (vsubq_f32): Likewise.
2567 (vsubq_u8): Likewise.
2568 (vsubq_u16): Likewise.
2569 (vsubq_u32): Likewise.
2570 (vsubq_u64): Likewise.
2571 (vand_s8): Likewise.
2572 (vand_s16): Likewise.
2573 (vand_s32): Likewise.
2574 (vand_u8): Likewise.
2575 (vand_u16): Likewise.
2576 (vand_u32): Likewise.
2577 (vand_s64): Likewise.
2578 (vand_u64): Likewise.
2579 (vandq_s8): Likewise.
2580 (vandq_s16): Likewise.
2581 (vandq_s32): Likewise.
2582 (vandq_s64): Likewise.
2583 (vandq_u8): Likewise.
2584 (vandq_u16): Likewise.
2585 (vandq_u32): Likewise.
2586 (vandq_u64): Likewise.
2587 (vorr_s8): Likewise.
2588 (vorr_s16): Likewise.
2589 (vorr_s32): Likewise.
2590 (vorr_u8): Likewise.
2591 (vorr_u16): Likewise.
2592 (vorr_u32): Likewise.
2593 (vorr_s64): Likewise.
2594 (vorr_u64): Likewise.
2595 (vorrq_s8): Likewise.
2596 (vorrq_s16): Likewise.
2597 (vorrq_s32): Likewise.
2598 (vorrq_s64): Likewise.
2599 (vorrq_u8): Likewise.
2600 (vorrq_u16): Likewise.
2601 (vorrq_u32): Likewise.
2602 (vorrq_u64): Likewise.
2603 (veor_s8): Likewise.
2604 (veor_s16): Likewise.
2605 (veor_s32): Likewise.
2606 (veor_u8): Likewise.
2607 (veor_u16): Likewise.
2608 (veor_u32): Likewise.
2609 (veor_s64): Likewise.
2610 (veor_u64): Likewise.
2611 (veorq_s8): Likewise.
2612 (veorq_s16): Likewise.
2613 (veorq_s32): Likewise.
2614 (veorq_s64): Likewise.
2615 (veorq_u8): Likewise.
2616 (veorq_u16): Likewise.
2617 (veorq_u32): Likewise.
2618 (veorq_u64): Likewise.
2619 (vbic_s8): Likewise.
2620 (vbic_s16): Likewise.
2621 (vbic_s32): Likewise.
2622 (vbic_u8): Likewise.
2623 (vbic_u16): Likewise.
2624 (vbic_u32): Likewise.
2625 (vbic_s64): Likewise.
2626 (vbic_u64): Likewise.
2627 (vbicq_s8): Likewise.
2628 (vbicq_s16): Likewise.
2629 (vbicq_s32): Likewise.
2630 (vbicq_s64): Likewise.
2631 (vbicq_u8): Likewise.
2632 (vbicq_u16): Likewise.
2633 (vbicq_u32): Likewise.
2634 (vbicq_u64): Likewise.
2635 (vorn_s8): Likewise.
2636 (vorn_s16): Likewise.
2637 (vorn_s32): Likewise.
2638 (vorn_u8): Likewise.
2639 (vorn_u16): Likewise.
2640 (vorn_u32): Likewise.
2641 (vorn_s64): Likewise.
2642 (vorn_u64): Likewise.
2643 (vornq_s8): Likewise.
2644 (vornq_s16): Likewise.
2645 (vornq_s32): Likewise.
2646 (vornq_s64): Likewise.
2647 (vornq_u8): Likewise.
2648 (vornq_u16): Likewise.
2649 (vornq_u32): Likewise.
2650 (vornq_u64): Likewise.
2652 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2654 * wide-int.cc (UTItype): Define.
2655 (UDWtype): Define for appropriate W_TYPE_SIZE.
2657 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2659 PR tree-optimization/59100
2660 * tree-ssa-phiopt.c: Include tree-inline.h.
2661 (neutral_element_p, absorbing_element_p): New functions.
2662 (value_replacement): Handle conditional binary operations with a
2663 neutral or absorbing element.
2665 2014-05-08 Richard Biener <rguenther@suse.de>
2667 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2668 folding the expression.
2669 (valueize_expr): Remove.
2670 (visit_reference_op_load): Do not valueize the result of
2672 (simplify_binary_expression): Likewise.
2673 (simplify_unary_expression): Likewise.
2675 2014-05-08 Richard Biener <rguenther@suse.de>
2677 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2678 looking at TYPE_ARG_TYPES.
2680 2014-05-08 Richard Biener <rguenther@suse.de>
2682 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2683 pointer propagation special-case.
2685 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2687 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2688 core part of address expressions.
2690 2014-05-08 Alan Modra <amodra@gmail.com>
2693 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2694 loads and stores when -mno-strict-align at any alignment.
2695 (expand_block_clear): Similarly. Also correct calculation of
2698 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2701 * tree-complex.c (expand_complex_move): Keep line info when expanding
2703 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2704 of complex expression. Use new argument to display correct location
2705 for values coming from phi statement.
2706 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2707 (warn_uninitialized_phi): Pass location of phi argument to
2709 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2710 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2712 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2714 * config/rs6000/predicates.md (indexed_address_mem): New.
2715 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2716 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2717 fpstore_ux, fpstore_u.
2718 (sign_extend, indexed, update): New.
2719 (cell_micro): Adjust.
2720 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2721 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2722 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2723 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2724 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2725 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2726 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2727 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2728 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2729 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2730 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2731 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2732 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2733 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2734 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2736 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2737 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2738 *vsx_extract_<mode>_store): Adjust.
2739 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2740 is_cracked_insn, insn_must_be_first_in_group,
2741 insn_must_be_last_in_group): Adjust.
2743 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2745 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2746 ppc440-fpstore): Adjust.
2747 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2748 ppc476-fpstore): Adjust.
2749 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2750 ppc601-fpstore): Adjust.
2751 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2753 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2755 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2756 ppc7450-fpstore): Adjust.
2757 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2758 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2759 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2761 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2762 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2763 cell-fpstore, cell-fpstore-update): Adjust.
2764 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2765 ppce300c3_store, ppce300c3_fpstore): Adjust.
2766 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2767 e500mc_fpstore): Adjust.
2768 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2769 e500mc64_store, e500mc64_fpstore): Adjust.
2770 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2771 e5500_fpstore): Adjust.
2772 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2773 e6500_fpstore): Adjust.
2774 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2776 * config/rs6000/power4.md (power4-load, power4-load-ext,
2777 power4-load-ext-update, power4-load-ext-update-indexed,
2778 power4-load-update-indexed, power4-load-update, power4-fpload,
2779 power4-fpload-update, power4-store, power4-store-update,
2780 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2782 * config/rs6000/power5.md (power5-load, power5-load-ext,
2783 power5-load-ext-update, power5-load-ext-update-indexed,
2784 power5-load-update-indexed, power5-load-update, power5-fpload,
2785 power5-fpload-update, power5-store, power5-store-update,
2786 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2788 * config/rs6000/power6.md (power6-load, power6-load-ext,
2789 power6-load-update, power6-load-update-indexed,
2790 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2791 power6-fpload-update, power6-store, power6-store-update,
2792 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2794 * config/rs6000/power7.md (power7-load, power7-load-ext,
2795 power7-load-update, power7-load-update-indexed,
2796 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2797 power7-fpload-update, power7-store, power7-store-update,
2798 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2800 * config/rs6000/power8.md (power8-load, power8-load-update,
2801 power8-load-ext, power8-load-ext-update, power8-fpload,
2802 power8-fpload-update, power8-store, power8-store-update-indexed,
2803 power8-fpstore, power8-fpstore-update): Adjust.
2804 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2806 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2807 titan_lsu_store, titan_lsu_fpstore): Adjust.
2808 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2810 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2813 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2814 unrolled byte insns. Emit address increments after move insns.
2816 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2818 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2819 const_gimple, rather than a gimple.
2820 (gimple_call_builtin_p): Likewise, for the three variants.
2822 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2823 (gimple_call_builtin_p): Likewise, for the three variants.
2825 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
2827 PR tree-optimization/61095
2828 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
2830 2014-05-07 Richard Biener <rguenther@suse.de>
2832 PR tree-optimization/61034
2833 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
2834 (maybe_skip_until): Use translate to take into account
2835 lattices when trying to do disambiguations.
2836 (get_continuation_for_phi_1): Likewise.
2837 (get_continuation_for_phi): Adjust for added translate arguments.
2838 (walk_non_aliased_vuses): Likewise.
2839 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
2840 (walk_non_aliased_vuses): Likewise.
2841 (call_may_clobber_ref_p_1): Declare.
2842 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
2843 calls. Stop early if we are only supposed to disambiguate.
2844 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
2846 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
2848 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
2849 Emit an error when the function has arguments.
2851 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2853 * cfgloop.h (unswitch_loops): Remove.
2854 * doc/passes.texi: Remove references to loop-unswitch.c
2855 * timevar.def (TV_LOOP_UNSWITCH): Remove.
2857 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
2859 * tree-vect-data-refs.c (vect_grouped_load_supported): New
2860 check for loads group of length 3.
2861 (vect_permute_load_chain): New permutations for loads group of
2863 * tree-vect-stmts.c (vect_model_load_cost): Change cost
2864 of vec_perm_shuffle for the new permutations.
2866 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
2868 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
2869 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
2870 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
2871 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
2872 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
2873 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
2874 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
2875 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
2877 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
2879 * loop-unswitch.c: Delete.
2881 2014-05-07 Richard Biener <rguenther@suse.de>
2883 * config.gcc: Always set need_64bit_hwint to yes.
2885 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
2887 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
2888 of using optimize_size.
2890 2014-05-06 Mike Stump <mikestump@comcast.net>
2892 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
2894 2014-05-06 Joseph Myers <joseph@codesourcery.com>
2896 * config/i386/sse.md (*mov<mode>_internal)
2897 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
2898 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
2899 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
2900 (*<code><mode>3, *andnot<mode>3<mask_name>)
2901 (<mask_codefor><code><mode>3<mask_name>): Only consider
2902 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
2904 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2907 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
2909 * lra-constraints.c (valid_address_p): Move earlier in file.
2910 Add a constraint argument to the address_info version.
2911 (satisfies_memory_constraint_p): New function.
2912 (satisfies_address_constraint_p): Likewise.
2913 (process_alt_operands, curr_insn_transform): Use them.
2914 (process_address): Pass the constraint to valid_address_p when
2915 checking address operands.
2917 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
2919 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
2920 to their respective blocks. Fix inadvertent use of "node".
2922 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
2924 * emit-rtl.c (init_derived_machine_modes): New functionm, split
2926 (init_emit_once): ...here.
2927 * rtl.h (init_derived_machine_modes): Declare.
2928 * toplev.c (do_compile): Call it even if no_backend.
2930 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
2931 Mike Stump <mikestump@comcast.net>
2932 Richard Sandiford <rdsandiford@googlemail.com>
2933 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2935 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
2936 (rtx_equal_for_memref_p): Update comment.
2937 (adjust_offset_for_component_ref): Use wide-int interfaces.
2938 * builtins.c (get_object_alignment_2): Likewise.
2939 (c_readstr): Likewise.
2940 (target_char_cast): Add comment.
2941 (determine_block_size): Use wide-int interfaces.
2942 (expand_builtin_signbit): Likewise.
2943 (fold_builtin_int_roundingfn): Likewise.
2944 (fold_builtin_bitop): Likewise.
2945 (fold_builtin_bswap): Likewise.
2946 (fold_builtin_logarithm): Use signop.
2947 (fold_builtin_pow): Likewise.
2948 (fold_builtin_memory_op): Use wide-int interfaces.
2949 (fold_builtin_object_size): Likewise.
2950 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
2951 nb_iterations_estimate.
2952 (record_niter_bound): Use wide-int interfaces.
2953 (get_estimated_loop_iterations_int): Likewise.
2954 (get_estimated_loop_iterations): Likewise.
2955 (get_max_loop_iterations): Likewise.
2956 * cfgloop.h: Include wide-int.h.
2957 (struct nb_iter_bound): Change bound to widest_int.
2958 (struct loop): Change nb_iterations_upper_bound and
2959 nb_iterations_estimate to widest_int.
2960 (record_niter_bound): Switch to use widest_int.
2961 (get_estimated_loop_iterations): Likewise.
2962 (get_max_loop_iterations): Likewise.
2963 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
2964 update for wide-int.
2965 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
2966 * combine.c (try_combine): Likewise.
2967 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
2968 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
2970 (aarch64_float_const_representable_p): Likewise.
2971 * config/arc/arc.c: Include wide-int.h.
2972 (arc_can_use_doloop_p): Use wide-int interfaces.
2973 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
2974 (vfp3_const_double_index): Likewise.
2975 * config/avr/avr.c (avr_out_round): Likewise.
2976 (avr_fold_builtin): Likewise.
2977 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
2978 (bfin_can_use_doloop_p): Likewise.
2979 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
2980 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
2981 * config/i386/i386.c: Include wide-int.h.
2982 (ix86_data_alignment): Use wide-int interfaces.
2983 (ix86_local_alignment): Likewise.
2984 (ix86_emit_swsqrtsf): Update real_from_integer.
2985 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
2986 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
2987 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
2988 (zero_constant): Likewise.
2989 (input_operand): Likewise.
2990 (splat_input_operand): Likewise.
2991 (non_logical_cint_operand): Change const_double to const_wide_int.
2992 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
2993 (easy_altivec_constant): Remove comment.
2994 (paired_expand_vector_init): Use CONSTANT_P.
2995 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
2996 (rs6000_emit_move): Update checks.
2997 (rs6000_aggregate_candidate): Use wide-int interfaces.
2998 (rs6000_expand_ternop_builtin): Likewise.
2999 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3000 (rs6000_assemble_integer): Likewise.
3001 (rs6000_hash_constant): Likewise.
3002 (output_toc): Likewise.
3003 (rs6000_rtx_costs): Likewise.
3004 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3005 * config/rs6000/rs6000-c.c: Include wide-int.h.
3006 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3007 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3008 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3009 Handle CONST_WIDE_INT.
3010 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3011 Use tree_fits_uhwi_p.
3012 * config/sparc/sparc.c: Include wide-int.h.
3013 (sparc_fold_builtin): Use wide-int interfaces.
3014 * config/vax/vax.c: Include wide-int.h.
3015 (vax_float_literal): Use real_from_integer.
3016 * coretypes.h (struct hwivec_def): New.
3018 (const_hwivec): New.
3019 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3020 (equiv_constant): Handle CONST_WIDE_INT.
3021 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3022 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3023 * dbxout.c (stabstr_U): Use wide-int interfaces.
3024 (dbxout_type): Update to use cst_fits_shwi_p.
3025 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3026 (TARGET_SUPPORTS_WIDE_INT): Add default.
3027 * dfp.c: Include wide-int.h.
3028 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3029 decimal_real_to_integer.
3030 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3031 decimal_real_to_integer.
3032 * doc/generic.texi (Constant expressions): Update for wide_int.
3033 * doc/rtl.texi (const_double): Likewise.
3034 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3035 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3036 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3037 (REAL_VALUE_FROM_INT): Remove.
3038 (TARGET_SUPPORTS_WIDE_INT): New.
3039 * doc/tm.texi: Regenerate.
3040 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3041 * double-int.h: Include wide-int.h.
3042 (struct wi::int_traits): New.
3043 * dwarf2out.c (get_full_len): New.
3044 (dw_val_equal_p): Add case dw_val_class_wide_int.
3045 (size_of_loc_descr): Likewise.
3046 (output_loc_operands): Likewise.
3047 (insert_double): Remove.
3048 (insert_wide_int): New.
3050 (print_die): Add case dw_val_class_wide_int.
3051 (attr_checksum): Likewise.
3052 (attr_checksum_ordered): Likewise.
3053 (same_dw_val_p): Likewise.
3054 (size_of_die): Likewise.
3055 (value_format): Likewise.
3056 (output_die): Likewise.
3057 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3059 (clz_loc_descriptor): Use wide-int interfaces.
3060 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3061 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3062 (round_up_to_align): Use wide-int interfaces.
3063 (field_byte_offset): Likewise.
3064 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3065 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3066 CONST_DOUBLE handling. Use wide-int interfaces.
3067 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3068 (gen_enumeration_type_die): Use add_AT_wide.
3069 (hash_loc_operands): Add case dw_val_class_wide_int.
3070 (compare_loc_operands): Likewise.
3071 * dwarf2out.h: Include wide-int.h.
3072 (wide_int_ptr): New.
3073 (enum dw_val_class): Add dw_val_class_wide_int.
3074 (struct dw_val_struct): Add val_wide.
3075 * emit-rtl.c (const_wide_int_htab): New.
3076 (const_wide_int_htab_hash): New.
3077 (const_wide_int_htab_eq): New.
3078 (lookup_const_wide_int): New.
3079 (const_double_htab_hash): Use wide-int interfaces.
3080 (const_double_htab_eq): Likewise.
3081 (rtx_to_double_int): Conditionally compile for wide-int.
3082 (immed_double_int_const): Rename to immed_wide_int_const and
3083 update for wide-int.
3084 (immed_double_const): Conditionally compile for wide-int.
3085 (init_emit_once): Use wide-int interfaces.
3086 * explow.c (plus_constant): Likewise.
3087 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3088 (lshift_value): Use wide-int interfaces.
3089 (expand_mult): Likewise.
3090 (choose_multiplier): Likewise.
3091 (expand_smod_pow2): Likewise.
3092 (make_tree): Likewise.
3093 * expr.c (convert_modes): Consolidate handling of constants.
3094 Use wide-int interfaces.
3095 (emit_group_load_1): Add note.
3096 (store_expr): Update comment.
3097 (get_inner_reference): Use wide-int interfaces.
3098 (expand_constructor): Update comment.
3099 (expand_expr_real_2): Use wide-int interfaces.
3100 (expand_expr_real_1): Likewise.
3101 (reduce_to_bit_field_precision): Likewise.
3102 (const_vector_from_tree): Likewise.
3103 * final.c: Include wide-int-print.h.
3104 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3105 * fixed-value.c: Include wide-int.h.
3106 (fixed_from_string): Use wide-int interfaces.
3107 (fixed_to_decimal): Likewise.
3108 (fixed_convert_from_real): Likewise.
3109 (real_convert_from_fixed): Likewise.
3110 * fold-const.h (mem_ref_offset): Return an offset_int.
3111 (div_if_zero_remainder): Remove code parameter.
3112 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3113 Use wide-int interfaces.
3114 (may_negate_without_overflow_p): Use wide-int interfaces.
3115 (negate_expr_p): Likewise.
3116 (fold_negate_expr): Likewise.
3117 (int_const_binop_1): Likewise.
3118 (const_binop): Likewise.
3119 (fold_convert_const_int_from_int): Likewise.
3120 (fold_convert_const_int_from_real): Likewise.
3121 (fold_convert_const_int_from_fixed): Likewise.
3122 (fold_convert_const_fixed_from_int): Likewise.
3123 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3124 (sign_bit_p): Use wide-int interfaces.
3125 (make_range_step): Likewise.
3126 (build_range_check): Likewise. Pass an integer of the correct type
3127 instead of using integer_one_node.
3128 (range_predecessor): Pass an integer of the correct type instead
3129 of using integer_one_node.
3130 (range_successor): Likewise.
3131 (merge_ranges): Likewise.
3132 (unextend): Use wide-int interfaces.
3133 (extract_muldiv_1): Likewise.
3134 (fold_div_compare): Likewise.
3135 (fold_single_bit_test): Likewise.
3136 (fold_sign_changed_comparison): Likewise.
3137 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3138 (fold_plusminus_mult_expr): Use wide-int interfaces.
3139 (native_encode_int): Likewise.
3140 (native_interpret_int): Likewise.
3141 (fold_unary_loc): Likewise.
3142 (pointer_may_wrap_p): Likewise.
3143 (size_low_cst): Likewise.
3144 (mask_with_tz): Likewise.
3145 (fold_binary_loc): Likewise.
3146 (fold_ternary_loc): Likewise.
3147 (multiple_of_p): Likewise.
3148 (tree_call_nonnegative_warnv_p): Update calls to
3149 tree_int_cst_min_precision and real_from_integer.
3150 (fold_negate_const): Use wide-int interfaces.
3151 (fold_abs_const): Likewise.
3152 (fold_relational_const): Use tree_int_cst_lt.
3153 (round_up_loc): Use wide-int interfaces.
3154 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3155 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3156 * gengtype.c: Remove include of double-int.h.
3157 (do_typedef): Use wide-int interfaces.
3158 (open_base_files): Add wide-int.h.
3159 (main): Add offset_int and widest_int typedefs.
3160 * gengtype-lex.l: Handle "^".
3161 (CXX_KEYWORD): Add "static".
3162 * gengtype-parse.c (require3): New.
3163 (require_template_declaration): Handle constant template arguments
3164 and nested templates.
3165 * gengtype-state.c: Don't include "double-int.h".
3166 * genpreds.c (write_one_predicate_function): Update comment.
3167 (write_tm_constrs_h): Add check for hval and lval use in
3169 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3170 (add_to_sequence): Likewise.
3171 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3172 and const_double_operand.
3173 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3175 * gimple-fold.c (get_base_constructor): Likewise.
3176 (fold_array_ctor_reference): Likewise.
3177 (fold_nonarray_ctor_reference): Likewise.
3178 (fold_const_aggregate_ref_1): Likewise.
3179 (gimple_val_nonnegative_real_p): Likewise.
3180 (gimple_fold_indirect_ref): Likewise.
3181 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3182 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3183 (struct slsr_cand_d): Change index to be widest_int.
3184 (struct incr_info_d): Change incr to be widest_int.
3185 (alloc_cand_and_find_basis): Use wide-int interfaces.
3186 (slsr_process_phi): Likewise.
3187 (backtrace_base_for_ref): Likewise. Return a widest_int.
3188 (restructure_reference): Take a widest_int instead of a double_int.
3189 (slsr_process_ref): Use wide-int interfaces.
3190 (create_mul_ssa_cand): Likewise.
3191 (create_mul_imm_cand): Likewise.
3192 (create_add_ssa_cand): Likewise.
3193 (create_add_imm_cand): Take a widest_int instead of a double_int.
3194 (slsr_process_add): Use wide-int interfaces.
3195 (slsr_process_cast): Likewise.
3196 (slsr_process_copy): Likewise.
3197 (dump_candidate): Likewise.
3198 (dump_incr_vec): Likewise.
3199 (replace_ref): Likewise.
3200 (cand_increment): Likewise. Return a widest_int.
3201 (cand_abs_increment): Likewise.
3202 (replace_mult_candidate): Take a widest_int instead of a double_int.
3203 (replace_unconditional_candidate): Use wide-int interfaces.
3204 (incr_vec_index): Take a widest_int instead of a double_int.
3205 (create_add_on_incoming_edge): Likewise.
3206 (create_phi_basis): Use wide-int interfaces.
3207 (replace_conditional_candidate): Likewise.
3208 (record_increment): Take a widest_int instead of a double_int.
3209 (record_phi_increments): Use wide-int interfaces.
3210 (phi_incr_cost): Take a widest_int instead of a double_int.
3211 (lowest_cost_path): Likewise.
3212 (total_savings): Likewise.
3213 (analyze_increments): Use wide-int interfaces.
3214 (ncd_with_phi): Take a widest_int instead of a double_int.
3215 (ncd_of_cand_and_phis): Likewise.
3216 (nearest_common_dominator_for_cands): Likewise.
3217 (insert_initializers): Use wide-int interfaces.
3218 (all_phi_incrs_profitable): Likewise.
3219 (replace_one_candidate): Likewise.
3220 (replace_profitable_candidates): Likewise.
3221 * godump.c: Include wide-int-print.h.
3222 (go_output_typedef): Use wide-int interfaces.
3223 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3224 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3225 (build_loop_iteration_domains): Likewise.
3226 * hooks.h: Include wide-int.h rather than double-int.h.
3227 (hook_bool_dint_dint_uint_bool_true): Delete.
3228 (hook_bool_wint_wint_uint_bool_true): Declare.
3229 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3230 (hook_bool_wint_wint_uint_bool_true): New.
3231 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3233 (ubsan_expand_si_overflow_mul_check): Likewise.
3234 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3235 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3236 (get_ancestor_addr_info): Likewise.
3237 (ipa_modify_call_arguments): Likewise.
3238 * loop-doloop.c (doloop_modify): Likewise.
3239 (doloop_optimize): Likewise.
3240 * loop-iv.c (iv_number_of_iterations): Likewise.
3241 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3242 (unroll_loop_constant_iterations): Likewise.
3243 (decide_unroll_runtime_iterations): Likewise.
3244 (unroll_loop_runtime_iterations): Likewise.
3245 (decide_peel_simple): Likewise.
3246 (decide_unroll_stupid): Likewise.
3247 * lto-streamer-in.c (streamer_read_wi): Add.
3248 (input_cfg): Use wide-int interfaces.
3249 (lto_input_tree_1): Likewise.
3250 * lto-streamer-out.c (streamer_write_wi): Add.
3251 (hash_tree): Use wide-int interfaces.
3252 (output_cfg): Likewise.
3253 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3254 (GTFILES): Add wide-int.h and signop.h.
3255 (TAGS): Look for .cc files too.
3256 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3257 * optabs.c (expand_subword_shift): Likewise.
3258 (expand_doubleword_shift): Likewise.
3259 (expand_absneg_bit): Likewise.
3260 (expand_copysign_absneg): Likewise.
3261 (expand_copysign_bit): Likewise.
3262 * postreload.c (reload_cse_simplify_set): Likewise.
3263 * predict.c (predict_iv_comparison): Likewise.
3264 * pretty-print.h: Include wide-int-print.h.
3266 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3267 * print-tree.c: Include wide-int-print.h.
3268 (print_node_brief): Use wide-int interfaces.
3269 (print_node): Likewise.
3270 * read-rtl.c (validate_const_wide_int): New.
3271 (read_rtx_code): Add CONST_WIDE_INT case.
3272 * real.c: Include wide-int.h.
3273 (real_to_integer2): Delete.
3274 (real_to_integer): New function, returning a wide_int.
3275 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3276 (ten_to_ptwo): Update call to real_from_integer.
3277 (real_digit): Likewise.
3278 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3279 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3280 (REAL_VALUE_TO_INT): Delete.
3281 (real_to_integer): Declare a wide-int form.
3282 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3283 * recog.c (const_int_operand): Improve comment.
3284 (const_scalar_int_operand): New.
3285 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3286 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3287 (split_double): Likewise.
3288 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3289 (rtx_size): Likewise.
3290 (rtx_alloc_stat_v): New.
3291 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3292 (cwi_output_hex): New.
3293 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3294 (cwi_check_failed_bounds): New.
3295 * rtl.def (CONST_WIDE_INT): New.
3296 * rtl.h: Include <utility> and wide-int.h.
3297 (struct hwivec_def): New.
3298 (CWI_GET_NUM_ELEM): New.
3299 (CWI_PUT_NUM_ELEM): New.
3300 (struct rtx_def): Add num_elem and hwiv.
3301 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3302 (CASE_CONST_UNIQUE): Likewise.
3303 (CASE_CONST_ANY): Likewise.
3304 (CONST_SCALAR_INT_P): Likewise.
3305 (CONST_WIDE_INT_P): New.
3307 (HWIVEC_CHECK): New.
3308 (cwi_check_failed_bounds): New.
3310 (HWIVEC_CHECK): New.
3311 (CONST_WIDE_INT_VEC) New.
3312 (CONST_WIDE_INT_NUNITS) New.
3313 (CONST_WIDE_INT_ELT) New.
3314 (rtx_mode_t): New type.
3315 (wi::int_traits <rtx_mode_t>): New.
3317 (wi::min_value): New.
3318 (wi::max_value): New.
3320 (const_wide_int_alloc): New.
3321 (immed_wide_int_const): New.
3322 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3323 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3324 * signop.h: New file.
3325 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3326 (simplify_const_unary_operation): Use wide-int interfaces.
3327 (simplify_binary_operation_1): Likewise.
3328 (simplify_const_binary_operation): Likewise.
3329 (simplify_const_relational_operation): Likewise.
3330 (simplify_immed_subreg): Likewise.
3331 * stmt.c (expand_case): Likewise.
3332 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3333 signop rather than a bool.
3334 * stor-layout.c (layout_type): Use wide-int interfaces.
3335 (initialize_sizetypes): Update calls to
3336 set_min_and_max_values_for_integral_type.
3337 (set_min_and_max_values_for_integral_type): Take a signop rather
3338 than a bool. Use wide-int interfaces.
3339 (fixup_signed_type): Update accordingly. Remove
3340 HOST_BITS_PER_DOUBLE_INT limit.
3341 (fixup_unsigned_type): Likewise.
3342 * system.h (STATIC_CONSTANT_P): New.
3343 (STATIC_ASSERT): New.
3344 * target.def (can_use_doloop_p): Take widest_ints rather than
3346 * target.h: Include wide-int.h rather than double-int.h.
3347 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3349 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3350 rather than INT_CST_LT_UNSIGNED.
3351 (can_use_doloop_if_innermost): Take widest_ints rather than
3353 * tree-affine.c: Include wide-int-print.h.
3354 (double_int_ext_for_comb): Delete.
3355 (wide_int_ext_for_comb): New.
3356 (aff_combination_zero): Use wide-int interfaces.
3357 (aff_combination_const): Take a widest_int instead of a double_int.
3358 (aff_combination_elt): Use wide-int interfaces.
3359 (aff_combination_scale): Take a widest_int instead of a double_int.
3360 (aff_combination_add_elt): Likewise.
3361 (aff_combination_add_cst): Likewise.
3362 (aff_combination_add): Use wide-int interfaces.
3363 (aff_combination_convert): Likewise.
3364 (tree_to_aff_combination): Likewise.
3365 (add_elt_to_tree): Take a widest_int instead of a double_int.
3366 (aff_combination_to_tree): Use wide-int interfaces.
3367 (aff_combination_remove_elt): Likewise.
3368 (aff_combination_add_product): Take a widest_int instead of
3370 (aff_combination_mult): Use wide-int interfaces.
3371 (aff_combination_expand): Likewise.
3372 (double_int_constant_multiple_p): Delete.
3373 (wide_int_constant_multiple_p): New.
3374 (aff_combination_constant_multiple_p): Take a widest_int pointer
3375 instead of a double_int pointer.
3376 (print_aff): Use wide-int interfaces.
3377 (get_inner_reference_aff): Take a widest_int pointer
3378 instead of a double_int pointer.
3379 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3380 * tree-affine.h: Include wide-int.h.
3381 (struct aff_comb_elt): Change type of coef to widest_int.
3382 (struct affine_tree_combination): Change type of offset to widest_int.
3383 (double_int_ext_for_comb): Delete.
3384 (wide_int_ext_for_comb): New.
3385 (aff_combination_const): Use widest_int instead of double_int.
3386 (aff_combination_scale): Likewise.
3387 (aff_combination_add_elt): Likewise.
3388 (aff_combination_constant_multiple_p): Likewise.
3389 (get_inner_reference_aff): Likewise.
3390 (aff_comb_cannot_overlap_p): Likewise.
3391 (aff_combination_zero_p): Use wide-int interfaces.
3392 * tree.c: Include tree.h.
3393 (init_ttree): Use make_int_cst.
3394 (tree_code_size): Removed code for INTEGER_CST case.
3395 (tree_size): Add INTEGER_CST case.
3396 (make_node_stat): Update comment.
3397 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3398 (build_int_cst_type): Use wide-int interfaces.
3399 (double_int_to_tree): Likewise.
3400 (double_int_fits_to_tree_p): Delete.
3401 (force_fit_type_double): Delete.
3402 (force_fit_type): New.
3403 (int_cst_hash_hash): Use wide-int interfaces.
3404 (int_cst_hash_eq): Likewise.
3405 (build_int_cst_wide): Delete.
3406 (wide_int_to_tree): New.
3407 (cache_integer_cst): Use wide-int interfaces.
3408 (build_low_bits_mask): Likewise.
3409 (cst_and_fits_in_hwi): Likewise.
3410 (real_value_from_int_cst): Likewise.
3411 (make_int_cst_stat): New.
3412 (integer_zerop): Use wide_int interfaces.
3413 (integer_onep): Likewise.
3414 (integer_all_onesp): Likewise.
3415 (integer_pow2p): Likewise.
3416 (integer_nonzerop): Likewise.
3417 (tree_log2): Likewise.
3418 (tree_floor_log2): Likewise.
3419 (tree_ctz): Likewise.
3420 (int_size_in_bytes): Likewise.
3421 (mem_ref_offset): Return an offset_int rather than a double_int.
3422 (build_type_attribute_qual_variant): Use wide_int interfaces.
3423 (type_hash_eq): Likewise
3424 (tree_int_cst_equal): Likewise.
3425 (tree_int_cst_lt): Delete.
3426 (tree_int_cst_compare): Likewise.
3427 (tree_fits_shwi_p): Use wide_int interfaces.
3428 (tree_fits_uhwi_p): Likewise.
3429 (tree_int_cst_sign_bit): Likewise.
3430 (tree_int_cst_sgn): Likewise.
3431 (tree_int_cst_min_precision): Take a signop rather than a bool.
3432 (simple_cst_equal): Use wide_int interfaces.
3433 (compare_tree_int): Likewise.
3434 (iterative_hash_expr): Likewise.
3435 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3437 (get_type_static_bounds): Use wide_int interfaces.
3438 (tree_int_cst_elt_check_failed): New.
3439 (build_common_tree_nodes): Reordered to set prec before filling in
3441 (int_cst_value): Check cst_and_fits_in_hwi.
3442 (widest_int_cst_value): Use wide_int interfaces.
3443 (upper_bound_in_type): Likewise.
3444 (lower_bound_in_type): Likewise.
3445 (num_ending_zeros): Likewise.
3446 (drop_tree_overflow): Likewise.
3447 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3448 (gen_conditions_for_pow_cst_base): Likewise.
3449 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3450 (group_case_labels_stmt): Use wide-int interfaces.
3451 (verify_gimple_assign_binary): Likewise.
3452 (print_loop): Likewise.
3453 * tree-chrec.c (tree_fold_binomial): Likewise.
3454 * tree-core.h (struct tree_base): Add int_length.
3455 (struct tree_int_cst): Change rep of value.
3456 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3457 (dr_may_alias_p): Likewise.
3458 (max_stmt_executions_tree): Likewise.
3459 * tree.def (INTEGER_CST): Update comment.
3460 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3461 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3462 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3463 (dequeue_and_dump): Use wide-int interfaces.
3464 * tree.h: Include wide-int.h.
3465 (NULL_TREE): Moved to earlier loc in file.
3466 (TREE_INT_CST_ELT_CHECK): New.
3467 (tree_int_cst_elt_check_failed): New.
3469 (TREE_INT_CST): Delete.
3470 (TREE_INT_CST_LOW): Use wide-int interfaces.
3471 (TREE_INT_CST_HIGH): Delete.
3472 (TREE_INT_CST_NUNITS): New.
3473 (TREE_INT_CST_EXT_NUNITS): Likewise.
3474 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3475 (TREE_INT_CST_ELT): Likewise.
3476 (INT_CST_LT): Delete.
3477 (tree_int_cst_elt_check): New (two forms).
3478 (type_code_size): Update comment.
3479 (make_int_cst_stat, make_int_cst): New.
3480 (tree_to_double_int): Delete.
3481 (double_int_fits_to_tree_p): Delete.
3482 (force_fit_type_double): Delete.
3483 (build_int_cstu): Replace with out-of-line function.
3484 (build_int_cst_wide): Delete.
3485 (tree_int_cst_lt): Define inline.
3486 (tree_int_cst_le): New.
3487 (tree_int_cst_compare): Define inline.
3488 (tree_int_cst_min_precision): Take a signop rather than a bool.
3489 (wi::int_traits <const_tree>): New.
3490 (wi::int_traits <tree>): New.
3491 (wi::extended_tree): New.
3492 (wi::int_traits <wi::extended_tree>): New.
3493 (wi::to_widest): New.
3494 (wi::to_offset): New.
3495 (wi::fits_to_tree_p): New.
3496 (wi::min_value): New.
3497 (wi::max_value): New.
3498 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3499 (copy_tree_body_r): Likewise.
3500 * tree-object-size.c (compute_object_offset): Likewise.
3501 (addr_object_size): Likewise.
3502 * tree-predcom.c: Include wide-int-print.h.
3503 (struct dref_d): Change type of offset to widest_int.
3504 (dump_dref): Call wide-int printer.
3505 (aff_combination_dr_offset): Use wide-int interfaces.
3506 (determine_offset): Take a widest_int pointer rather than a
3508 (split_data_refs_to_components): Use wide-int interfaces.
3509 (suitable_component_p): Likewise.
3510 (order_drefs): Likewise.
3511 (add_ref_to_chain): Likewise.
3512 (valid_initializer_p): Likewise.
3513 (determine_roots_comp): Likewise.
3514 * tree-pretty-print.c: Include wide-int-print.h.
3515 (dump_generic_node): Use wide-int interfaces.
3516 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3517 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3518 (move_fixed_address_to_symbol): Likewise.
3519 (move_hint_to_base): Likewise.
3520 (move_pointer_to_base): Likewise.
3521 (move_variant_to_index): Likewise.
3522 (most_expensive_mult_to_index): Likewise.
3523 (addr_to_parts): Likewise.
3524 (copy_ref_info): Likewise.
3525 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3526 (indirect_refs_may_alias_p): Likewise.
3527 (stmt_kills_ref_p_1): Likewise.
3528 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3529 * tree-ssa-ccp.c: Update comment at top of file. Include
3531 (struct prop_value_d): Change type of mask to widest_int.
3532 (extend_mask): New function.
3533 (dump_lattice_value): Use wide-int interfaces.
3534 (get_default_value): Likewise.
3535 (set_constant_value): Likewise.
3536 (set_value_varying): Likewise.
3537 (valid_lattice_transition): Likewise.
3538 (set_lattice_value): Likewise.
3539 (value_to_double_int): Delete.
3540 (value_to_wide_int): New.
3541 (get_value_from_alignment): Use wide-int interfaces.
3542 (get_value_for_expr): Likewise.
3543 (do_dbg_cnt): Likewise.
3544 (ccp_finalize): Likewise.
3545 (ccp_lattice_meet): Likewise.
3546 (bit_value_unop_1): Use widest_ints rather than double_ints.
3547 (bit_value_binop_1): Likewise.
3548 (bit_value_unop): Use wide-int interfaces.
3549 (bit_value_binop): Likewise.
3550 (bit_value_assume_aligned): Likewise.
3551 (evaluate_stmt): Likewise.
3552 (ccp_fold_stmt): Likewise.
3553 (visit_cond_stmt): Likewise.
3554 (ccp_visit_stmt): Likewise.
3555 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3556 (constant_pointer_difference): Likewise.
3557 (associate_pointerplus): Likewise.
3558 (combine_conversions): Likewise.
3559 * tree-ssa-loop.h: Include wide-int.h.
3560 (struct tree_niter_desc): Change type of max to widest_int.
3561 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3562 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3563 (remove_redundant_iv_tests): Likewise.
3564 (canonicalize_loop_induction_variables): Likewise.
3565 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3566 (constant_multiple_of): Take a widest_int pointer instead of
3567 a double_int pointer.
3568 (get_computation_aff): Use wide-int interfaces.
3569 (ptr_difference_cost): Likewise.
3570 (difference_cost): Likewise.
3571 (get_loop_invariant_expr_id): Likewise.
3572 (get_computation_cost_at): Likewise.
3573 (iv_elimination_compare_lt): Likewise.
3574 (may_eliminate_iv): Likewise.
3575 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3576 instead of double_int.
3577 (max_loop_iterations): Likewise.
3578 (max_stmt_executions): Likewise.
3579 (estimated_stmt_executions): Likewise.
3580 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3581 (split_to_var_and_offset): Use wide-int interfaces.
3582 (determine_value_range): Likewise.
3583 (bound_difference_of_offsetted_base): Likewise.
3584 (bounds_add): Take a widest_int instead of a double_int.
3585 (number_of_iterations_ne_max): Use wide-int interfaces.
3586 (number_of_iterations_ne): Likewise.
3587 (number_of_iterations_lt_to_ne): Likewise.
3588 (assert_loop_rolls_lt): Likewise.
3589 (number_of_iterations_lt): Likewise.
3590 (number_of_iterations_le): Likewise.
3591 (number_of_iterations_cond): Likewise.
3592 (number_of_iterations_exit): Likewise.
3593 (finite_loop_p): Likewise.
3594 (derive_constant_upper_bound_assign): Likewise.
3595 (derive_constant_upper_bound): Return a widest_int.
3596 (derive_constant_upper_bound_ops): Likewise.
3597 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3598 (record_estimate): Take a widest_int rather than a double_int.
3599 (record_nonwrapping_iv): Use wide-int interfaces.
3600 (double_int_cmp): Delete.
3601 (wide_int_cmp): New.
3602 (bound_index): Take a widest_int rather than a double_int.
3603 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3604 (maybe_lower_iteration_bound): Likewise.
3605 (estimate_numbers_of_iterations_loop): Likewise.
3606 (estimated_loop_iterations): Take a widest_int pointer than than
3607 a double_int pointer.
3608 (estimated_loop_iterations_int): Use wide-int interfaces.
3609 (max_loop_iterations): Take a widest_int pointer than than
3610 a double_int pointer.
3611 (max_loop_iterations_int): Use wide-int interfaces.
3612 (max_stmt_executions): Take a widest_int pointer than than
3613 a double_int pointer.
3614 (estimated_stmt_executions): Likewise.
3615 (n_of_executions_at_most): Use wide-int interfaces.
3616 (scev_probably_wraps_p): Likewise.
3617 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3619 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3621 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3622 double_ints. Adjust for trailing_wide_ints <3> representation.
3623 (set_nonzero_bits): Likewise.
3624 (get_range_info): Return wide_ints rather than double_ints.
3625 Adjust for trailing_wide_ints <3> representation.
3626 (get_nonzero_bits): Likewise.
3627 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3629 * tree-ssanames.h (struct range_info_def): Replace min, max and
3630 nonzero_bits with a trailing_wide_ints <3>.
3631 (set_range_info): Use wide_int_refs rather than double_ints.
3632 (set_nonzero_bits): Likewise.
3633 (get_range_info): Return wide_ints rather than double_ints.
3634 (get_nonzero_bits): Likewise.
3635 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3636 * tree-ssa-pre.c (phi_translate_1): Likewise.
3637 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3638 (acceptable_pow_call): Likewise.
3639 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3641 (vn_reference_fold_indirect): Likewise.
3642 (vn_reference_maybe_forwprop_address): Likewise.
3643 (valueize_refs_1): Likewise.
3644 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3645 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3646 tree_int_cst_lt and tree_int_cst_le.
3647 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3649 (streamer_alloc_tree): Likewise.
3650 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3651 (streamer_write_tree_header): Likewise.
3652 (streamer_write_integer_cst): Likewise.
3653 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3654 (build_constructors): Likewise.
3655 (array_value_type): Likewise.
3656 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3657 (vect_check_gather): Likewise.
3658 * tree-vect-generic.c (build_replicated_const): Likewise.
3659 (expand_vector_divmod): Likewise.
3660 * tree-vect-loop.c (vect_transform_loop): Likewise.
3661 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3662 (vect_do_peeling_for_alignment): Likewise.
3663 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3664 * tree-vrp.c: Include wide-int.h.
3665 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3666 (extract_range_from_assert): Use wide-int interfaces.
3667 (vrp_int_const_binop): Likewise.
3668 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3669 double_int pointers.
3670 (ranges_from_anti_range): Use wide-int interfaces.
3671 (quad_int_cmp): Delete.
3672 (quad_int_pair_sort): Likewise.
3673 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3674 (extract_range_from_unary_expr_1): Likewise.
3675 (adjust_range_with_scev): Likewise.
3676 (masked_increment): Take and return wide_ints rather than double_ints.
3677 (register_edge_assert_for_2): Use wide-int interfaces.
3678 (check_array_ref): Likewise.
3679 (search_for_addr_array): Likewise.
3680 (maybe_set_nonzero_bits): Likewise.
3681 (union_ranges): Pass an integer of the correct type instead of
3682 using integer_one_node.
3683 (intersect_ranges): Likewise.
3684 (simplify_truth_ops_using_ranges): Likewise.
3685 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3686 (range_fits_type_p): Likewise.
3687 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3689 (simplify_conversion_using_ranges): Use wide-int interfaces.
3690 (simplify_float_conversion_using_ranges): Likewise.
3691 (vrp_finalize): Likewise.
3692 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3693 (gimple_stringops_transform): Likewise.
3694 * varasm.c (decode_addr_const): Likewise.
3695 (const_hash_1): Likewise.
3696 (const_rtx_hash_1): Likewise
3697 (output_constant): Likewise.
3698 (array_size_for_constructor): Likewise.
3699 (output_constructor_regular_field): Likewise.
3700 (output_constructor_bitfield): Likewise.
3701 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3702 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3704 * gencheck.c: Define BITS_PER_UNIT.
3707 * wide-int-print.cc: New.
3708 * wide-int-print.h: New.
3710 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3712 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3714 2014-05-06 Richard Biener <rguenther@suse.de>
3716 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3717 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3718 (TODO_verify_all): Adjust.
3719 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3720 TODO_verify_stmts and TODO_verify_rtl_sharing.
3721 * bb-reorder.c: Likewise.
3722 * cfgexpand.c: Likewise.
3723 * cprop.c: Likewise.
3725 * function.c: Likewise.
3726 * fwprop.c: Likewise.
3728 * gimple-ssa-isolate-paths.c: Likewise.
3729 * gimple-ssa-strength-reduction.c: Likewise.
3730 * ipa-split.c: Likewise.
3731 * loop-init.c: Likewise.
3732 * loop-unroll.c: Likewise.
3733 * lower-subreg.c: Likewise.
3734 * modulo-sched.c: Likewise.
3735 * postreload-gcse.c: Likewise.
3736 * predict.c: Likewise.
3737 * recog.c: Likewise.
3738 * sched-rgn.c: Likewise.
3739 * store-motion.c: Likewise.
3740 * tracer.c: Likewise.
3741 * trans-mem.c: Likewise.
3742 * tree-call-cdce.c: Likewise.
3743 * tree-cfg.c: Likewise.
3744 * tree-cfgcleanup.c: Likewise.
3745 * tree-complex.c: Likewise.
3746 * tree-eh.c: Likewise.
3747 * tree-emutls.c: Likewise.
3748 * tree-if-conv.c: Likewise.
3749 * tree-into-ssa.c: Likewise.
3750 * tree-loop-distribution.c: Likewise.
3751 * tree-object-size.c: Likewise.
3752 * tree-parloops.c: Likewise.
3753 * tree-pass.h: Likewise.
3754 * tree-sra.c: Likewise.
3755 * tree-ssa-ccp.c: Likewise.
3756 * tree-ssa-copy.c: Likewise.
3757 * tree-ssa-copyrename.c: Likewise.
3758 * tree-ssa-dce.c: Likewise.
3759 * tree-ssa-dom.c: Likewise.
3760 * tree-ssa-dse.c: Likewise.
3761 * tree-ssa-forwprop.c: Likewise.
3762 * tree-ssa-ifcombine.c: Likewise.
3763 * tree-ssa-loop-ch.c: Likewise.
3764 * tree-ssa-loop-ivcanon.c: Likewise.
3765 * tree-ssa-loop.c: Likewise.
3766 * tree-ssa-math-opts.c: Likewise.
3767 * tree-ssa-phiopt.c: Likewise.
3768 * tree-ssa-phiprop.c: Likewise.
3769 * tree-ssa-pre.c: Likewise.
3770 * tree-ssa-reassoc.c: Likewise.
3771 * tree-ssa-sink.c: Likewise.
3772 * tree-ssa-strlen.c: Likewise.
3773 * tree-ssa-tail-merge.c: Likewise.
3774 * tree-ssa-uncprop.c: Likewise.
3775 * tree-switch-conversion.c: Likewise.
3776 * tree-tailcall.c: Likewise.
3777 * tree-vect-generic.c: Likewise.
3778 * tree-vectorizer.c: Likewise.
3779 * tree-vrp.c: Likewise.
3781 * var-tracking.c: Likewise.
3782 * bt-load.c: Likewise.
3783 * cfgcleanup.c: Likewise.
3784 * combine-stack-adj.c: Likewise.
3785 * combine.c: Likewise.
3786 * compare-elim.c: Likewise.
3787 * config/epiphany/resolve-sw-modes.c: Likewise.
3788 * config/i386/i386.c: Likewise.
3789 * config/mips/mips.c: Likewise.
3790 * config/s390/s390.c: Likewise.
3791 * config/sh/sh_treg_combine.cc: Likewise.
3792 * config/sparc/sparc.c: Likewise.
3795 * final.c: Likewise.
3796 * ifcvt.c: Likewise.
3797 * mode-switching.c: Likewise.
3798 * passes.c: Likewise.
3799 * postreload.c: Likewise.
3801 * reg-stack.c: Likewise.
3802 * regcprop.c: Likewise.
3803 * regrename.c: Likewise.
3806 2014-05-06 Richard Biener <rguenther@suse.de>
3809 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3810 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3812 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3815 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3817 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3818 Tom de Vries <tom@codesourcery.com>
3820 * target.def (call_fusage_contains_non_callee_clobbers): New
3822 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3824 (@node Miscellaneous Register Hooks): New node.
3825 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
3826 * doc/tm.texi: Regenerate.
3828 2014-05-05 Marek Polacek <polacek@redhat.com>
3831 * opts.c (common_handle_option): Call error_at instead of warning_at.
3833 2014-05-05 Richard Biener <rguenther@suse.de>
3835 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
3836 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
3837 under the TODO_verify_il umbrella.
3839 2014-05-05 Richard Biener <rguenther@suse.de>
3841 * passes.c (execute_function_todo): Move TODO_verify_flow under
3842 the TODO_verify_ul umbrella.
3844 2014-05-05 Richard Biener <rguenther@suse.de>
3847 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
3848 X & CST away from a CST that is the mask of a mode.
3850 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3852 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
3853 int argument to enum machine_mode.
3854 (picochip_class_max_nregs): Ditto.
3855 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
3856 (picochip_class_max_nregs): Ditto.
3858 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3860 * target.def: Add new target hook.
3861 * doc/tm.texi: Regenerate.
3862 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
3863 * targhooks.c (default_keep_leaf_when_profiled): New function.
3865 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
3866 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
3868 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3870 PR tree-optimization/60363
3871 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
3872 (copy_phi_args): New parameters. Call get_value_locus_in_path.
3873 (update_destination_phis): New parameter.
3874 (create_edge_and_update_destination_phis): Ditto.
3875 (ssa_fix_duplicate_block_edges): Pass new arguments.
3876 (thread_single_edge): Ditto.
3878 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
3880 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
3881 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
3882 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
3883 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
3884 Use RS6000_BTM_HARD_FLOAT.
3885 (BU_MISC_2): Likewise.
3886 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
3887 RS6000_BTM_HARD_FLOAT.
3888 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
3890 (rs6000_invalid_builtin): Add hard floating builtin support.
3891 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
3892 hard float builtins.
3893 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
3895 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3897 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
3898 Add missing function* argument.
3900 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3902 * lra-constraints.c (valid_address_p): Move earlier in file.
3903 Add a constraint argument to the address_info version.
3904 (satisfies_memory_constraint_p): New function.
3905 (satisfies_address_constraint_p): Likewise.
3906 (process_alt_operands, curr_insn_transform): Use them.
3907 (process_address): Pass the constraint to valid_address_p when
3908 checking address operands.
3910 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3912 * config/mips/mips.c (mips_isa_rev): New variable.
3913 (mips_set_architecture): Set it.
3914 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
3916 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
3917 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
3918 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
3919 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
3920 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
3921 conditions in terms of mips_isa_rev.
3922 (mips_isa_rev): Declare.
3924 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3926 * config/sh/sh-mem.cc: Use tabs instead of spaces.
3927 (prob_unlikely, prob_likely): Make variables const.
3929 2014-05-03 Denis Chertykov <chertykov@gmail.com>
3931 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
3933 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3935 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
3937 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3939 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
3940 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
3941 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
3943 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
3944 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
3946 Replace usage of ROUND_REG with sh_round_reg.
3947 Use CEIL instead of ROUND_ADVANCE.
3949 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
3952 * config/sh/sh.c: Include stdlib headers before everything else.
3954 2014-05-02 Jakub Jelinek <jakub@redhat.com>
3956 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
3957 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
3958 (gimplify_adjust_omp_clauses): Simd region is never
3959 directly nested in combined parallel. Instead, for linear
3960 with copyin/copyout, if in combined for simd loop, make decl
3961 firstprivate/lastprivate on OMP_FOR.
3962 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
3963 expand_omp_for_static_chunk): When setting endvar, also set
3964 fd->loop.v to the same value.
3966 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3968 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
3970 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
3972 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
3975 2014-05-02 Marek Polacek <polacek@redhat.com>
3977 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
3979 2014-05-02 Kito Cheng <kito@0xlab.org>
3981 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
3982 to a C expression marco.
3983 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
3984 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
3985 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
3986 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
3987 HONOR_REG_ALLOC_ORDER.
3988 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
3990 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3992 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
3994 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3996 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
3998 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4000 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4001 (convert_scalar_cond_reduction): Likewise.
4002 (predicate_scalar_phi): Add recognition and transformation
4003 of simple conditioanl reduction to be vectorizable.
4005 2014-05-01 Marek Polacek <polacek@redhat.com>
4008 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4010 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4012 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4013 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4014 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4015 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4016 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4017 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4018 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4019 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4021 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4023 * config/arc/arc.opt (mlra): Move comment above option name
4024 to avoid mis-parsing as language options.
4026 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4028 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4029 * config/sol2.h: ... here.
4030 * config/sol2-10.h: Remove.
4032 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4033 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4034 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4035 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4036 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4037 * config/sol2.h: ... here.
4038 (SECTION_NAME_FORMAT): Don't redefine.
4039 (STARTFILE_ARCH32_SPEC): Rename to ...
4040 (STARTFILE_ARCH_SPEC): ... this.
4041 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4042 * config/sparc/sol2.h: ... here.
4043 (SECTION_NAME_FORMAT): Don't undef.
4044 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4045 (SUBTARGET_EXTRA_SPECS): Remove.
4046 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4048 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4049 (MD_STARTFILE_PREFIX): Remove.
4050 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4051 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4052 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4053 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4054 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4055 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4056 * config/i386/sol2.h: ... here.
4057 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4058 * config/i386/sol2-bi.h: Remove.
4059 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4060 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4062 * config/i386/t-sol2-64: Rename to ...
4063 * config/i386/t-sol2: ... this.
4064 * config/sparc/t-sol2-64: Rename to ...
4065 * config/sparc/t-sol2: ... this.
4067 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4068 sol2_tm_file_head, sol2_tm_file_tail.
4069 Include ${cpu_type}/sol2.h before sol2.h.
4071 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4072 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4073 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4074 Reflect i386/t-sol2-64 renaming.
4075 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4076 Reflect sparc/t-sol2-64 renaming.
4078 2014-04-30 Richard Biener <rguenther@suse.de>
4080 * passes.c (execute_function_todo): Move TODO_verify_stmts
4081 and TODO_verify_ssa under the TODO_verify_il umbrella.
4082 * tree-ssa.h (verify_ssa): Adjust prototype.
4083 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4084 we should verify SSA operands.
4085 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4086 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4087 whether we should verify whether not throwing stmts have EH info.
4088 * graphite-scop-detection.c (create_sese_edges): Adjust.
4089 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4090 * tree-eh.c (lower_try_finally_switch): Do not add the
4091 default case label twice.
4093 2014-04-30 Marek Polacek <polacek@redhat.com>
4095 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4096 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4097 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4098 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4100 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4102 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4103 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4104 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4105 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4106 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4107 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4108 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4109 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4111 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4113 * tree-cfg.c (dump_function_to_file): Dump the return type of
4114 functions, in a line to itself before the function body, mimicking
4115 the layout of a C function.
4117 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4119 PR tree-optimization/60971
4120 * tree-tailcall.c (process_assignment): Reject conversions which
4123 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4125 * calls.c (initialize_argument_information): Always treat
4126 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4127 (expand_call): Likewise.
4128 (emit_library_call_calue_1): Likewise.
4129 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4130 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4133 2014-04-29 Nick Clifton <nickc@redhat.com>
4135 * config/msp430/msp430.md (umulsidi): Fix typo.
4136 (mulhisi3): Enable even inside interrupt handlers.
4137 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4138 bigger return address pushed in large mode.
4140 2014-04-29 Nick Clifton <nickc@redhat.com>
4142 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4143 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4145 * config/m32r/m32r.c (init_reg_tables): Likewise.
4146 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4147 enum to hold the modes.
4149 2014-04-29 Richard Biener <rguenther@suse.de>
4151 * dominance.c (free_dominance_info): Add overload with
4153 (dom_info_state): Likewise.
4154 (dom_info_available_p): Likewise.
4155 * basic-block.h (free_dominance_info, dom_info_state,
4156 dom_info_available_p): Declare overloads.
4157 * passes.c (execute_function_todo): Verify that verifiers
4158 don't change dominator info state. Drop dominator info
4159 for IPA pass invocations.
4160 * cgraph.c (release_function_body): Restore asserts that
4161 dominator information is released.
4163 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4165 * doc/invoke.texi: Fix typo.
4166 * tree-vrp.c: Fix typos.
4167 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4169 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4171 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4173 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4175 * config/aarch64/aarch64-builtins.c
4176 (aarch64_types_storestruct_lane_qualifiers): New.
4177 (TYPES_STORESTRUCT_LANE): Likewise.
4178 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4179 (st3_lane): Likewise.
4180 (st4_lane): Likewise.
4181 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4182 (vec_store_lanesci_lane<mode>): Likewise.
4183 (vec_store_lanesxi_lane<mode>): Likewise.
4184 (aarch64_st2_lane<VQ:mode>): Likewise.
4185 (aarch64_st3_lane<VQ:mode>): Likewise.
4186 (aarch64_st4_lane<VQ:mode>): Likewise.
4187 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4188 * config/aarch64/arm_neon.h
4189 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4190 use new macro arguments.
4191 (__ST3_LANE_FUNC): Likewise.
4192 (__ST4_LANE_FUNC): Likewise.
4193 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4194 (V_THREE_ELEM): Likewise.
4195 (V_FOUR_ELEM): Likewise.
4197 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4199 * doc/gimple.texi: Replace the description of the now-defunct
4200 union gimple_statement_d with a diagram showing the
4201 gimple_statement_base class hierarchy and its relationships to
4202 the GSS_ and GIMPLE_ enums.
4204 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4206 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4207 * config/aarch64/aarch64.c
4208 (aarch64_cannot_change_mode_class): Weaken conditions.
4209 (aarch64_modes_tieable_p): New.
4210 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4212 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4214 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4215 (loadsync_<mode>): Change mode.
4216 (load_quadpti, store_quadpti): New.
4217 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4218 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4220 2014-04-28 Martin Jambor <mjambor@suse.cz>
4222 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4223 same alias type as the original statement.
4224 (subreplacement_assignment_data): New type.
4225 (handle_unscalarized_data_in_subtree): New type of parameter,
4226 generate new memory accesses with same alias type as the original
4228 (load_assign_lhs_subreplacements): Likewise.
4229 (sra_modify_constructor_assign): Generate new memory accesses with
4230 same alias type as the original statement.
4232 2014-04-28 Richard Biener <rguenther@suse.de>
4234 * tree-pass.h (TODO_verify_il): Define.
4235 (TODO_verify_all): Complete properly.
4236 * passes.c (execute_function_todo): Move existing loop-closed
4237 SSA verification under TODO_verify_il.
4238 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4239 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4240 Fix tree sharing issue.
4242 2014-04-28 Richard Biener <rguenther@suse.de>
4245 * builtins.def (DEF_C11_BUILTIN): Add.
4246 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4247 * coretypes.h (enum function_class): Add function_c11_misc.
4248 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4249 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4250 (call_may_clobber_ref_p_1): Likewise.
4251 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4252 (mark_all_reaching_defs_necessary_1): Likewise.
4253 (propagate_necessity): Likewise.
4254 (eliminate_unnecessary_stmts): Likewise.
4255 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4257 2014-04-28 Richard Biener <rguenther@suse.de>
4259 * tree-vrp.c (vrp_var_may_overflow): Remove.
4260 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4261 with overflow immediately bump to one before that value and
4262 let iteration figure out overflow status.
4264 2014-04-28 Richard Biener <rguenther@suse.de>
4266 * configure.ac: Do valgrind header checks unconditionally.
4267 Add --enable-valgrind-annotations.
4268 * system.h: Guard valgrind header inclusion with
4269 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4270 * alloc-pool.c (pool_alloc, pool_free): Use
4271 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4272 to guard possibly dead code.
4273 * config.in: Regenerated.
4274 * configure: Likewise.
4276 2014-04-28 Jeff Law <law@redhat.com>
4278 PR tree-optimization/60902
4279 * tree-ssa-threadedge.c
4280 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4281 over real defs when invalidating outputs from statements that do not
4282 produce useful outputs for threading.
4284 2014-04-28 Richard Biener <rguenther@suse.de>
4286 PR tree-optimization/60979
4287 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4288 SCOPs that end in a block with a successor with abnormal
4291 2014-04-28 Richard Biener <rguenther@suse.de>
4293 * tree-pass.h (execute_pass_list): Adjust prototype.
4294 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4295 (do_per_function): Change callback signature, push all actual
4296 work to the callbals.
4297 (do_per_function_toporder): Likewise.
4298 (execute_function_dump): Adjust.
4299 (execute_function_todo): Likewise.
4300 (clear_last_verified): Likewise.
4301 (verify_curr_properties): Likewise.
4302 (update_properties_after_pass): Likewise.
4303 (execute_pass_list_1): Split out from ...
4304 (execute_pass_list): ... here. Adjust.
4305 (execute_ipa_pass_list): Likewise.
4306 * cgraphunit.c (cgraph_add_new_function): Adjust.
4307 (analyze_function): Likewise.
4308 (expand_function): Likewise.
4309 * cgraph.c (release_function_body): Free dominance info
4310 here instead of asserting it was magically freed elsewhere.
4312 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4314 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4315 * configure: Regenerate.
4316 * config/sparc/sparc.opt (muser-mode): New option.
4317 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4319 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4320 * doc/invoke.texi (SPARC options): Document -muser-mode.
4322 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4324 * cselib.c (find_slot_memmode): Delete.
4325 (cselib_hasher): Change compare_type to a struct.
4326 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4328 (preserve_constants_and_equivs): Adjust for new compare_type.
4329 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4330 (wrap_constant): Delete.
4331 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4333 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4335 * doc/install.texi (Building with profile feedback): Remove
4338 2014-04-26 Tom de Vries <tom@codesourcery.com>
4340 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4343 2014-04-25 Cary Coutant <ccoutant@google.com>
4346 * dwarf2out.c (should_move_die_to_comdat): A type definition
4347 can contain a subprogram definition, but don't move it to a
4349 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4350 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4352 (clone_tree_hash): Rename to...
4353 (clone_tree_partial): ...this; change callers. Copy
4354 DW_TAG_subprogram DIEs as declarations.
4355 (copy_decls_walk): Don't copy children of a declaration into a
4358 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4361 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4364 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4366 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4367 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4369 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4372 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4374 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4376 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4378 PR tree-optimization/60930
4379 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4380 creating a multiply candidate by folding two constant
4381 multiplicands when the result overflows.
4383 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4385 PR tree-optimization/60960
4386 * tree-vect-generic.c (expand_vector_operation): Only call
4387 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4389 2014-04-25 Tom de Vries <tom@codesourcery.com>
4391 * expr.c (clobber_reg_mode): New function.
4392 * expr.h (clobber_reg): New function.
4394 2014-04-25 Tom de Vries <tom@codesourcery.com>
4396 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4399 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4400 Tom de Vries <tom@codesourcery.com>
4402 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4404 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4405 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4406 new argument to find_all_hard_reg_sets call.
4408 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4410 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4411 Use HOST_WIDE_INT_C for mask literal.
4412 (aarch_rev16_shleft_mask_imm_p): Likewise.
4414 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4417 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4419 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4421 PR preprocessor/56540
4422 * config/i386/i386-c.c (ix86_target_macros): Define
4423 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4425 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4427 * configure.ac (tga_func): Remove.
4428 (LIB_TLS_SPEC): Remove.
4429 * configure: Regenerate.
4430 * config.in: Regenerate.
4431 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4433 2014-04-25 Richard Biener <rguenther@suse.de>
4436 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4437 call stmt use/clobber sets during stmt walk instead of
4438 walking the possibly incomplete set of caller edges.
4440 2014-04-25 Richard Biener <rguenther@suse.de>
4443 * passes.c (apply_ipa_transforms): Inline into only caller ...
4444 (execute_one_pass): ... here. Properly bring in function
4445 bodies for nodes we want to apply IPA transforms to.
4447 2014-04-24 Cong Hou <congh@google.com>
4449 PR tree-optimization/60896
4450 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4451 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4452 (vect_mark_pattern_stmts): Set the def type of all statements in
4453 PATTERN_DEF_SEQ as vect_internal_def.
4455 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4457 * doc/extend.texi (PowerPC Built-in Functions): Document new
4458 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4459 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4461 * config/rs6000/predicates.md (const_0_to_3_operand): New
4462 predicate to match 0..3 integer constants.
4464 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4465 to support adding miscellaneous builtin functions.
4466 (BU_DFP_MISC_2): Likewise.
4467 (BU_P7_MISC_1): Likewise.
4468 (BU_P7_MISC_2): Likewise.
4469 (BU_P8V_MISC_3): Likewise.
4470 (BU_MISC_1): Likewise.
4471 (BU_MISC_2): Likewise.
4472 (DIVWE): Add extended divide builtin functions.
4475 (DIVWEUO): Likewise.
4479 (DIVDEUO): Likewise.
4480 (DXEX): Add decimal floating-point builtin functions.
4483 (DDEDPDQ): Likewise.
4485 (DENBCDQ): Likewise.
4492 (CDTBCD): Add new BCD builtin functions.
4496 (BCDADD_LT): Likewise.
4497 (BCDADD_EQ): Likewise.
4498 (BCDADD_GT): Likewise.
4499 (BCDADD_OV): Likewise.
4501 (BCDSUB_LT): Likewise.
4502 (BCDSUB_EQ): Likewise.
4503 (BCDSUB_GT): Likewise.
4504 (BCDSUB_OV): Likewise.
4505 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4506 (UNPACK_TD): Likewise.
4507 (PACK_TF): Likewise.
4508 (UNPACK_TF): Likewise.
4509 (UNPACK_TF_0): Likewise.
4510 (UNPACK_TF_1): Likewise.
4511 (PACK_V1TI): Likewise.
4512 (UNPACK_V1TI): Likewise.
4514 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4515 support for decimal floating point builtin functions.
4516 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4517 functions that take constant arguments.
4518 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4519 (rs6000_init_builtins): Setup long double, _Decimal64, and
4520 _Decimal128 types for new builtin functions.
4521 (builtin_function_type): Set the unsigned flags appropriately for
4522 the new builtin functions.
4523 (rs6000_opt_masks): Add support for decimal floating point builtin
4526 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4527 floating point builtin functions.
4528 (RS6000_BTM_COMMON): Likewise.
4529 (RS6000_BTI_long_double): Likewise.
4530 (RS6000_BTI_dfloat64): Likewise.
4531 (RS6000_BTI_dfloat128): Likewise.
4532 (long_double_type_internal_node): Likewise.
4533 (dfloat64_type_internal_node): Likewise.
4534 (dfloat128_type_internal_node): Likewise.
4536 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4537 2.07 bcd arithmetic instructions.
4538 (UNSPEC_BCDSUB): Likewise.
4539 (UNSPEC_BCD_OVERFLOW): Likewise.
4540 (UNSPEC_BCD_ADD_SUB): Likewise.
4541 (bcd_add_sub): Likewise.
4542 (BCD_TEST): Likewise.
4543 (bcd<bcd_add_sub>): Likewise.
4544 (bcd<bcd_add_sub>_test): Likewise.
4545 (bcd<bcd_add_sub>_test2): Likewise.
4546 (bcd<bcd_add_sub>_<code>): Likewise.
4547 (peephole2 for combined bcd ops): Likewise.
4549 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4550 decimal floating point builtin functions.
4551 (UNSPEC_DENBCD): Likewise.
4552 (UNSPEC_DXEX): Likewise.
4553 (UNSPEC_DIEX): Likewise.
4554 (UNSPEC_DSCLI): Likewise.
4555 (UNSPEC_DSCRI): Likewise.
4556 (D64_D128): Likewise.
4557 (dfp_suffix): Likewise.
4558 (dfp_ddedpd_<mode>): Likewise.
4559 (dfp_denbcd_<mode>): Likewise.
4560 (dfp_dxex_<mode>): Likewise.
4561 (dfp_diex_<mode>): Likewise.
4562 (dfp_dscli_<mode>): Likewise.
4563 (dfp_dscri_<mode>): Likewise.
4565 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4567 (UNSPEC_CDTBCD): Likewise.
4568 (UNSPEC_CBCDTD): Likewise.
4569 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4570 (UNSPEC_DIVEO): Likewise.
4571 (UNSPEC_DIVEU): Likewise.
4572 (UNSPEC_DIVEUO): Likewise.
4573 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4574 pack/unpack 128-bit types.
4575 (UNSPEC_PACK_128BIT): Likewise.
4576 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4577 (udiv<mode>3): Use idiv_ldiv mode attribute.
4578 (div<mode>3): Likewise.
4579 (addg6s): Add new BCD builtin functions.
4582 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4583 (div_extend): Likewise.
4584 (div<div_extend>_<mode>"): Likewise.
4585 (FP128_64): Add support for new builtin functions to pack/unpack
4587 (unpack<mode>): Likewise.
4588 (unpacktf_0): Likewise.
4589 (unpacktf_1): Likewise.
4590 (unpack<mode>_dm): Likewise.
4591 (unpack<mode>_nodm): Likewise.
4592 (pack<mode>): Likewise.
4593 (unpackv1ti): Likewise.
4594 (packv1ti): Likewise.
4596 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4598 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4601 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4603 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4604 * gimplify.c (omp_is_private): Change last argument's type to int.
4605 Only diagnose lastprivate if the simd argument is 1, only diagnose
4606 linear if the simd argument is 2.
4607 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4608 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4609 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4610 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4611 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4612 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4613 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4614 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4615 * tree-nested.c (convert_nonlocal_omp_clauses,
4616 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4618 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4621 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4624 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4626 * flag-types.h (enum ivar_visibility): Add.
4628 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4630 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4631 function * argument.
4633 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4635 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4637 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4638 Tom de Vries <tom@codesourcery.com>
4640 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4641 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4643 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4644 * emit-rtl.c (try_split): Same.
4646 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4647 Tom de Vries <tom@codesourcery.com>
4649 * common.opt (fuse-caller-save): New option.
4651 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4653 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4654 elements for big-endian.
4656 2014-04-24 Richard Biener <rguenther@suse.de>
4658 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4659 during TER and instead use the sepops interface for expanding
4660 non-GIMPLE_SINGLE_RHS.
4662 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4664 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4665 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4667 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4669 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4670 assembler 64-bit option.
4671 * configure: Regenerate.
4673 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4675 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4676 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4677 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4678 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4679 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4681 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4683 * config/aarch64/aarch64-builtins.c
4684 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4685 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4686 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4687 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4689 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4690 (Vrevsuff): New mode attribute.
4692 2014-04-24 Terry Guo <terry.guo@arm.com>
4694 * config/arm/arm.h (machine_function): Define variable
4695 after_arm_reorg here.
4696 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4697 (arm_split_constant): Update the way to access variable
4700 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4702 2014-04-23 Tom de Vries <tom@codesourcery.com>
4704 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4706 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4708 * is-a.h: Update comments to reflect the following changes to the
4709 "pointerness" of the API, making the template parameter match the
4710 return type, allowing use of is-a.h with typedefs of pointers.
4711 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4712 that the return type matches the parameter to the is_a_helper.
4714 (dyn_cast): Likewise.
4716 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4717 pointer from the is-a.h API.
4719 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4720 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4722 (is_a_helper <varpool_node>::test): Likewise, convert to...
4723 (is_a_helper <varpool_node *>::test): ...this.
4725 (varpool_first_variable): Update for removal of implicit pointer
4726 from the is-a.h API.
4727 (varpool_next_variable): Likewise.
4728 (varpool_first_static_initializer): Likewise.
4729 (varpool_next_static_initializer): Likewise.
4730 (varpool_first_defined_variable): Likewise.
4731 (varpool_next_defined_variable): Likewise.
4732 (cgraph_first_defined_function): Likewise.
4733 (cgraph_next_defined_function): Likewise.
4734 (cgraph_first_function): Likewise.
4735 (cgraph_next_function): Likewise.
4736 (cgraph_first_function_with_gimple_body): Likewise.
4737 (cgraph_next_function_with_gimple_body): Likewise.
4738 (cgraph_alias_target): Likewise.
4739 (varpool_alias_target): Likewise.
4740 (cgraph_function_or_thunk_node): Likewise.
4741 (varpool_variable_node): Likewise.
4742 (symtab_real_symbol_p): Likewise.
4743 * cgraphunit.c (referred_to_p): Likewise.
4744 (analyze_functions): Likewise.
4745 (handle_alias_pairs): Likewise.
4746 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4747 * gimple-ssa.h (gimple_vuse_op): Likewise.
4748 (gimple_vdef_op): Likewise.
4749 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4750 * gimple.c (gimple_build_asm_1): Likewise.
4751 (gimple_build_try): Likewise.
4752 (gimple_build_resx): Likewise.
4753 (gimple_build_eh_dispatch): Likewise.
4754 (gimple_build_omp_for): Likewise.
4755 (gimple_omp_for_set_clauses): Likewise.
4757 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4758 (is_a_helper <gimple_statement_asm *>::test): ...this.
4759 (is_a_helper <gimple_statement_bind>::test): Convert to...
4760 (is_a_helper <gimple_statement_bind *>::test): ...this.
4761 (is_a_helper <gimple_statement_call>::test): Convert to...
4762 (is_a_helper <gimple_statement_call *>::test): ...this.
4763 (is_a_helper <gimple_statement_catch>::test): Convert to...
4764 (is_a_helper <gimple_statement_catch *>::test): ...this.
4765 (is_a_helper <gimple_statement_resx>::test): Convert to...
4766 (is_a_helper <gimple_statement_resx *>::test): ...this.
4767 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4768 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4769 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4770 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4771 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4772 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4773 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4774 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4775 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4776 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4777 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4778 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4779 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4780 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4781 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4782 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4783 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4784 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4785 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4786 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4787 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4788 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4789 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4790 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4791 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4792 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4793 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4794 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4795 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4796 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4797 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4798 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4799 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4800 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4801 (is_a_helper <gimple_statement_phi>::test): Convert to...
4802 (is_a_helper <gimple_statement_phi *>::test): ...this.
4803 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4804 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4805 (is_a_helper <gimple_statement_try>::test): Convert to...
4806 (is_a_helper <gimple_statement_try *>::test): ...this.
4807 (is_a_helper <gimple_statement_wce>::test): Convert to...
4808 (is_a_helper <gimple_statement_wce *>::test): ...this.
4809 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4810 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4811 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4812 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4813 (is_a_helper <const gimple_statement_call>::test): Convert to...
4814 (is_a_helper <const gimple_statement_call *>::test): ...this.
4815 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4816 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4817 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4818 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4819 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4820 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4821 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4822 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4823 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
4825 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
4827 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
4829 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
4831 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
4832 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
4833 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
4835 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
4836 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
4838 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
4839 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
4840 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
4841 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
4842 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
4843 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
4845 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
4846 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
4847 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
4848 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
4850 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
4851 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
4852 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
4853 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
4854 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
4855 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
4856 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
4857 (is_a_helper <const gimple_statement_phi>::test): Convert to...
4858 (is_a_helper <const gimple_statement_phi *>::test): ...this.
4859 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
4860 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
4861 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
4862 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
4863 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
4864 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
4865 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
4867 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
4869 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
4870 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
4872 (gimple_use_ops): Update for removal of implicit pointer from the
4874 (gimple_set_use_ops): Likewise.
4875 (gimple_vuse): Likewise.
4876 (gimple_vdef): Likewise.
4877 (gimple_vuse_ptr): Likewise.
4878 (gimple_vdef_ptr): Likewise.
4879 (gimple_set_vuse): Likewise.
4880 (gimple_set_vdef): Likewise.
4881 (gimple_omp_return_set_lhs): Likewise.
4882 (gimple_omp_return_lhs): Likewise.
4883 (gimple_omp_return_lhs_ptr): Likewise.
4884 (gimple_call_fntype): Likewise.
4885 (gimple_call_set_fntype): Likewise.
4886 (gimple_call_set_internal_fn): Likewise.
4887 (gimple_call_use_set): Likewise.
4888 (gimple_call_clobber_set): Likewise.
4889 (gimple_bind_vars): Likewise.
4890 (gimple_bind_set_vars): Likewise.
4891 (gimple_bind_body_ptr): Likewise.
4892 (gimple_bind_set_body): Likewise.
4893 (gimple_bind_add_stmt): Likewise.
4894 (gimple_bind_block): Likewise.
4895 (gimple_bind_set_block): Likewise.
4896 (gimple_asm_ninputs): Likewise.
4897 (gimple_asm_noutputs): Likewise.
4898 (gimple_asm_nclobbers): Likewise.
4899 (gimple_asm_nlabels): Likewise.
4900 (gimple_asm_input_op): Likewise.
4901 (gimple_asm_input_op_ptr): Likewise.
4902 (gimple_asm_output_op): Likewise.
4903 (gimple_asm_output_op_ptr): Likewise.
4904 (gimple_asm_set_output_op): Likewise.
4905 (gimple_asm_clobber_op): Likewise.
4906 (gimple_asm_set_clobber_op): Likewise.
4907 (gimple_asm_label_op): Likewise.
4908 (gimple_asm_set_label_op): Likewise.
4909 (gimple_asm_string): Likewise.
4910 (gimple_catch_types): Likewise.
4911 (gimple_catch_types_ptr): Likewise.
4912 (gimple_catch_handler_ptr): Likewise.
4913 (gimple_catch_set_types): Likewise.
4914 (gimple_catch_set_handler): Likewise.
4915 (gimple_eh_filter_types): Likewise.
4916 (gimple_eh_filter_types_ptr): Likewise.
4917 (gimple_eh_filter_failure_ptr): Likewise.
4918 (gimple_eh_filter_set_types): Likewise.
4919 (gimple_eh_filter_set_failure): Likewise.
4920 (gimple_eh_must_not_throw_fndecl): Likewise.
4921 (gimple_eh_must_not_throw_set_fndecl): Likewise.
4922 (gimple_eh_else_n_body_ptr): Likewise.
4923 (gimple_eh_else_e_body_ptr): Likewise.
4924 (gimple_eh_else_set_n_body): Likewise.
4925 (gimple_eh_else_set_e_body): Likewise.
4926 (gimple_try_eval_ptr): Likewise.
4927 (gimple_try_cleanup_ptr): Likewise.
4928 (gimple_try_set_eval): Likewise.
4929 (gimple_try_set_cleanup): Likewise.
4930 (gimple_wce_cleanup_ptr): Likewise.
4931 (gimple_wce_set_cleanup): Likewise.
4932 (gimple_phi_capacity): Likewise.
4933 (gimple_phi_num_args): Likewise.
4934 (gimple_phi_result): Likewise.
4935 (gimple_phi_result_ptr): Likewise.
4936 (gimple_phi_set_result): Likewise.
4937 (gimple_phi_arg): Likewise.
4938 (gimple_phi_set_arg): Likewise.
4939 (gimple_resx_region): Likewise.
4940 (gimple_resx_set_region): Likewise.
4941 (gimple_eh_dispatch_region): Likewise.
4942 (gimple_eh_dispatch_set_region): Likewise.
4943 (gimple_omp_critical_name): Likewise.
4944 (gimple_omp_critical_name_ptr): Likewise.
4945 (gimple_omp_critical_set_name): Likewise.
4946 (gimple_omp_for_clauses): Likewise.
4947 (gimple_omp_for_clauses_ptr): Likewise.
4948 (gimple_omp_for_set_clauses): Likewise.
4949 (gimple_omp_for_collapse): Likewise.
4950 (gimple_omp_for_index): Likewise.
4951 (gimple_omp_for_index_ptr): Likewise.
4952 (gimple_omp_for_set_index): Likewise.
4953 (gimple_omp_for_initial): Likewise.
4954 (gimple_omp_for_initial_ptr): Likewise.
4955 (gimple_omp_for_set_initial): Likewise.
4956 (gimple_omp_for_final): Likewise.
4957 (gimple_omp_for_final_ptr): Likewise.
4958 (gimple_omp_for_set_final): Likewise.
4959 (gimple_omp_for_incr): Likewise.
4960 (gimple_omp_for_incr_ptr): Likewise.
4961 (gimple_omp_for_set_incr): Likewise.
4962 (gimple_omp_for_pre_body_ptr): Likewise.
4963 (gimple_omp_for_set_pre_body): Likewise.
4964 (gimple_omp_parallel_clauses): Likewise.
4965 (gimple_omp_parallel_clauses_ptr): Likewise.
4966 (gimple_omp_parallel_set_clauses): Likewise.
4967 (gimple_omp_parallel_child_fn): Likewise.
4968 (gimple_omp_parallel_child_fn_ptr): Likewise.
4969 (gimple_omp_parallel_set_child_fn): Likewise.
4970 (gimple_omp_parallel_data_arg): Likewise.
4971 (gimple_omp_parallel_data_arg_ptr): Likewise.
4972 (gimple_omp_parallel_set_data_arg): Likewise.
4973 (gimple_omp_task_clauses): Likewise.
4974 (gimple_omp_task_clauses_ptr): Likewise.
4975 (gimple_omp_task_set_clauses): Likewise.
4976 (gimple_omp_task_child_fn): Likewise.
4977 (gimple_omp_task_child_fn_ptr): Likewise.
4978 (gimple_omp_task_set_child_fn): Likewise.
4979 (gimple_omp_task_data_arg): Likewise.
4980 (gimple_omp_task_data_arg_ptr): Likewise.
4981 (gimple_omp_task_set_data_arg): Likewise.
4982 (gimple_omp_taskreg_clauses): Likewise.
4983 (gimple_omp_taskreg_clauses_ptr): Likewise.
4984 (gimple_omp_taskreg_set_clauses): Likewise.
4985 (gimple_omp_taskreg_child_fn): Likewise.
4986 (gimple_omp_taskreg_child_fn_ptr): Likewise.
4987 (gimple_omp_taskreg_set_child_fn): Likewise.
4988 (gimple_omp_taskreg_data_arg): Likewise.
4989 (gimple_omp_taskreg_data_arg_ptr): Likewise.
4990 (gimple_omp_taskreg_set_data_arg): Likewise.
4991 (gimple_omp_task_copy_fn): Likewise.
4992 (gimple_omp_task_copy_fn_ptr): Likewise.
4993 (gimple_omp_task_set_copy_fn): Likewise.
4994 (gimple_omp_task_arg_size): Likewise.
4995 (gimple_omp_task_arg_size_ptr): Likewise.
4996 (gimple_omp_task_set_arg_size): Likewise.
4997 (gimple_omp_task_arg_align): Likewise.
4998 (gimple_omp_task_arg_align_ptr): Likewise.
4999 (gimple_omp_task_set_arg_align): Likewise.
5000 (gimple_omp_single_clauses): Likewise.
5001 (gimple_omp_single_clauses_ptr): Likewise.
5002 (gimple_omp_single_set_clauses): Likewise.
5003 (gimple_omp_target_clauses): Likewise.
5004 (gimple_omp_target_clauses_ptr): Likewise.
5005 (gimple_omp_target_set_clauses): Likewise.
5006 (gimple_omp_target_child_fn): Likewise.
5007 (gimple_omp_target_child_fn_ptr): Likewise.
5008 (gimple_omp_target_set_child_fn): Likewise.
5009 (gimple_omp_target_data_arg): Likewise.
5010 (gimple_omp_target_data_arg_ptr): Likewise.
5011 (gimple_omp_target_set_data_arg): Likewise.
5012 (gimple_omp_teams_clauses): Likewise.
5013 (gimple_omp_teams_clauses_ptr): Likewise.
5014 (gimple_omp_teams_set_clauses): Likewise.
5015 (gimple_omp_sections_clauses): Likewise.
5016 (gimple_omp_sections_clauses_ptr): Likewise.
5017 (gimple_omp_sections_set_clauses): Likewise.
5018 (gimple_omp_sections_control): Likewise.
5019 (gimple_omp_sections_control_ptr): Likewise.
5020 (gimple_omp_sections_set_control): Likewise.
5021 (gimple_omp_for_set_cond): Likewise.
5022 (gimple_omp_for_cond): Likewise.
5023 (gimple_omp_atomic_store_set_val): Likewise.
5024 (gimple_omp_atomic_store_val): Likewise.
5025 (gimple_omp_atomic_store_val_ptr): Likewise.
5026 (gimple_omp_atomic_load_set_lhs): Likewise.
5027 (gimple_omp_atomic_load_lhs): Likewise.
5028 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5029 (gimple_omp_atomic_load_set_rhs): Likewise.
5030 (gimple_omp_atomic_load_rhs): Likewise.
5031 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5032 (gimple_omp_continue_control_def): Likewise.
5033 (gimple_omp_continue_control_def_ptr): Likewise.
5034 (gimple_omp_continue_set_control_def): Likewise.
5035 (gimple_omp_continue_control_use): Likewise.
5036 (gimple_omp_continue_control_use_ptr): Likewise.
5037 (gimple_omp_continue_set_control_use): Likewise.
5038 (gimple_transaction_body_ptr): Likewise.
5039 (gimple_transaction_label): Likewise.
5040 (gimple_transaction_label_ptr): Likewise.
5041 (gimple_transaction_set_body): Likewise.
5042 (gimple_transaction_set_label): Likewise.
5044 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5045 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5046 * ipa-ref.c (ipa_record_reference): Likewise.
5047 * ipa-reference.c (analyze_function): Likewise.
5048 (ipa_reference_write_optimization_summary): Likewise.
5049 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5050 (address_taken_from_non_vtable_p): Likewise.
5051 (comdat_can_be_unshared_p_1): Likewise.
5052 * lto-cgraph.c (lto_output_ref): Likewise.
5053 (add_references): Likewise.
5054 (compute_ltrans_boundary): Likewise.
5055 (output_symtab): Likewise.
5056 (input_ref): Likewise.
5057 (input_cgraph_1): Likewise.
5058 (output_cgraph_opt_summary): Likewise.
5059 * lto-streamer-out.c (lto_output): Likewise.
5060 (output_symbol_p): Likewise.
5061 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5062 (lsei_start_function_in_partition): Likewise.
5063 (lsei_next_variable_in_partition): Likewise.
5064 (lsei_start_variable_in_partition): Likewise.
5065 * symtab.c (insert_to_assembler_name_hash): Likewise.
5066 (unlink_from_assembler_name_hash): Likewise.
5067 (symtab_unregister_node): Likewise.
5068 (symtab_remove_node): Likewise.
5069 (dump_symtab_node): Likewise.
5070 (verify_symtab_base): Likewise.
5071 (verify_symtab_node): Likewise.
5072 (symtab_make_decl_local): Likewise.
5073 (symtab_alias_ultimate_target): Likewise.
5074 (symtab_resolve_alias): Likewise.
5075 (symtab_get_symbol_partitioning_class): Likewise.
5076 * tree-phinodes.c (allocate_phi_node): Likewise.
5077 (reserve_phi_args_for_new_edge): Likewise.
5078 (remove_phi_args): Likewise.
5079 * varpool.c (varpool_node_for_asm): Likewise.
5080 (varpool_remove_unreferenced_decls): Likewise.
5082 2014-04-23 Jeff Law <law@redhat.com>
5084 PR tree-optimization/60902
5085 * tree-ssa-threadedge.c
5086 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5087 invalidate outputs from statements that do not produce useful
5088 outputs for threading.
5090 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5092 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5093 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5094 machine descriptions for Stack Smashing Protector.
5096 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5098 * aarch64.md (<optab>_rol<mode>3): New pattern.
5099 (<optab>_rolsi3_uxtw): Likewise.
5100 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5102 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5104 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5105 (arm_cortex_a12_tune): Likewise.
5107 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5109 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5111 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5113 * config/arm/arm.md (arm_rev16si2): New pattern.
5114 (arm_rev16si2_alt): Likewise.
5115 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5117 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5119 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5120 (rev16<mode>2_alt): Likewise.
5121 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5122 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5123 (aarch_rev16_shleft_mask_imm_p): Likewise.
5124 (aarch_rev16_p_1): Likewise.
5125 (aarch_rev16_p): Likewise.
5126 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5127 (aarch_rev16_shright_mask_imm_p): Likewise.
5128 (aarch_rev16_shleft_mask_imm_p): Likewise.
5130 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5132 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5133 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5135 (cortex_a53_extra_costs): Likewise.
5136 (cortex_a57_extra_costs): Likewise.
5137 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5138 (cortexa7_extra_costs): Likewise.
5139 (cortexa8_extra_costs): Likewise.
5140 (cortexa12_extra_costs): Likewise.
5141 (cortexa15_extra_costs): Likewise.
5142 (v7m_extra_costs): Likewise.
5143 (arm_new_rtx_costs): Handle BSWAP.
5145 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5147 * config/arm/arm.c (cortexa8_extra_costs): New table.
5148 (arm_cortex_a8_tune): New tuning struct.
5149 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5151 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5153 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5155 2014-04-23 Richard Biener <rguenther@suse.de>
5157 * Makefile.in (OBJS): Remove loop-unswitch.o.
5158 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5159 * passes.def (pass_rtl_unswitch): Likewise.
5160 * loop-init.c (gate_rtl_unswitch): Likewise.
5161 (rtl_unswitch): Likewise.
5162 (pass_data_rtl_unswitch): Likewise.
5163 (pass_rtl_unswitch): Likewise.
5164 (make_pass_rtl_unswitch): Likewise.
5165 * rtl.h (reversed_condition): Likewise.
5166 (compare_and_jump_seq): Likewise.
5167 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5169 * loop-unroll.c (compare_and_jump_seq): Likewise.
5171 2014-04-23 Richard Biener <rguenther@suse.de>
5173 PR tree-optimization/60903
5174 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5175 commented code block.
5176 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5177 loop flags to newly created BBs and edges.
5179 2014-04-23 Nick Clifton <nickc@redhat.com>
5181 * config/msp430/msp430.c (msp430_handle_option): Move function
5183 (msp430_option_override): Simplify mcu and mcpu option handling.
5184 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5185 support for -mhwmult command line option.
5186 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5187 -mhwmult command line option.
5188 (msp430_hwmult_enabled): Delete.
5189 (msp43o_output_labelref): Add support for -mhwmult command line option.
5190 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5191 (umulsidi3): Likewise.
5192 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5193 (mcpu, mlarge, msmall): Likewise.
5194 (mhwmult): New option.
5195 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5197 (msp430_is_f5_mcu): Remove prototype.
5198 (msp430_use_f5_series_hwmult): Add prototype.
5199 * config/msp430/msp430-opts.h: New file.
5200 * common/config/msp430: New directory.
5201 * common/config/msp430/msp430-common.c: New file.
5202 * config.gcc (msp430): Remove target_has_targetm_common.
5203 * doc/invoke.texi: Document -mhwmult command line option.
5205 2014-04-23 Nick Clifton <nickc@redhat.com>
5207 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5208 default-manifest.o if it can be found in the search path.
5209 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5211 2014-04-23 Terry Guo <terry.guo@arm.com>
5213 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5215 2014-04-23 Richard Biener <rguenther@suse.de>
5218 * tree-inline.c (declare_return_variable): Use mark_addressable.
5220 2014-04-23 Richard Biener <rguenther@suse.de>
5223 * loop-init.c (loop_optimizer_init): Make sure to apply
5224 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5226 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5229 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5231 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5232 if flag_sanitize_undefined_trap_on_error.
5233 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5234 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5235 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5236 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5237 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5238 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5239 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5240 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5241 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5242 * ubsan.c (ubsan_instrument_unreachable): Return
5243 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5244 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5245 if flag_sanitize_undefined_trap_on_error and
5246 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5247 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5248 instrument_bool_enum_load): Emit __builtin_trap () if
5249 flag_sanitize_undefined_trap_on_error and
5250 __builtin_handle_*_abort () if !flag_sanitize_recover.
5251 * doc/invoke.texi (-fsanitize-recover,
5252 -fsanitize-undefined-trap-on-error): Document.
5254 2014-04-22 Christian Bruel <christian.bruel@st.com>
5256 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5257 Force immediates to SImode.
5259 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5261 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5263 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5264 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5265 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5266 (nios2_fpu_insn): Add entry for round.
5267 (N2FPU_NO_ERRNO_P): Define.
5268 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5270 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5272 2014-04-22 Richard Henderson <rth@redhat.com>
5274 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5275 (add<GPI>3_compare0): Remove leading * from name.
5276 (add<GPI>3_carryin): Likewise.
5277 (sub<GPI>3_compare0): Likewise.
5278 (sub<GPI>3_carryin): Likewise.
5279 (<su_optab>mulditi3): New expander.
5280 (multi3): New expander.
5281 (madd<GPI>): Remove leading * from name.
5283 2014-04-22 Martin Jambor <mjambor@suse.cz>
5285 * cgraphclones.c (cgraph_function_versioning): Copy
5286 ipa_transforms_to_apply instead of asserting it is empty.
5288 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5291 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5292 on count_exp to get mode.
5294 2014-04-22 Andrew Pinski <apinski@cavium.com>
5296 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5297 Handle TLS for ILP32.
5298 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5299 (tlsie_small_<mode>): this and handle PTR.
5300 (tlsie_small_sidi): New pattern.
5301 (tlsle_small): Change to an expand to handle ILP32.
5302 (tlsle_small_<mode>): New pattern.
5303 (tlsdesc_small): Rename to ...
5304 (tlsdesc_small_<mode>): this and handle PTR.
5306 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5308 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5310 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5312 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5313 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5314 (aarch64_types_signed_poly_qualifiers): Likewise.
5315 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5316 (aarch64_types_poly_signed_qualifiers): Likewise.
5317 (TYPES_REINTERP_SS): Type macro added.
5318 (TYPES_REINTERP_SU): Likewise.
5319 (TYPES_REINTERP_SP): Likewise.
5320 (TYPES_REINTERP_US): Likewise.
5321 (TYPES_REINTERP_PS): Likewise.
5322 (aarch64_fold_builtin): New expression folding added.
5323 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5324 Declarations removed.
5325 (REINTERP_SS): Declarations added.
5326 (REINTERP_US): Likewise.
5327 (REINTERP_PS): Likewise.
5328 (REINTERP_SU): Likewise.
5329 (REINTERP_SP): Likewise.
5330 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5331 (vreinterpretq_p8_f64): Likewise.
5332 (vreinterpret_p16_f64): Likewise.
5333 (vreinterpretq_p16_f64): Likewise.
5334 (vreinterpret_f32_f64): Likewise.
5335 (vreinterpretq_f32_f64): Likewise.
5336 (vreinterpret_f64_f32): Likewise.
5337 (vreinterpret_f64_p8): Likewise.
5338 (vreinterpret_f64_p16): Likewise.
5339 (vreinterpret_f64_s8): Likewise.
5340 (vreinterpret_f64_s16): Likewise.
5341 (vreinterpret_f64_s32): Likewise.
5342 (vreinterpret_f64_s64): Likewise.
5343 (vreinterpret_f64_u8): Likewise.
5344 (vreinterpret_f64_u16): Likewise.
5345 (vreinterpret_f64_u32): Likewise.
5346 (vreinterpret_f64_u64): Likewise.
5347 (vreinterpretq_f64_f32): Likewise.
5348 (vreinterpretq_f64_p8): Likewise.
5349 (vreinterpretq_f64_p16): Likewise.
5350 (vreinterpretq_f64_s8): Likewise.
5351 (vreinterpretq_f64_s16): Likewise.
5352 (vreinterpretq_f64_s32): Likewise.
5353 (vreinterpretq_f64_s64): Likewise.
5354 (vreinterpretq_f64_u8): Likewise.
5355 (vreinterpretq_f64_u16): Likewise.
5356 (vreinterpretq_f64_u32): Likewise.
5357 (vreinterpretq_f64_u64): Likewise.
5358 (vreinterpret_s64_f64): Likewise.
5359 (vreinterpretq_s64_f64): Likewise.
5360 (vreinterpret_u64_f64): Likewise.
5361 (vreinterpretq_u64_f64): Likewise.
5362 (vreinterpret_s8_f64): Likewise.
5363 (vreinterpretq_s8_f64): Likewise.
5364 (vreinterpret_s16_f64): Likewise.
5365 (vreinterpretq_s16_f64): Likewise.
5366 (vreinterpret_s32_f64): Likewise.
5367 (vreinterpretq_s32_f64): Likewise.
5368 (vreinterpret_u8_f64): Likewise.
5369 (vreinterpretq_u8_f64): Likewise.
5370 (vreinterpret_u16_f64): Likewise.
5371 (vreinterpretq_u16_f64): Likewise.
5372 (vreinterpret_u32_f64): Likewise.
5373 (vreinterpretq_u32_f64): Likewise.
5375 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5377 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5378 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5379 (vreinterpret_p8_s8): Likewise.
5380 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5381 (vreinterpret_p8_s16): Likewise.
5382 (vreinterpret_p8_s32): Likewise.
5383 (vreinterpret_p8_s64): Likewise.
5384 (vreinterpret_p8_f32): Likewise.
5385 (vreinterpret_p8_u8): Likewise.
5386 (vreinterpret_p8_u16): Likewise.
5387 (vreinterpret_p8_u32): Likewise.
5388 (vreinterpret_p8_u64): Likewise.
5389 (vreinterpret_p8_p16): Likewise.
5390 (vreinterpretq_p8_s8): Likewise.
5391 (vreinterpretq_p8_s16): Likewise.
5392 (vreinterpretq_p8_s32): Likewise.
5393 (vreinterpretq_p8_s64): Likewise.
5394 (vreinterpretq_p8_f32): Likewise.
5395 (vreinterpretq_p8_u8): Likewise.
5396 (vreinterpretq_p8_u16): Likewise.
5397 (vreinterpretq_p8_u32): Likewise.
5398 (vreinterpretq_p8_u64): Likewise.
5399 (vreinterpretq_p8_p16): Likewise.
5400 (vreinterpret_p16_s8): Likewise.
5401 (vreinterpret_p16_s16): Likewise.
5402 (vreinterpret_p16_s32): Likewise.
5403 (vreinterpret_p16_s64): Likewise.
5404 (vreinterpret_p16_f32): Likewise.
5405 (vreinterpret_p16_u8): Likewise.
5406 (vreinterpret_p16_u16): Likewise.
5407 (vreinterpret_p16_u32): Likewise.
5408 (vreinterpret_p16_u64): Likewise.
5409 (vreinterpret_p16_p8): Likewise.
5410 (vreinterpretq_p16_s8): Likewise.
5411 (vreinterpretq_p16_s16): Likewise.
5412 (vreinterpretq_p16_s32): Likewise.
5413 (vreinterpretq_p16_s64): Likewise.
5414 (vreinterpretq_p16_f32): Likewise.
5415 (vreinterpretq_p16_u8): Likewise.
5416 (vreinterpretq_p16_u16): Likewise.
5417 (vreinterpretq_p16_u32): Likewise.
5418 (vreinterpretq_p16_u64): Likewise.
5419 (vreinterpretq_p16_p8): Likewise.
5420 (vreinterpret_f32_s8): Likewise.
5421 (vreinterpret_f32_s16): Likewise.
5422 (vreinterpret_f32_s32): Likewise.
5423 (vreinterpret_f32_s64): Likewise.
5424 (vreinterpret_f32_u8): Likewise.
5425 (vreinterpret_f32_u16): Likewise.
5426 (vreinterpret_f32_u32): Likewise.
5427 (vreinterpret_f32_u64): Likewise.
5428 (vreinterpret_f32_p8): Likewise.
5429 (vreinterpret_f32_p16): Likewise.
5430 (vreinterpretq_f32_s8): Likewise.
5431 (vreinterpretq_f32_s16): Likewise.
5432 (vreinterpretq_f32_s32): Likewise.
5433 (vreinterpretq_f32_s64): Likewise.
5434 (vreinterpretq_f32_u8): Likewise.
5435 (vreinterpretq_f32_u16): Likewise.
5436 (vreinterpretq_f32_u32): Likewise.
5437 (vreinterpretq_f32_u64): Likewise.
5438 (vreinterpretq_f32_p8): Likewise.
5439 (vreinterpretq_f32_p16): Likewise.
5440 (vreinterpret_s64_s8): Likewise.
5441 (vreinterpret_s64_s16): Likewise.
5442 (vreinterpret_s64_s32): Likewise.
5443 (vreinterpret_s64_f32): Likewise.
5444 (vreinterpret_s64_u8): Likewise.
5445 (vreinterpret_s64_u16): Likewise.
5446 (vreinterpret_s64_u32): Likewise.
5447 (vreinterpret_s64_u64): Likewise.
5448 (vreinterpret_s64_p8): Likewise.
5449 (vreinterpret_s64_p16): Likewise.
5450 (vreinterpretq_s64_s8): Likewise.
5451 (vreinterpretq_s64_s16): Likewise.
5452 (vreinterpretq_s64_s32): Likewise.
5453 (vreinterpretq_s64_f32): Likewise.
5454 (vreinterpretq_s64_u8): Likewise.
5455 (vreinterpretq_s64_u16): Likewise.
5456 (vreinterpretq_s64_u32): Likewise.
5457 (vreinterpretq_s64_u64): Likewise.
5458 (vreinterpretq_s64_p8): Likewise.
5459 (vreinterpretq_s64_p16): Likewise.
5460 (vreinterpret_u64_s8): Likewise.
5461 (vreinterpret_u64_s16): Likewise.
5462 (vreinterpret_u64_s32): Likewise.
5463 (vreinterpret_u64_s64): Likewise.
5464 (vreinterpret_u64_f32): Likewise.
5465 (vreinterpret_u64_u8): Likewise.
5466 (vreinterpret_u64_u16): Likewise.
5467 (vreinterpret_u64_u32): Likewise.
5468 (vreinterpret_u64_p8): Likewise.
5469 (vreinterpret_u64_p16): Likewise.
5470 (vreinterpretq_u64_s8): Likewise.
5471 (vreinterpretq_u64_s16): Likewise.
5472 (vreinterpretq_u64_s32): Likewise.
5473 (vreinterpretq_u64_s64): Likewise.
5474 (vreinterpretq_u64_f32): Likewise.
5475 (vreinterpretq_u64_u8): Likewise.
5476 (vreinterpretq_u64_u16): Likewise.
5477 (vreinterpretq_u64_u32): Likewise.
5478 (vreinterpretq_u64_p8): Likewise.
5479 (vreinterpretq_u64_p16): Likewise.
5480 (vreinterpret_s8_s16): Likewise.
5481 (vreinterpret_s8_s32): Likewise.
5482 (vreinterpret_s8_s64): Likewise.
5483 (vreinterpret_s8_f32): Likewise.
5484 (vreinterpret_s8_u8): Likewise.
5485 (vreinterpret_s8_u16): Likewise.
5486 (vreinterpret_s8_u32): Likewise.
5487 (vreinterpret_s8_u64): Likewise.
5488 (vreinterpret_s8_p8): Likewise.
5489 (vreinterpret_s8_p16): Likewise.
5490 (vreinterpretq_s8_s16): Likewise.
5491 (vreinterpretq_s8_s32): Likewise.
5492 (vreinterpretq_s8_s64): Likewise.
5493 (vreinterpretq_s8_f32): Likewise.
5494 (vreinterpretq_s8_u8): Likewise.
5495 (vreinterpretq_s8_u16): Likewise.
5496 (vreinterpretq_s8_u32): Likewise.
5497 (vreinterpretq_s8_u64): Likewise.
5498 (vreinterpretq_s8_p8): Likewise.
5499 (vreinterpretq_s8_p16): Likewise.
5500 (vreinterpret_s16_s8): Likewise.
5501 (vreinterpret_s16_s32): Likewise.
5502 (vreinterpret_s16_s64): Likewise.
5503 (vreinterpret_s16_f32): Likewise.
5504 (vreinterpret_s16_u8): Likewise.
5505 (vreinterpret_s16_u16): Likewise.
5506 (vreinterpret_s16_u32): Likewise.
5507 (vreinterpret_s16_u64): Likewise.
5508 (vreinterpret_s16_p8): Likewise.
5509 (vreinterpret_s16_p16): Likewise.
5510 (vreinterpretq_s16_s8): Likewise.
5511 (vreinterpretq_s16_s32): Likewise.
5512 (vreinterpretq_s16_s64): Likewise.
5513 (vreinterpretq_s16_f32): Likewise.
5514 (vreinterpretq_s16_u8): Likewise.
5515 (vreinterpretq_s16_u16): Likewise.
5516 (vreinterpretq_s16_u32): Likewise.
5517 (vreinterpretq_s16_u64): Likewise.
5518 (vreinterpretq_s16_p8): Likewise.
5519 (vreinterpretq_s16_p16): Likewise.
5520 (vreinterpret_s32_s8): Likewise.
5521 (vreinterpret_s32_s16): Likewise.
5522 (vreinterpret_s32_s64): Likewise.
5523 (vreinterpret_s32_f32): Likewise.
5524 (vreinterpret_s32_u8): Likewise.
5525 (vreinterpret_s32_u16): Likewise.
5526 (vreinterpret_s32_u32): Likewise.
5527 (vreinterpret_s32_u64): Likewise.
5528 (vreinterpret_s32_p8): Likewise.
5529 (vreinterpret_s32_p16): Likewise.
5530 (vreinterpretq_s32_s8): Likewise.
5531 (vreinterpretq_s32_s16): Likewise.
5532 (vreinterpretq_s32_s64): Likewise.
5533 (vreinterpretq_s32_f32): Likewise.
5534 (vreinterpretq_s32_u8): Likewise.
5535 (vreinterpretq_s32_u16): Likewise.
5536 (vreinterpretq_s32_u32): Likewise.
5537 (vreinterpretq_s32_u64): Likewise.
5538 (vreinterpretq_s32_p8): Likewise.
5539 (vreinterpretq_s32_p16): Likewise.
5540 (vreinterpret_u8_s8): Likewise.
5541 (vreinterpret_u8_s16): Likewise.
5542 (vreinterpret_u8_s32): Likewise.
5543 (vreinterpret_u8_s64): Likewise.
5544 (vreinterpret_u8_f32): Likewise.
5545 (vreinterpret_u8_u16): Likewise.
5546 (vreinterpret_u8_u32): Likewise.
5547 (vreinterpret_u8_u64): Likewise.
5548 (vreinterpret_u8_p8): Likewise.
5549 (vreinterpret_u8_p16): Likewise.
5550 (vreinterpretq_u8_s8): Likewise.
5551 (vreinterpretq_u8_s16): Likewise.
5552 (vreinterpretq_u8_s32): Likewise.
5553 (vreinterpretq_u8_s64): Likewise.
5554 (vreinterpretq_u8_f32): Likewise.
5555 (vreinterpretq_u8_u16): Likewise.
5556 (vreinterpretq_u8_u32): Likewise.
5557 (vreinterpretq_u8_u64): Likewise.
5558 (vreinterpretq_u8_p8): Likewise.
5559 (vreinterpretq_u8_p16): Likewise.
5560 (vreinterpret_u16_s8): Likewise.
5561 (vreinterpret_u16_s16): Likewise.
5562 (vreinterpret_u16_s32): Likewise.
5563 (vreinterpret_u16_s64): Likewise.
5564 (vreinterpret_u16_f32): Likewise.
5565 (vreinterpret_u16_u8): Likewise.
5566 (vreinterpret_u16_u32): Likewise.
5567 (vreinterpret_u16_u64): Likewise.
5568 (vreinterpret_u16_p8): Likewise.
5569 (vreinterpret_u16_p16): Likewise.
5570 (vreinterpretq_u16_s8): Likewise.
5571 (vreinterpretq_u16_s16): Likewise.
5572 (vreinterpretq_u16_s32): Likewise.
5573 (vreinterpretq_u16_s64): Likewise.
5574 (vreinterpretq_u16_f32): Likewise.
5575 (vreinterpretq_u16_u8): Likewise.
5576 (vreinterpretq_u16_u32): Likewise.
5577 (vreinterpretq_u16_u64): Likewise.
5578 (vreinterpretq_u16_p8): Likewise.
5579 (vreinterpretq_u16_p16): Likewise.
5580 (vreinterpret_u32_s8): Likewise.
5581 (vreinterpret_u32_s16): Likewise.
5582 (vreinterpret_u32_s32): Likewise.
5583 (vreinterpret_u32_s64): Likewise.
5584 (vreinterpret_u32_f32): Likewise.
5585 (vreinterpret_u32_u8): Likewise.
5586 (vreinterpret_u32_u16): Likewise.
5587 (vreinterpret_u32_u64): Likewise.
5588 (vreinterpret_u32_p8): Likewise.
5589 (vreinterpret_u32_p16): Likewise.
5590 (vreinterpretq_u32_s8): Likewise.
5591 (vreinterpretq_u32_s16): Likewise.
5592 (vreinterpretq_u32_s32): Likewise.
5593 (vreinterpretq_u32_s64): Likewise.
5594 (vreinterpretq_u32_f32): Likewise.
5595 (vreinterpretq_u32_u8): Likewise.
5596 (vreinterpretq_u32_u16): Likewise.
5597 (vreinterpretq_u32_u64): Likewise.
5598 (vreinterpretq_u32_p8): Likewise.
5599 (vreinterpretq_u32_p16): Likewise.
5601 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5603 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5605 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5607 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5608 (vqnegd_s64): Likewise.
5609 (vqabs_s64): Likewise.
5610 (vqabsd_s64): Likewise.
5612 2014-04-22 Richard Henderson <rth@redhat.com>
5614 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5615 computation to the top of the loop.
5617 2014-04-22 Renlin <renlin.li@arm.com>
5618 Jiong Wang <jiong.wang@arm.com>
5620 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5621 * config/aarch64/aarch64.c (aarch64_layout_frame)
5622 (aarch64_initial_elimination_offset): Likewise.
5624 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5626 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5629 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5631 * machmode.h (bitwise_mode_for_mode): Declare.
5632 * stor-layout.h (bitwise_type_for_mode): Likewise.
5633 * stor-layout.c (bitwise_mode_for_mode): New function.
5634 (bitwise_type_for_mode): Likewise.
5635 * builtins.c (fold_builtin_memory_op): Use it instead of
5636 int_mode_for_mode and build_nonstandard_integer_type.
5638 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5640 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5641 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5642 (*-*-solaris2*): Simplify.
5643 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5644 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5645 *-*-solaris2.9* handling.
5647 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5649 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5650 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5652 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5653 * configure: Regenerate.
5655 * config/i386/sol2-9.h: Remove.
5657 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5658 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5659 Remove Solaris 9 references.
5661 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5663 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5664 (floatuns<GPI:mode><GPF:mode>2): Remove.
5665 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5666 and floatuns conversions.
5667 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5668 and floatuns conversions.
5669 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5670 (w1,w2): New mode attributes for inequal width conversions.
5672 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5674 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5675 the output asm format.
5677 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5679 * config/aarch64/aarch64-simd.md
5680 (aarch64_cm<optab>di): Always split.
5681 (*aarch64_cm<optab>di): New.
5682 (aarch64_cmtstdi): Always split.
5683 (*aarch64_cmtstdi): New.
5685 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5687 PR tree-optimization/60823
5688 * omp-low.c (ipa_simd_modify_function_body): Go through
5689 all SSA_NAMEs and for those refering to vector arguments
5690 which are going to be replaced adjust SSA_NAME_VAR and,
5691 if it is a default definition, change it into a non-default
5692 definition assigned at the beginning of function from new_decl.
5693 (ipa_simd_modify_stmt_ops): Rewritten.
5694 * tree-dfa.c (set_ssa_default_def): When removing default def,
5695 check for NULL loc instead of NULL *loc.
5697 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5699 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5700 restrictions on core registers for DImode values in Thumb2.
5702 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5704 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5705 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5707 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5709 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5710 (*iordi_notzesidi_di): Likewise.
5711 (*iordi_notsesidi_di): Likewise.
5713 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5715 * config/arm/arm-protos.h (tune_params): New struct members.
5716 * config/arm/arm.c: Initialise tune_params per processor.
5717 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5718 for speed, based on new tune_params.
5720 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5722 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5723 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5724 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5725 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5726 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5727 (vrnda_f64): Likewise.
5728 (vrndi_f64): Likewise.
5729 (vrndm_f64): Likewise.
5730 (vrndn_f64): Likewise.
5731 (vrndp_f64): Likewise.
5732 (vrndx_f64): Likewise.
5734 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5736 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5737 GET_MODE_SIZE argument is enum machine_mode.
5739 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5742 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5743 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5745 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5748 * asan.c (asan_emit_stack_protection): Force the base to align to
5749 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5750 appropriate bits if STRICT_ALIGNMENT.
5751 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5753 (expand_used_vars): Leave a space in the stack frame for alignment
5754 if STRICT_ALIGNMENT.
5756 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5758 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5760 (gimple_store_p): Likewise.
5761 (gimple_assign_load_p): Likewise.
5762 (gimple_assign_cast_p): Likewise.
5763 (gimple_clobber_p): Likewise.
5765 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5766 rather than a gimple.
5767 (gimple_assign_cast_p): Likewise.
5769 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5772 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5773 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5775 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5776 more debug information for E500 if -mdebug=reg.
5778 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5781 * config/i386/i386.c (ix86_expand_builtin)
5782 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5783 register for target RTX.
5784 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5786 2014-04-18 Cong Hou <congh@google.com>
5788 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5789 the widen-mult pattern by handling two operands with different sizes,
5790 and operands whose size is smaller than half of the result type.
5792 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5794 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5795 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5796 (do_estimate_edge_time): Compute it.
5797 * ipa-inline.c (want_inline_small_function_p): Bypass
5798 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5800 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5802 * ipa-inline.c (spec_rem): New static variable.
5803 (dump_overall_stats): New function.
5804 (dump_inline_stats): New function.
5806 2014-04-18 Richard Henderson <rth@redhat.com>
5808 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5809 to GET_MODE_SIZE, not a reg_class_t.
5811 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5813 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5814 (vsx_xxmrglw_<mode>): Likewise.
5816 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5819 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5820 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5821 (rs6000_init_hard_regno_mode_ok): Likewise.
5823 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
5825 * ipa-inline.c (inline_small_functions): Account only non-cold
5827 * doc/invoke.texi (inline-unit-growth): Update documentation.
5829 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
5831 * config/rs6000/rs6000.md (addti3, subti3): New.
5833 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
5836 * config/i386/i386.c (ix86_expand_clear): Remove outdated
5837 comment. Check optimize_insn_for_size_p instead of
5838 optimize_insn_for_speed_p.
5840 2014-04-17 Martin Jambor <mjambor@suse.cz>
5842 * gimple-iterator.c (gsi_start_edge): New function.
5843 * gimple-iterator.h (gsi_start_edge): Declare.
5844 * tree-sra.c (single_non_eh_succ): New function.
5845 (disqualify_ops_if_throwing_stmt): Renamed to
5846 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
5847 having one non-EH successor BB.
5848 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
5849 generate loads into replacements.
5850 (sra_modify_assign): Likewise and and also use the simple path for
5852 (sra_modify_function_body): Commit statements on edges.
5854 2014-04-17 Richard Biener <rguenther@suse.de>
5857 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
5858 comparison results and add clarifying comment.
5860 2014-04-17 Jakub Jelinek <jakub@redhat.com>
5862 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
5863 (blank_mode): Initialize it.
5864 (emit_mode_size_inline, emit_mode_nunits_inline,
5865 emit_mode_inner_inline): New functions.
5866 (emit_insn_modes_h): Call them and surround their output with
5867 #if GCC_VERSION >= 4001 ... #endif.
5868 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
5869 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
5870 mode_* arrays if the argument is __builtin_constant_p.
5871 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
5872 is enum machine_mode.
5874 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5876 * passes.c (opt_pass::execute): Adjust.
5877 (pass_manager::execute_pass_mode_switching): Likewise.
5878 (early_local_passes::execute): Likewise.
5879 (execute_one_pass): Pass cfun to the pass's execute method.
5880 * tree-pass.h (opt_pass::execute): Add function * argument.
5881 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5882 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5883 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5884 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5885 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5886 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
5887 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
5888 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5889 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5890 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
5891 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
5892 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
5893 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
5894 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
5895 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5896 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5897 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5898 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
5899 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
5900 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5901 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5902 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5903 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5904 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5905 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5906 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5907 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
5908 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
5909 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
5912 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5914 * passes.c (opt_pass::gate): Take function * argument.
5915 (gate_all_early_local_passes): Merge into
5916 (early_local_passes::gate): this.
5917 (gate_all_early_optimizations): Merge into
5918 (all_early_optimizations::gate): this.
5919 (gate_all_optimizations): Mege into
5920 (all_optimizations::gate): this.
5921 (gate_all_optimizations_g): Merge into
5922 (all_optimizations_g::gate): this.
5923 (gate_rest_of_compilation): Mege into
5924 (rest_of_compilation::gate): this.
5925 (gate_postreload): Merge into
5926 (postreload::gate): this.
5927 (dump_one_pass): Pass cfun to the pass's gate method.
5928 (execute_ipa_summary_passes): Likewise.
5929 (execute_one_pass): Likewise.
5930 (ipa_write_summaries_2): Likewise.
5931 (ipa_write_optimization_summaries_1): Likewise.
5932 (ipa_read_summaries_1): Likewise.
5933 (ipa_read_optimization_summaries_1): Likewise.
5934 (execute_ipa_stmt_fixups): Likewise.
5935 * tree-pass.h (opt_pass::gate): Add function * argument.
5936 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
5937 combine-stack-adj.c, combine.c, compare-elim.c,
5938 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5939 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
5940 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
5941 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
5942 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
5943 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
5944 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5945 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
5946 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
5947 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
5948 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
5949 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
5950 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
5951 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5952 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
5953 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
5954 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
5955 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
5956 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
5957 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
5958 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
5959 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
5960 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
5961 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
5962 var-tracking.c, vtable-verify.c, web.c: Adjust.
5964 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5966 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
5967 * configure: Regenerate.
5969 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
5971 * passes.c (dump_one_pass): don't check pass->has_gate.
5972 (execute_ipa_summary_passes): Likewise.
5973 (execute_one_pass): Likewise.
5974 (ipa_write_summaries_2): Likewise.
5975 (ipa_write_optimization_summaries_1): Likewise.
5976 (ipa_read_optimization_summaries_1): Likewise.
5977 (execute_ipa_stmt_fixups): Likewise.
5978 * tree-pass.h (pass_data::has_gate): Remove.
5979 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
5980 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
5981 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
5982 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
5983 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
5984 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
5985 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
5986 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
5987 gimple-low.c, gimple-ssa-isolate-paths.c,
5988 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
5989 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
5990 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
5991 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
5992 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
5993 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
5994 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
5995 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5996 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
5997 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
5998 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
5999 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6000 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6001 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6002 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6003 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6004 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6005 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6006 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6007 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6008 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6009 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6012 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6014 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6016 * passes.c (pass_manager::register_dump_files_1): Merge into
6017 (pass_manager::register_dump_files): this, and remove its handling of
6018 properties since the pass always has the properties anyway.
6019 (pass_manager::pass_manager): Adjust.
6021 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6023 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6024 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6025 dealing with properties.
6026 (pass_manager::register_dump_files): Adjust.
6028 2014-03-20 Mark Wielaard <mjw@redhat.com>
6030 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6031 then represent the bound as normal constant value.
6033 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6036 Forward port from 4.8 branch
6037 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6039 * config/i386/bmiintrin.h (_blsi_u32): New.
6043 (_blsmsk_u32): Ditto.
6044 (_blsmsk_u64): Ditto.
6045 (_tzcnt_u32): Ditto.
6046 (_tzcnt_u64): Ditto.
6048 2014-04-17 Kito Cheng <kito@0xlab.org>
6050 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6052 2014-04-17 Richard Biener <rguenther@suse.de>
6055 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6056 boolean results for comparisons.
6058 2014-04-17 Richard Biener <rguenther@suse.de>
6060 PR tree-optimization/60836
6061 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6062 initial PHI args to be gimple values.
6064 2014-04-17 Richard Biener <rguenther@suse.de>
6066 PR tree-optimization/60841
6067 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6068 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6069 of stmts to SLP build.
6070 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6071 (vect_analyze_slp): Likewise.
6072 (vect_analyze_slp_instance): Likewise.
6073 (vect_build_slp_tree): Limit overall SLP tree growth.
6074 * tree-vectorizer.h (vect_analyze_data_refs,
6075 vect_analyze_slp): Adjust prototypes.
6077 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6079 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6082 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6084 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6085 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6086 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6087 for TARGET_SLOW_PSHUFB
6089 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6091 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6092 * config/i386/i386.c (intel_cost): Ditto.
6094 2014-04-17 Joey Ye <joey.ye@arm.com>
6096 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6098 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6100 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6103 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6105 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6106 (type_all_derivations_known_p): New predicate.
6107 (type_all_ctors_visible_p): New predicate.
6108 (type_possibly_instantiated_p): New predicate.
6109 (get_odr_type): Compute all_derivations_known.
6110 (dump_odr_type): Dump the flag.
6111 (maybe_record_type): Cleanup.
6112 (record_target_from_binfo): Add bases_to_consider array;
6113 record bases for types w/o instances and skip CXX destructor.
6114 (possible_polymorphic_call_targets_1): Add bases_to_consider
6115 and consider_construction parameters; check if type may have instance.
6116 (get_polymorphic_call_info): Set maybe_in_construction to true
6117 when we know nothing.
6118 (record_targets_from_bases): Skip CXX destructors; they are
6119 never called for types in construction.
6120 (possible_polymorphic_call_targets): Do not record target when
6121 type may not have instance.
6123 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6126 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6127 external aliases alive, too.
6129 2014-04-16 Andrew Pinski <apinski@cavium.com>
6131 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6134 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6136 * final.c (compute_alignments): Do not apply loop alignment to a block
6137 falling through to the exit.
6139 2014-04-16 Catherine Moore <clm@codesourcery.com>
6141 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6142 Adjust constraints for microMIPS store patterns.
6144 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6146 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6148 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6150 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6151 (append_use): Run at -O0.
6152 (append_vdef): Likewise.
6153 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6154 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6156 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6158 PR tree-optimization/60844
6159 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6160 (propagate_op_to_single_use, remove_visited_stmt_chain,
6161 linearize_expr, repropagate_negates, reassociate_bb): Use it
6162 instead of gsi_remove.
6164 2014-04-16 Martin Jambor <mjambor@suse.cz>
6166 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6167 ipa_transforms_to_apply.
6168 (cgraph_function_versioning): Assert that old_node has empty
6169 ipa_transforms_to_apply.
6170 * trans-mem.c (ipa_tm_create_version): Likewise.
6171 * tree-inline.c (tree_function_versioning): Do not duplicate
6172 ipa_transforms_to_apply.
6174 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6177 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6179 Pass necessary as flags on 64-bit Solaris/x86.
6180 Use lowercase relocs for x86_64-*-*.
6181 * configure: Regenerate.
6183 2014-04-15 Jan Hubicka <jh@suse.cz>
6185 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6186 (maybe_record_node, likely_target_p): Use it.
6188 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6191 Revert following patch
6193 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6196 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6197 software floating point or no floating point registers, do not
6198 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6199 in GPRs that occurs after we tested for GPRs that would never be
6202 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6203 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6204 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6205 specifically allow DDmode, since that does not use the SPE SIMD
6208 2014-03-21 Mark Wielaard <mjw@redhat.com>
6210 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6211 as unsigned or int depending on type and value used.
6213 2014-04-15 Richard Biener <rguenther@suse.de>
6215 PR rtl-optimization/56965
6216 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6217 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6219 * alias.c (true_dependence_1): Do not call
6220 nonoverlapping_component_refs_p.
6221 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6222 nonoverlapping_component_refs_p.
6223 (indirect_refs_may_alias_p): Likewise.
6225 2014-04-15 Teresa Johnson <tejohnson@google.com>
6227 * cfg.c (dump_bb_info): Fix flags check.
6228 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6230 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6232 PR rtl-optimization/60663
6233 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6236 2014-04-15 Richard Biener <rguenther@suse.de>
6238 * lto-streamer.h (LTO_major_version): Bump to 4.
6240 2014-04-15 Richard Biener <rguenther@suse.de>
6242 * common.opt (lto_partition_model): New enum.
6243 (flto-partition=): Merge separate options with a single with argument,
6244 add -flto-partition=one support.
6245 * flag-types.h (enum lto_partition_model): Declare.
6246 * opts.c (finish_options): Remove duplicate -flto-partition=
6248 * lto-wrapper.c (run_gcc): Adjust.
6250 2014-04-15 Richard Biener <rguenther@suse.de>
6252 * alias.c (ncr_compar): New function.
6253 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6255 2014-04-15 Richard Biener <rguenther@suse.de>
6257 * alias.c (record_component_aliases): Do not walk BINFOs.
6259 2014-04-15 Richard Biener <rguenther@suse.de>
6261 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6262 Add struct function argument and adjust.
6263 (find_func_aliases_for_call): Likewise.
6264 (find_func_aliases): Likewise.
6265 (find_func_clobbers): Likewise.
6266 (intra_create_variable_infos): Likewise.
6267 (compute_points_to_sets): Likewise.
6268 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6270 2014-04-15 Richard Biener <rguenther@suse.de>
6272 * tree.c (iterative_hash_expr): Use enum tree_code_class
6273 to store TREE_CODE_CLASS.
6274 (tree_block): Likewise.
6275 (tree_set_block): Likewise.
6276 * tree.h (fold_build_pointer_plus_loc): Use
6277 convert_to_ptrofftype_loc.
6279 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6282 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6285 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6287 * cfgloop.h (struct loop): Move force_vectorize down.
6288 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6289 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6290 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6291 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6292 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6293 * tree-core.h (enum annot_expr_kind): Add new kind values.
6294 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6295 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6297 * tree.def (ANNOTATE_EXPR): Tweak comment.
6299 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6301 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6304 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6306 * tree.h (TYPE_IDENTIFIER): Declare.
6307 * tree.c (subrange_type_for_debug_p): Use it.
6308 * godump.c (go_format_type): Likewise.
6309 * dwarf2out.c (is_cxx_auto, modified_type_die,
6310 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6311 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6313 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6316 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6318 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6320 * config/i386/i386.c (examine_argument): Return bool. Return true if
6321 parameter should be passed in memory.
6322 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6323 (construct_container): Update calls to examine_argument.
6324 (function_arg_advance_64): Ditto.
6325 (return_in_memory_32): Merge with ix86_return_in_memory.
6326 (return_in_memory_64): Ditto.
6327 (return_in_memory_ms_64): Ditto.
6329 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6331 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6332 * coverage.c (coverage_compute_profile_id): Handle externally visible
6335 2014-04-14 Martin Jambor <mjambor@suse.cz>
6337 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6338 DECL_DISREGARD_INLINE_LIMITS functions.
6340 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6343 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6345 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6348 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6349 optimize_insn_for_speed_p instead of
6350 optimize_function_for_speed_p.
6352 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6354 * doc/invoke.texi (free): Document AArch64.
6356 2014-04-14 Richard Biener <rguenther@suse.de>
6358 PR tree-optimization/60042
6359 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6360 (insert_into_preds_of_block): Do not prevent PHI insertion
6361 for REFERENCE exprs here ...
6362 (eliminate_dom_walker::before_dom_children): ... but prevent
6363 their use here under similar conditions when applied to the
6364 IL after PRE optimizations.
6366 2014-04-14 Richard Biener <rguenther@suse.de>
6368 * passes.def: Move early points-to after early SRA.
6370 2014-04-14 Richard Biener <rguenther@suse.de>
6372 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6373 check for which sign-changes we allow when forwarding
6374 a converted value into a switch.
6376 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6378 * stor-layout.c (place_field): Finalize non-constant offset for the
6381 2014-04-14 Richard Biener <rguenther@suse.de>
6383 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6385 (expand_switch_using_bit_tests_p): Likewise.
6386 (process_switch): Compute and pass on speed_p based on the
6388 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6389 optimize_bb_for_speed_p.
6391 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6393 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6394 * function.h (struct function): Rename has_force_vect_loops into
6395 has_force_vectorize_loops.
6396 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6397 (input_struct_function_base): Likewise.
6398 * lto-streamer-out.c (output_cfg): Likewise.
6399 (output_struct_function_base): Likewise.
6400 * omp-low.c (expand_omp_simd): Likewise.
6401 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6402 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6403 (version_loop_for_if_conversion): Likewise.
6404 (tree_if_conversion): Likewise.
6405 (main_tree_if_conversion): Likewise.
6406 (gate_tree_if_conversion): Likewise.
6407 * tree-inline.c (copy_loops): Likewise.
6408 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6409 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6410 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6411 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6412 * tree-vectorizer.c (vectorize_loops): Likewise.
6413 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6415 2014-04-14 Richard Biener <rguenther@suse.de>
6418 * lto-streamer-out.c (wrap_refs): New function.
6419 (lto_output): Wrap symbol references in global initializes in
6420 type-preserving MEM_REFs.
6422 2014-04-14 Christian Bruel <christian.bruel@st.com>
6424 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6426 2014-04-14 Christian Bruel <christian.bruel@st.com>
6428 * config/sh/sh.md (setmemqi): New expand pattern.
6429 * config/sh/sh.h (CLEAR_RATIO): Define.
6430 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6431 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6433 2014-04-14 Richard Biener <rguenther@suse.de>
6436 * fold-const.c (negate_expr_p): Don't negate directional rounding
6438 (fold_negate_expr): Likewise.
6440 2014-04-14 Richard Biener <rguenther@suse.de>
6442 PR tree-optimization/59817
6443 PR tree-optimization/60453
6444 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6445 recursion to catch all CHRECs in the scalar evolution and restrict
6446 the predicate for the remains appropriately.
6448 2014-04-12 Catherine Moore <clm@codesourcery.com>
6450 * config/mips/constraints.md: Add new register constraint "kb".
6451 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6452 (*movhi_internal): Likewise.
6453 (*movqi_internal): Likewise.
6454 * config/mips/mips.h (M16_STORE_REGS): New register class.
6455 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6456 (REG_CLASS_CONTENTS): Likewise.
6457 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6459 2014-04-11 Tobias Burnus <burnus@net-b.de>
6462 * doc/invoke.texi (-Wformat-signedness): Document it.
6463 (Wformat=2): Mention that this enables -Wformat-signedness.
6465 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6467 * common/config/epiphany/epiphany-common.c
6468 (epiphany_option_optimization_table): Enable section anchors by
6469 default at -O1 or higher.
6470 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6471 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6472 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6473 carries no extra cost.
6474 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6475 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6476 * config/epiphany/predicates.md (memclob_operand): New predicate.
6477 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6478 Use memclob_operand predicate and X constraint for operand 3.
6480 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6482 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6483 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6486 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6488 PR rtl-optimization/60651
6489 * mode-switching.c (optimize_mode_switching): Make sure to emit
6490 sets of a lower numbered entity before sets of a higher numbered
6491 entity to a mode of the same or lower priority.
6492 When creating a seginfo for a basic block that starts with a code
6493 label, move the insertion point past the code label.
6494 (new_seginfo): Document and enforce requirement that
6495 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6496 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6497 * doc/tm.texi: Regenerate.
6499 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6502 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6504 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6506 * BASE-VER: Set to 4.10.0.
6508 2014-04-11 Tobias Burnus <burnus@net-b.de>
6511 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6512 * doc/gcc.texi (Service): Update description in the @menu
6513 * doc/invoke.texi (Option Summary): Remove misplaced and
6516 2014-04-11 Steve Ellcey <sellcey@mips.com>
6517 Jakub Jelinek <jakub@redhat.com>
6520 * expr.c (convert_move): Use emit_store_flag_force instead of
6521 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6524 2014-04-11 Richard Biener <rguenther@suse.de>
6527 * varasm.c (assemble_alias): Avoid endless error reporting
6528 recursion by setting TREE_ASM_WRITTEN.
6530 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6532 * config/s390/s390.md: Add a splitter for NOT rtx.
6534 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6536 PR rtl-optimization/60663
6537 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6539 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6540 Jakub Jelinek <jakub@redhat.com>
6543 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6544 flag from decl_node to node.
6546 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6549 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6550 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6551 ameliorating the cases where it can be.
6553 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6556 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6558 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6559 (loadsync_<mode>): Change mode.
6560 (load_quadpti, store_quadpti): New.
6561 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6562 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6563 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6565 2014-04-09 Cong Hou <congh@google.com>
6568 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6571 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6573 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6574 instead of vnor to exploit possible fusion opportunity in the
6576 (altivec_expand_vec_perm_const_le): Likewise.
6578 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6580 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6581 (loadsync_<mode>): Change mode.
6582 (load_quadpti, store_quadpti): New.
6583 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6584 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6586 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6589 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6590 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6591 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6593 2014-04-08 Richard Biener <rguenther@suse.de>
6596 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6597 a 64bit widest int print double-int similar to on HWI64 hosts.
6599 2014-04-08 Richard Biener <rguenther@suse.de>
6601 PR tree-optimization/60785
6602 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6603 default defs properly.
6605 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6607 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6608 (Weffc++): Likewise.
6610 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6612 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6613 set completep to false rather than true.
6615 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6618 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6619 ARM_TARGET2_DWARF_FORMAT.
6621 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6624 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6625 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6628 2014-04-07 Richard Biener <rguenther@suse.de>
6630 PR tree-optimization/60766
6631 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6632 (may_eliminate_iv): Convert cand_value_at result to desired type.
6634 2014-04-07 Jason Merrill <jason@redhat.com>
6637 * common.opt (-fno-gnu-unique): Add.
6638 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6640 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6642 * haifa-sched.c: Fix outdated function reference and minor
6643 grammar errors in introductory comment.
6645 2014-04-07 Richard Biener <rguenther@suse.de>
6648 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6650 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6652 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6655 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6656 size accounting for thunks.
6657 (pa_asm_output_mi_thunk): Use final_start_function() and
6658 final_end_function() to output function start and end directives.
6660 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6662 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6663 device specific ISA/ feature information. Remove short_sp and
6664 errata_skip ds. Add avr_device_specific_features enum to have device
6666 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6667 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6668 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6669 updated device specific info.
6670 * config/avr/avr-mcus.def: Merge device specific details to
6671 dev_attribute field.
6672 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6674 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6675 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6676 assembler if RMW isa supported by current device.
6677 * config/avr/genmultilib.awk: Update as device info structure changed.
6678 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6680 2014-04-04 Cong Hou <congh@google.com>
6682 PR tree-optimization/60656
6683 * tree-vect-stmts.c (supportable_widening_operation):
6684 Fix a bug that elements in a vector with vect_used_by_reduction
6685 property are incorrectly reordered when the operation on it is not
6686 consistant with the one in reduction operation.
6688 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6690 PR rtl-optimization/60155
6691 * gcse.c (record_set_data): New function.
6692 (single_set_gcse): New function.
6693 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6694 (hoist_code): Likewise.
6695 (get_pressure_class_and_nregs): Likewise.
6697 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6699 * explow.c (probe_stack_range): Emit a final optimization blockage.
6701 2014-04-04 Anthony Green <green@moxielogic.com>
6703 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6706 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6709 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6710 flags are set only during streaming.
6711 * ipa.c (process_references, walk_polymorphic_call_targets,
6712 symtab_remove_unreachable_nodes): Drop bodies of always inline
6713 after early inlining.
6714 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6716 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6717 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6720 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6723 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6726 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6728 (cortex_a53_fdivd): Likewise.
6730 2014-04-04 Martin Jambor <mjambor@suse.cz>
6733 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6735 * cgraph.c (clone_of_p): Also return true if thunks match.
6736 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6737 cgraph_function_or_thunk_node and an obsolete comment.
6738 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6740 (build_function_decl_skip_args): Likewise.
6741 (set_new_clone_decl_and_node_flags): New function.
6742 (duplicate_thunk_for_node): Likewise.
6743 (redirect_edge_duplicating_thunks): Likewise.
6744 (cgraph_clone_node): New parameter args_to_skip, pass it to
6745 redirect_edge_duplicating_thunks which is called instead of
6746 cgraph_redirect_edge_callee.
6747 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6748 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6750 2014-04-04 Jeff Law <law@redhat.com>
6753 * config/arm/predicates.md (const_int_I_operand): New predicate.
6754 (const_int_M_operand): Similarly.
6755 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6757 (insv_t2, extv_reg, extzv_t2): Likewise.
6758 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6759 (pop_multiple_with_writeback_and_return): Likewise.
6760 (vfp_pop_multiple_with_writeback): Likewise
6762 2014-04-04 Richard Biener <rguenther@suse.de>
6765 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6766 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6768 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6769 * gimplify.c (gimple_add_tmp_var_fn): New function.
6770 * gimple-expr.h (create_tmp_reg_fn): Declare.
6771 * gimple-expr.c (create_tmp_reg_fn): New function.
6772 * gimple-low.c (record_vars_into): Don't change cfun.
6773 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6774 code generation without cfun.
6776 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6779 * Makefile.in (install-driver): Fix shell scripting.
6781 2014-04-03 Cong Hou <congh@google.com>
6783 PR tree-optimization/60505
6784 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6785 threshold of number of iterations below which no vectorization
6787 * tree-vect-loop.c (new_loop_vec_info):
6788 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6789 * tree-vect-loop.c (vect_analyze_loop_operations):
6790 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6791 * tree-vect-loop.c (vect_transform_loop):
6792 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6793 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6794 of iterations of the loop and see if we should build the epilogue.
6796 2014-04-03 Richard Biener <rguenther@suse.de>
6798 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6799 (streamer_tree_cache_create): Adjust.
6800 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6801 to allow optional nodes array.
6802 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6803 (streamer_tree_cache_append): Likewise.
6804 (streamer_tree_cache_create): Create nodes array optionally
6805 as specified by parameter.
6806 * lto-streamer-out.c (create_output_block): Avoid maintaining
6807 the node array in the writer cache.
6808 (DFS_write_tree): Remove assertion.
6809 (produce_asm_for_decls): Free the out decl state hash table early.
6810 * lto-streamer-in.c (lto_data_in_create): Adjust for
6811 streamer_tree_cache_create prototype change.
6813 2014-04-03 Richard Biener <rguenther@suse.de>
6815 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6816 set TREE_CHAIN to NULL_TREE.
6818 2014-04-03 Richard Biener <rguenther@suse.de>
6820 PR tree-optimization/60740
6821 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6822 over all GIMPLE_COND operands.
6824 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
6826 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
6827 (Weffc++): Remove Scott's numbering, merge lists and reference
6830 2014-04-03 Nick Clifton <nickc@redhat.com>
6832 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
6835 2014-04-03 Martin Jambor <mjambor@suse.cz>
6837 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
6838 mention gcc_unreachable before failing.
6839 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
6842 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
6845 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
6846 inconsistent code and instead mark the context inconsistent.
6847 (possible_polymorphic_call_targets): For inconsistent contexts
6848 return empty complete list.
6850 2014-04-02 Anthony Green <green@moxielogic.com>
6852 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
6853 (extendqisi2, extendhisi2): Define.
6854 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
6855 (WCHAR_TYPE): Change to unsigned int.
6857 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6859 PR tree-optimization/60733
6860 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
6861 insertion point for PHI candidates to be the end of the feeding
6862 block for the PHI argument.
6864 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
6866 PR rtl-optimization/60650
6867 * lra-constraints.c (process_alt_operands): Decrease reject for
6868 earlyclobber matching.
6870 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6872 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
6874 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6876 * config/spu/spu.c (pad_bb): Do not crash when the last
6877 insn is CODE_FOR_blockage.
6879 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
6881 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
6882 lies outside the target mode.
6884 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6887 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6888 software floating point or no floating point registers, do not
6889 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6890 in GPRs that occurs after we tested for GPRs that would never be
6893 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6894 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6895 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6896 specifically allow DDmode, since that does not use the SPE SIMD
6899 2014-04-02 Richard Biener <rguenther@suse.de>
6902 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
6903 MODE_INTs. Properly use negv_optab.
6904 (expand_abs): Likewise.
6906 2014-04-02 Richard Biener <rguenther@suse.de>
6909 * Makefile.in (install-driver): Guard extra installs with special
6912 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
6914 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
6917 2014-04-01 Richard Henderson <rth@redhat.com>
6920 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
6921 alternative enabled before register allocation.
6923 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
6925 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
6926 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
6928 (nios2_large_got_address): Remove unneeded 'sym' parameter.
6929 (nios2_got_address): Update nios2_large_got_address call site.
6930 (nios2_delegitimize_address): New function.
6931 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
6932 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
6933 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
6935 2014-04-01 Martin Husemann <martin@duskware.de>
6937 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
6940 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
6942 PR rtl-optimization/60604
6943 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
6944 check from register_operand.
6945 (register_operand): Redefine in terms of general_operand.
6946 (nonmemory_operand): Use register_operand for the non-constant cases.
6948 2014-04-01 Richard Biener <rguenther@suse.de>
6950 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
6952 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
6954 * doc/invoke.texi (mapp-regs): Clarify.
6956 2014-03-31 Ulrich Drepper <drepper@gmail.com>
6958 * config/i386/avx512fintrin.h (__v32hi): Define type.
6959 (__v64qi): Likewise.
6960 (_mm512_set1_epi8): Define.
6961 (_mm512_set1_epi16): Define.
6962 (_mm512_set4_epi32): Define.
6963 (_mm512_set4_epi64): Define.
6964 (_mm512_set4_pd): Define.
6965 (_mm512_set4_ps): Define.
6966 (_mm512_setr4_epi64): Define.
6967 (_mm512_setr4_epi32): Define.
6968 (_mm512_setr4_pd): Define.
6969 (_mm512_setr4_ps): Define.
6970 (_mm512_setzero_epi32): Define.
6972 2014-03-31 Martin Jambor <mjambor@suse.cz>
6975 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
6976 callsite_arguments_match_p. Updated all callers. Also check types of
6977 corresponding formal parameters and actual arguments.
6978 (not_all_callers_have_enough_arguments_p) Renamed to
6979 some_callers_have_mismatched_arguments_p.
6981 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
6983 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
6985 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
6988 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
6991 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
6993 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
6994 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
6996 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
6997 Use FMAMODE_NOVF512 mode iterator.
6998 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
6999 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7000 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7001 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7003 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7004 Use VF_128_256 mode iterator.
7005 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7008 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7010 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7011 static chain if needed.
7013 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7016 * lra-constraints.c (index_part_to_reg): New.
7017 (process_address): Use it.
7019 2014-03-27 Jeff Law <law@redhat.com>
7020 Jakub Jelinek <jakub@redhat.com>
7023 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7024 gen_rtx_{PLUS,MULT} to build up the address expression.
7026 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7027 creating non-canonical RTL.
7029 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7032 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7033 functions; reorganize to make cheap checks first.
7034 (inline_small_functions): Do not estimate growth when dumping;
7036 * ipa-inline.h (inline_summary): Add min_size.
7037 (growth_likely_positive): New function.
7038 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7039 (set_cond_stmt_execution_predicate): Cleanup.
7040 (estimate_edge_size_and_time): Compute min_size.
7041 (estimate_calls_size_and_time): Likewise.
7042 (estimate_node_size_and_time): Likewise.
7043 (inline_update_overall_summary): Update min_size.
7044 (do_estimate_edge_time): Likewise.
7045 (do_estimate_edge_size): Update.
7046 (do_estimate_edge_hints): Update.
7047 (growth_likely_positive): New function.
7049 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7052 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7053 also if addr has VOIDmode.
7055 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7057 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7058 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7060 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7061 instructions as well as AdvancedSIMD loads.
7063 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7065 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7066 Use crypto_aese type.
7067 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7068 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7069 crypto_aese, crypto_aesmc. Move to types.md.
7070 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7072 * config/arm/iterators.md (crypto_type): Likewise.
7074 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7076 * cgraph.c: Include expr.h and tree-dfa.h.
7077 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7080 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7083 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7084 regs from checking multi-reg pseudos.
7086 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7088 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7090 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7092 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7093 if it would clobber the stack pointer, even temporarily.
7095 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7097 * mode-switching.c: Make small adjustments to the top comment.
7099 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7101 * config/rs6000/constraints.md (wD constraint): New constraint to
7102 match the constant integer to get the top DImode/DFmode out of a
7103 vector in a VSX register.
7105 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7106 match the constant integer to get the top DImode/DFmode out of a
7107 vector in a VSX register.
7109 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7112 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7115 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7116 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7118 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7119 Optimize vec_extract of 64-bit values, where the value being
7120 extracted is in the top word, where we can use scalar
7121 instructions. Add direct move and store support. Combine the big
7122 endian/little endian vector select load support into a single insn.
7123 (vsx_extract_<mode>_internal1): Likewise.
7124 (vsx_extract_<mode>_internal2): Likewise.
7125 (vsx_extract_<mode>_load): Likewise.
7126 (vsx_extract_<mode>_store): Likewise.
7127 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7128 combined into vsx_extract_<mode>_load.
7129 (vsx_extract_<mode>_one_le): Likewise.
7131 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7132 define the top 64-bit vector element.
7134 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7137 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7138 Document vec_vbpermq builtin.
7141 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7142 enable use of xxsldwi and xxpermdi builtin functions.
7143 (vec_xxpermdi): Likewise.
7145 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7146 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7148 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7150 PR rtl-optimization/60650
7151 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7153 (find_spills_for): New.
7154 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7155 Spill all pseudos on the second iteration.
7157 2014-03-27 Marek Polacek <polacek@redhat.com>
7160 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7163 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7165 * config/s390/s390.c (s390_can_use_return_insn): Check for
7166 call-saved FPRs on 31 bit.
7168 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7171 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7172 if they need regimplification, just drop them instead of
7173 calling gimple_regimplify_operands on them.
7175 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7178 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7179 (aarch64_frame_pointer_required): Adjust logic.
7180 (aarch64_can_eliminate): Adjust logic.
7181 (aarch64_override_options_after_change): Adjust logic.
7183 2014-03-27 Dehao Chen <dehao@google.com>
7185 * ipa-inline.c (early_inliner): Update node's inline info.
7187 2014-03-26 Dehao Chen <dehao@google.com>
7189 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7190 compiler inserted conditional jumps for NAN float check.
7192 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7194 * ubsan.h (ubsan_create_data): Change second argument's type
7195 to const location_t *.
7196 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7198 (ubsan_create_data): Change second argument to const location_t *PLOC.
7199 Create Loc field whenever PLOC is non-NULL.
7200 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7201 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7205 * real.c (real_to_integer2): Change type of low to UHWI.
7207 2014-03-26 Tobias Burnus <burnus@net-b.de>
7209 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7210 (CILK_SELF_SPECS): New define.
7211 (driver_self_specs): Use it.
7213 2014-03-26 Richard Biener <rguenther@suse.de>
7215 * tree-pretty-print.c (percent_K_format): Implement special
7216 case for LTO and its stripped down BLOCK tree.
7218 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7221 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7223 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7224 one range is range_int_cst_p, but not both, at least optimize
7225 addition/subtraction of 0 and multiplication by 0 or 1.
7226 * gimple-fold.c (gimple_fold_call): Fold
7227 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7228 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7229 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7231 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7233 PR rtl-optimization/60452
7234 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7235 <case REG>: Return 1 for invalid offsets from the frame pointer.
7237 2014-03-26 Marek Polacek <polacek@redhat.com>
7240 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7243 2014-03-26 Marek Polacek <polacek@redhat.com>
7246 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7249 2014-03-26 Marek Polacek <polacek@redhat.com>
7252 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7253 multiplication in unsigned type.
7255 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7257 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7259 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7261 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7263 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7266 * cif-code.def (UNREACHABLE) New code.
7267 * ipa-inline.c (inline_small_functions): Skip edges to
7268 __builtlin_unreachable.
7269 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7270 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7271 predicate to __bulitin_unreachable.
7272 (set_cond_stmt_execution_predicate): Fix issue when
7273 invert_tree_comparison returns ERROR_MARK.
7274 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7275 propagate to inline clones.
7276 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7278 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7279 * cgraphclones.c (cgraph_clone_node): If call destination is already
7280 ureachable, do not redirect it back.
7281 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7284 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7286 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7287 Do not modify inline clones.
7289 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7291 * config/i386/i386.md (general_sext_operand): New mode attr.
7292 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7293 don't generate (sign_extend (const_int)).
7294 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7295 operands[2]. Use We constraint instead of <i> and
7296 <general_sext_operand> predicate instead of <general_operand>.
7297 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7298 * config/i386/constraints.md (We): New constraint.
7299 * config/i386/predicates.md (x86_64_sext_operand,
7300 sext_operand): New predicates.
7302 2014-03-25 Martin Jambor <mjambor@suse.cz>
7305 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7306 inconsistent devirtualizations to __builtin_unreachable.
7308 2014-03-25 Marek Polacek <polacek@redhat.com>
7311 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7313 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7315 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7316 order of elements for big-endian.
7318 2014-03-25 Richard Biener <rguenther@suse.de>
7321 * gimplify-me.c (gimple_regimplify_operands): Update the
7324 2014-03-25 Martin Jambor <mjambor@suse.cz>
7327 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7328 (lto_output_varpool_node): Likewise.
7329 (input_overwrite_node): Likewise.
7330 (input_varpool_node): Likewise.
7332 2014-03-25 Richard Biener <rguenther@suse.de>
7334 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7335 (run_gcc): Likewise.
7337 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7339 * combine.c (simplify_compare_const): Add MODE argument.
7340 Handle mode_width 0 as very large mode_width.
7341 (try_combine, simplify_comparison): Adjust callers.
7343 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7344 type to avoid signed integer overflow.
7345 * explow.c (plus_constant): Likewise.
7347 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7349 * doc/generic.texi: Correct typos.
7351 2014-03-24 Tobias Burnus <burnus@net-b.de>
7353 * doc/invoke.texi (-flto): Expand section about
7354 using static libraries with LTO.
7356 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7358 PR rtl-optimization/60501
7359 * optabs.def (addptr3_optab): New optab.
7360 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7361 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7362 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7364 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7366 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7368 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7370 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7373 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7374 (_mm256_undefined_ps): Define.
7375 (_mm256_undefined_pd): Define.
7376 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7377 (_mm_undefined_pd): Define.
7378 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7379 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7380 (_mm512_undefined_ps): Define.
7381 (_mm512_undefined_pd): Define.
7382 Use _mm*_undefined_*.
7383 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7385 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7387 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7388 (lshr_simd): DI mode added.
7389 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7390 (aarch64_ushr_simddi): Likewise.
7391 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7392 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7393 (vshrd_n_u64): Likewise.
7395 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7397 * Makefile.in (s-macro_list): Depend on cc1.
7399 2014-03-23 Teresa Johnson <tejohnson@google.com>
7401 * ipa-utils.c (ipa_print_order): Use specified dump file.
7403 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7405 PR rtl-optimization/60601
7406 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7408 * gcc.c (eval_spec_function): Initialize save_growing_value.
7410 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7413 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7414 code == MINUS_EXPR, never swap op0 with op1.
7416 * toplev.c (init_local_tick): Avoid signed integer multiplication
7418 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7419 shift by first operand's bitsize.
7421 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7424 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7426 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7427 TARGET_ISA_64BIT_P(x).
7429 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7431 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7432 pattern for vector nor instead of subtract from splat(-1).
7433 (altivec_expand_vec_perm_const_le): Likewise.
7435 2014-03-21 Richard Henderson <rth@twiddle.net>
7438 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7439 related insns after epilogue_completed.
7441 2014-03-21 Martin Jambor <mjambor@suse.cz>
7444 * cgraph.h (symtab_node): New flag body_removed.
7445 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7446 when removing bodies.
7447 * symtab.c (dump_symtab_base): Dump body_removed flag.
7448 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7449 had their bodies removed.
7451 2014-03-21 Martin Jambor <mjambor@suse.cz>
7454 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7457 2014-03-21 Richard Biener <rguenther@suse.de>
7459 PR tree-optimization/60577
7460 * tree-core.h (struct tree_base): Document nothrow_flag use
7461 in VAR_DECL_NONALIASED.
7462 * tree.h (VAR_DECL_NONALIASED): New.
7463 (may_be_aliased): Adjust.
7464 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7466 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7468 * expr.c (expand_expr_real_1): Remove outdated comment.
7470 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7473 * ira.c (adjust_cleared_regs): Call copy_rtx on
7474 *reg_equiv[REGNO (loc)].src_p before passing it to
7475 simplify_replace_fn_rtx.
7478 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7479 into CONST, put pic register as first operand of PLUS. Use
7480 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7482 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7484 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7486 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7488 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7489 around for store forwarding issue in the FPU on the UT699.
7490 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7491 loads and operations if -mfix-ut699 is specified.
7492 (divtf3_hq): Tweak attribute.
7493 (sqrttf2_hq): Likewise.
7495 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7497 * calls.c (store_one_arg): Remove incorrect const qualification on the
7498 type of the temporary.
7499 * cfgexpand.c (expand_return): Likewise.
7500 * expr.c (expand_constructor): Likewise.
7501 (expand_expr_real_1): Likewise.
7503 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7505 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7508 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7511 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7513 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7515 * config/arm/aarch-common-protos.h
7516 (alu_cost_table): Fix spelling of "extend".
7517 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7519 2014-03-19 Richard Biener <rguenther@suse.de>
7522 * tree-core.h (tree_type_common): Re-order pointer members
7523 to reduce recursion depth during GC walks.
7525 2014-03-19 Marek Polacek <polacek@redhat.com>
7528 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7529 before accessing it.
7531 2014-03-19 Richard Biener <rguenther@suse.de>
7534 * lto-streamer-in.c (input_function): In WPA stage do not drop
7537 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7539 PR tree-optimization/60559
7540 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7541 with build_zero_cst assignment.
7543 2014-03-18 Kai Tietz <ktietz@redhat.com>
7545 PR rtl-optimization/56356
7546 * sdbout.c (sdbout_parms): Verify that parms'
7547 incoming argument is valid.
7548 (sdbout_reg_parms): Likewise.
7550 2014-03-18 Richard Henderson <rth@redhat.com>
7553 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7554 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7555 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7557 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7559 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7560 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7561 Italicize plugin event names in description. Explain that
7562 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7563 Remind that no GCC functions should be called after PLUGIN_FINISH.
7564 Explain what pragmas with expansion are.
7566 2014-03-18 Martin Liska <mliska@suse.cz>
7568 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7569 gimple call statement is update.
7570 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7571 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7573 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7576 * ubsan.c (ubsan_instrument_unreachable): Call
7577 initialize_sanitizer_builtins.
7578 (ubsan_pass): Likewise.
7581 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7582 varpool_finalize_decl instead of rest_of_decl_compilation.
7584 2014-03-18 Richard Biener <rguenther@suse.de>
7586 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7587 by using bitmap_and_compl instead of bitmap_and_compl_into.
7588 (df_rd_transfer_function): Likewise.
7590 2014-03-18 Richard Biener <rguenther@suse.de>
7592 * doc/lto.texi (fresolution): Fix typo.
7594 2014-03-18 Richard Biener <rguenther@suse.de>
7596 * doc/invoke.texi (flto): Update for changes in 4.9.
7598 2014-03-18 Richard Biener <rguenther@suse.de>
7600 * doc/loop.texi: Remove section on the removed lambda framework.
7601 Update loop docs with recent changes in preserving loop structure.
7603 2014-03-18 Richard Biener <rguenther@suse.de>
7605 * doc/lto.texi (-fresolution): Document.
7607 2014-03-18 Richard Biener <rguenther@suse.de>
7609 * doc/contrib.texi: Adjust my name.
7611 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7614 * internal-fn.c: Include diagnostic-core.h.
7615 (expand_BUILTIN_EXPECT): New function.
7616 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7617 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7618 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7619 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7621 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7622 Revert 3 argument __builtin_expect code.
7623 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7624 * gimple-fold.c (gimple_fold_call): Likewise.
7625 * tree.h (fold_builtin_expect): New prototype.
7626 * builtins.c (build_builtin_expect_predicate): Add predictor
7627 argument, if non-NULL, create 3 argument __builtin_expect.
7628 (fold_builtin_expect): No longer static. Add ARG2 argument,
7629 pass it through to build_builtin_expect_predicate.
7630 (fold_builtin_2): Adjust caller.
7631 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7632 * internal-fn.def (BUILTIN_EXPECT): New.
7634 2014-03-18 Tobias Burnus <burnus@net-b.de>
7637 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7638 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7639 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7641 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7644 * predict.c (combine_predictions_for_bb): Fix up formatting.
7645 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7646 fill what it points to if non-NULL.
7647 (tree_predict_by_opcode): Adjust caller, use the predictor.
7648 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7650 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7652 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7653 proper constant for the store mode.
7655 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7657 * symtab.c (change_decl_assembler_name): Fix transparent alias
7660 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7662 * config/aarch64/aarch64.c: Correct the comments about the
7663 aarch64 stack layout.
7665 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7667 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7668 check for GF_OMP_FOR_KIND_FOR.
7670 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7672 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7673 ymm and zmm register names.
7675 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7678 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7679 note creation for the 2010-08-31 changes.
7681 2014-03-17 Marek Polacek <polacek@redhat.com>
7684 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7685 as -fno-tree-loop-vectorize.
7686 (expand_omp_simd): Likewise.
7688 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7690 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7691 (eligible_for_call_delay): New prototype.
7692 * config/sparc/sparc.c (tls_call_delay): Rename into...
7693 (eligible_for_call_delay): ...this. Return false if the instruction
7694 cannot be put in the delay slot of a branch.
7695 (eligible_for_restore_insn): Simplify.
7696 (eligible_for_return_delay): Return false if the instruction cannot be
7697 put in the delay slot of a branch and simplify.
7698 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7699 put in the delay slot of a branch.
7700 * config/sparc/sparc.md (fix_ut699): New attribute.
7701 (tls_call_delay): Delete.
7702 (in_call_delay): Reimplement.
7703 (eligible_for_sibcall_delay): Rename into...
7704 (in_sibcall_delay): ...this.
7705 (eligible_for_return_delay): Rename into...
7706 (in_return_delay): ...this.
7707 (in_branch_delay): Reimplement.
7708 (in_uncond_branch_delay): Delete.
7709 (in_annul_branch_delay): Delete.
7711 2014-03-14 Richard Henderson <rth@redhat.com>
7714 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7715 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7716 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7717 (floathi splitters): Remove.
7718 (float<SWI48x>xf2): New pattern.
7719 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7720 code that tried to handle DImode for 32-bit, but which was excluded
7721 by the pattern's condition. Drop allocation of stack temporary.
7722 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7723 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7724 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7725 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7726 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7727 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7728 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7729 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7730 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7731 (*float<SWI48x><X87MODEF>2_i387): Remove.
7732 (all float _with_temp splitters): Remove.
7733 (*float<SWI48x><MODEF>2_i387): New pattern.
7734 (*float<SWI48><MODEF>2_sse): New pattern.
7735 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7736 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7738 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7739 Marek Polacek <polacek@redhat.com>
7742 * common.opt (dump_base_name_prefixed): New Variable.
7743 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7744 if x_dump_base_name_prefixed is already set, set it at the end.
7746 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7748 PR rtl-optimization/60508
7749 * lra-constraints.c (get_reload_reg): Add new parameter
7751 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7752 Pass the new parameter values.
7754 2014-03-14 Richard Biener <rguenther@suse.de>
7756 * common.opt: Revert unintented changes from r205065.
7759 2014-03-14 Richard Biener <rguenther@suse.de>
7762 * cfghooks.c (split_block): Properly adjust all loops the
7763 block was a latch of.
7765 2014-03-14 Martin Jambor <mjambor@suse.cz>
7768 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7771 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7774 * config/avr/avr.c (avr_set_current_function): Pass function name
7775 through default_strip_name_encoding before sanity checking instead
7776 of skipping the first char of the assembler name.
7778 2014-03-13 Richard Henderson <rth@redhat.com>
7781 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7782 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7783 * config/i386/i386-protos.h: Likewise.
7784 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7785 in the expander instead of a splitter.
7786 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7787 any possibility of requiring a memory.
7788 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7789 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7790 (fp branch splitters): Update for ix86_split_fp_branch.
7791 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7792 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7793 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7794 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7795 (*fop_<MODEF>_3_i387): Likewise.
7796 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7797 (splitters for the fop_* register patterns): Remove.
7798 (fscalexf4_i387): Rename from *fscalexf4_i387.
7799 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7801 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7803 PR tree-optimization/59779
7804 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7805 type for bitsize and maxsize instead of HOST_WIDE_INT.
7807 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7809 PR rtl-optimization/57320
7810 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7811 the CFG after thread_prologue_and_epilogue_insns.
7813 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7815 PR rtl-optimization/57189
7816 * lra-constraints.c (process_alt_operands): Disfavor spilling
7819 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7821 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7823 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7825 PR tree-optimization/59025
7827 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
7828 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
7830 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
7833 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
7834 calls of avr_out_plus_1.
7836 2014-03-13 Bin Cheng <bin.cheng@arm.com>
7838 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
7839 BB's single pred and update the father loop's latch info later.
7841 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
7843 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
7847 (VEC_base): Likewise.
7848 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
7849 registers, we need to swap double words in little endian mode.
7851 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
7852 to be a container mode for 128-bit integer operations added in ISA
7853 2.07. Unlike TImode and PTImode, the preferred register set is
7854 the Altivec/VMX registers for the 128-bit operations.
7856 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
7858 (rs6000_split_128bit_ok_p): Likewise.
7860 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
7861 macros for creating ISA 2.07 normal and overloaded builtin
7862 functions with 3 arguments.
7863 (BU_P8V_OVERLOAD_3): Likewise.
7864 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
7865 for use as overloaded functions.
7866 (VPERM_1TI_UNS): Likewise.
7867 (VSEL_1TI): Likewise.
7868 (VSEL_1TI_UNS): Likewise.
7869 (ST_INTERNAL_1ti): Likewise.
7870 (LD_INTERNAL_1ti): Likewise.
7871 (XXSEL_1TI): Likewise.
7872 (XXSEL_1TI_UNS): Likewise.
7873 (VPERM_1TI): Likewise.
7874 (VPERM_1TI_UNS): Likewise.
7875 (XXPERMDI_1TI): Likewise.
7876 (SET_1TI): Likewise.
7877 (LXVD2X_V1TI): Likewise.
7878 (STXVD2X_V1TI): Likewise.
7879 (VEC_INIT_V1TI): Likewise.
7880 (VEC_SET_V1TI): Likewise.
7881 (VEC_EXT_V1TI): Likewise.
7882 (EQV_V1TI): Likewise.
7883 (NAND_V1TI): Likewise.
7884 (ORC_V1TI): Likewise.
7885 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
7886 added in ISA 2.07. Add both normal 'altivec' builtins, and the
7888 (VADDUQM): Likewise.
7889 (VSUBCUQ): Likewise.
7890 (VADDEUQM): Likewise.
7891 (VADDECUQ): Likewise.
7892 (VSUBEUQM): Likewise.
7893 (VSUBECUQ): Likewise.
7895 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
7896 __int128_t and __uint128_t types.
7897 (__uint128_type): Likewise.
7898 (altivec_categorize_keyword): Add support for vector __int128_t,
7899 vector __uint128_t, vector __int128, and vector unsigned __int128
7900 as a container type for TImode operations that need to be done in
7901 VSX/Altivec registers.
7902 (rs6000_macro_to_expand): Likewise.
7903 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
7904 to support 128-bit integer instructions vaddcuq, vadduqm,
7905 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
7906 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
7908 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
7909 for V1TImode, and set up preferences to use VSX/Altivec registers.
7910 Setup VSX reload handlers.
7911 (rs6000_debug_reg_global): Likewise.
7912 (rs6000_init_hard_regno_mode_ok): Likewise.
7913 (rs6000_preferred_simd_mode): Likewise.
7914 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
7915 (easy_altivec_constant): Likewise.
7916 (output_vec_const_move): Likewise.
7917 (rs6000_expand_vector_set): Convert V1TImode set and extract to
7919 (rs6000_expand_vector_extract): Likewise.
7920 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
7922 (rs6000_const_vec): Add support for V1TImode.
7923 (rs6000_emit_le_vsx_load): Swap double words when loading or
7924 storing TImode/V1TImode.
7925 (rs6000_emit_le_vsx_store): Likewise.
7926 (rs6000_emit_le_vsx_move): Likewise.
7927 (rs6000_emit_move): Add support for V1TImode.
7928 (altivec_expand_ld_builtin): Likewise.
7929 (altivec_expand_st_builtin): Likewise.
7930 (altivec_expand_vec_init_builtin): Likewise.
7931 (altivec_expand_builtin): Likewise.
7932 (rs6000_init_builtins): Add support for V1TImode type. Add
7933 support for ISA 2.07 128-bit integer builtins. Define type names
7934 for the VSX/Altivec vector types.
7935 (altivec_init_builtins): Add support for overloaded vector
7936 functions with V1TImode type.
7937 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
7938 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
7940 (rs6000_split_128bit_ok_p): Likewise.
7941 (rs6000_handle_altivec_attribute): Create V1TImode from vector
7942 __int128_t and vector __uint128_t.
7944 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
7945 and mode attributes.
7952 (VS_scalar): Likewise.
7953 (VS_double): Likewise.
7954 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
7956 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
7957 we support the ISA 2.07 128-bit integer arithmetic instructions.
7958 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
7959 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
7960 and TImode types for use with the builtin functions.
7961 (V1TI_type_node): Likewise.
7962 (unsigned_V1TI_type_node): Likewise.
7963 (intTI_type_internal_node): Likewise.
7964 (uintTI_type_internal_node): Likewise.
7966 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
7967 128-bit builtin functions.
7968 (UNSPEC_VADDEUQM): Likewise.
7969 (UNSPEC_VADDECUQ): Likewise.
7970 (UNSPEC_VSUBCUQ): Likewise.
7971 (UNSPEC_VSUBEUQM): Likewise.
7972 (UNSPEC_VSUBECUQ): Likewise.
7973 (VM): Add V1TImode to vector mode iterators.
7975 (VI_unit): Likewise.
7976 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
7977 (altivec_vaddcuq): Likewise.
7978 (altivec_vsubuqm): Likewise.
7979 (altivec_vsubcuq): Likewise.
7980 (altivec_vaddeuqm): Likewise.
7981 (altivec_vaddecuq): Likewise.
7982 (altivec_vsubeuqm): Likewise.
7983 (altivec_vsubecuq): Likewise.
7985 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
7987 (BOOL_128): Likewise.
7988 (BOOL_REGS_OUTPUT): Likewise.
7989 (BOOL_REGS_OP1): Likewise.
7990 (BOOL_REGS_OP2): Likewise.
7991 (BOOL_REGS_UNARY): Likewise.
7992 (BOOL_REGS_AND_CR0): Likewise.
7994 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
7995 128-bit integer builtin support.
7996 (vec_vadduqm): Likewise.
7997 (vec_vaddecuq): Likewise.
7998 (vec_vaddeuqm): Likewise.
7999 (vec_vsubecuq): Likewise.
8000 (vec_vsubeuqm): Likewise.
8001 (vec_vsubcuq): Likewise.
8002 (vec_vsubuqm): Likewise.
8004 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8005 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8006 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8007 128-bit integer add/subtract to ISA 2.07.
8009 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8011 * config/arc/arc.c (arc_predicate_delay_insns):
8012 Fix third argument passed to conditionalize_nonjump.
8014 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8016 * config/aarch64/aarch64-builtins.c
8017 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8018 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8019 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8020 instead of __builtin_lfloor.
8021 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8023 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8025 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8026 (tree_ssa_ifcombine_bb_1): New function.
8027 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8028 is an empty forwarder block to then_bb or vice versa and then_bb
8029 and else_bb are effectively swapped.
8031 2014-03-12 Christian Bruel <christian.bruel@st.com>
8034 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8035 REG_CFA_DEF_CFA note.
8036 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8037 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8039 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8041 PR tree-optimization/60454
8042 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8044 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8046 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8047 Do not define target_cpu_default2 to generic.
8048 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8049 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8050 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8052 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8053 Marc Glisse <marc.glisse@inria.fr>
8055 PR tree-optimization/60502
8056 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8057 instead of build_low_bits_mask.
8059 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8062 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8063 if there are multiple uses, but op doesn't live on E edge.
8064 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8065 clobber stmts before __builtin_unreachable.
8067 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8069 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8070 hard_frame_pointer_rtx.
8071 * cse.c (cse_insn): Remove volatile check.
8072 * cselib.c (cselib_process_insn): Likewise.
8073 * dse.c (scan_insn): Likewise.
8075 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8077 * config/arc/arc.c (conditionalize_nonjump): New function,
8079 (arc_ifcvt): ... this.
8080 (arc_predicate_delay_insns): Use it.
8082 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8084 * config/arc/predicates.md (extend_operand): During/after reload,
8085 allow const_int_operand.
8086 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8087 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8088 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8090 (umulsi3_highpart_i): Likewise.
8092 2014-03-11 Richard Biener <rguenther@suse.de>
8094 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8095 Add asserts to guard possible wrong-code bugs.
8097 2014-03-11 Richard Biener <rguenther@suse.de>
8099 PR tree-optimization/60429
8100 PR tree-optimization/60485
8101 * tree-ssa-structalias.c (set_union_with_increment): Properly
8102 take into account all fields that overlap the shifted vars.
8103 (do_sd_constraint): Likewise.
8104 (do_ds_constraint): Likewise.
8105 (get_constraint_for_ptr_offset): Likewise.
8107 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8109 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8110 (nios2_compute_frame_layout):
8111 Add calculation of cfun->machine->fp_save_offset.
8112 (nios2_expand_prologue): Correct setting of frame pointer register
8114 (nios2_expand_epilogue): Update recovery of stack pointer from
8115 frame pointer accordingly.
8116 (nios2_initial_elimination_offset): Update calculation of offset
8117 for eliminating to HARD_FRAME_POINTER_REGNUM.
8119 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8122 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8123 cgraph_get_create_node on VAR_DECLs.
8125 2014-03-10 Richard Biener <rguenther@suse.de>
8128 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8130 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8132 * config/vms/vms.opt (vms_float_format): New variable.
8134 2014-03-08 Tobias Burnus <burnus@net-b.de>
8136 * doc/invoke.texi (-fcilkplus): Update implementation status.
8138 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8139 Richard Biener <rguenther@suse.de>
8141 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8142 consistently accross all TUs.
8143 (run_gcc): Enable -fshort-double automatically at link at link-time
8144 and disallow override.
8146 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8149 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8150 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8151 if they can't be used.
8153 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8155 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8156 for Solaris 11/x86 ld.
8157 * configure: Regenerate.
8159 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8161 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8162 (LIB_TLS_SPEC): Save as ld_tls_libs.
8163 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8164 (HAVE_AS_IX86_TLSLDM): New test.
8165 * configure, config.in: Regenerate.
8166 * config/i386/i386.c (legitimize_tls_address): Fall back to
8167 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8168 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8169 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8170 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8172 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8174 * common.opt (fira-loop-pressure): Mark as optimization.
8176 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8178 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8179 an OpenMP mappable type.
8181 2014-03-06 Matthias Klose <doko@ubuntu.com>
8183 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8184 MULTILIB_OSDIRNAMES is not defined.
8186 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8187 Meador Inge <meadori@codesourcery.com>
8190 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8191 (arm_legitimize_address): Call legitimize_tls_address for any
8192 arm_tls_referenced_p expression, handle constant addend. Call it
8193 before testing for !TARGET_ARM.
8194 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8196 2014-03-06 Richard Biener <rguenther@suse.de>
8202 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8204 * tree-streamer.c (record_common_node): Assert we don't record
8205 nodes with type double.
8206 (preload_common_node): Skip type double, complex double and double
8207 pointer since it is now frontend dependent due to fshort-double option.
8209 2014-03-06 Richard Biener <rguenther@suse.de>
8211 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8212 or -fno-lto is specified and the linker has full plugin support.
8213 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8214 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8215 * lto-wrapper.c (merge_and_complain): Merge compile-time
8216 optimization levels.
8217 (run_gcc): And pass it through to the link options.
8219 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8223 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8225 * cselib.c (remove_useless_values): Skip to avoid quadratic
8226 behavior if the condition moved from...
8227 (cselib_process_insn): ... here holds.
8229 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8232 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8233 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8236 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8237 (TM_H): Add x86-tune.def.
8239 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8241 * config/aarch64/aarch64.c (generic_tunings):
8242 Use cortexa57_extra_costs.
8244 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8247 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8248 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8249 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8252 2014-03-04 Heiher <r@hev.cc>
8254 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8255 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8257 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8259 * config/i386/predicates.md (const2356_operand): Change to ...
8260 (const2367_operand): ... this.
8261 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8263 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8264 (*avx512pf_scatterpf<mode>sf): Ditto.
8265 (avx512pf_scatterpf<mode>df): Ditto.
8266 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8267 (*avx512pf_scatterpf<mode>df): Ditto.
8268 * config/i386/i386.c (ix86_expand_builtin): Update
8269 incorrect hint operand error message.
8271 2014-03-04 Richard Biener <rguenther@suse.de>
8273 * lto-section-in.c (lto_get_section_data): Fix const cast.
8275 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8277 * tree-streamer.c (record_common_node): Assert we don't record
8278 nodes with type double.
8279 (preload_common_node): Skip type double, complex double and double
8280 pointer since it is now frontend dependent due to fshort-double option.
8282 2014-03-04 Richard Biener <rguenther@suse.de>
8285 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8286 (lto_input_toplevel_asms): Likewise.
8287 * lto-section-in.c (lto_get_section_data): Instead do it here
8290 2014-03-04 Richard Biener <rguenther@suse.de>
8292 PR tree-optimization/60382
8293 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8294 dead PHIs a reduction.
8296 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8298 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8300 (_mm_prefetch): Move out of GCC target("sse") pragma.
8301 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8302 GCC target("prfchw") pragma.
8303 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8305 * config/i386/i386.c (ix86_option_override_internal): Enable
8306 -mprfchw with -mprefetchwt1.
8308 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8310 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8313 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8315 * opts.h (CL_PCH_IGNORE): Define.
8316 * targhooks.c (option_affects_pch_p):
8317 Return false for options that have CL_PCH_IGNORE set.
8318 * opt-functions.awk: Process PchIgnore.
8319 * doc/options.texi: Document PchIgnore.
8321 * config/arc/arc.opt (misize): Add PchIgnore property.
8323 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8325 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8326 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8327 constraint on constants to permit them being loaded into
8328 GENERAL_REGS or BASE_REGS.
8330 2014-03-03 Nick Clifton <nickc@redhat.com>
8332 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8333 anti-cacnonical alternatives.
8334 (negandhi3_real): New pattern.
8335 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8337 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8339 * config/avr/avr-mcus.def: Remove atxmega16x1.
8340 * config/avr/avr-tables.opt: Regenerate.
8341 * config/avr/t-multilib: Regenerate.
8342 * doc/avr-mmcu.texi: Regenerate.
8344 2014-03-03 Tobias Grosser <tobias@grosser.es>
8345 Mircea Namolaru <mircea.namolaru@inria.fr>
8347 PR tree-optimization/58028
8348 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8351 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8353 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8354 not handled by recognizers.
8356 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8359 * function.c (expand_function_end): Don't emit
8360 clobber_return_register sequence if clobber_after is a BARRIER.
8361 * cfgexpand.c (construct_exit_block): Append instructions before
8362 return_label to prev_bb.
8364 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8366 * config/rs6000/constraints.md: Document reserved use of "wc".
8368 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8371 * ipa.c (function_and_variable_visibility): When dissolving comdat
8372 group, also set all symbols to local.
8374 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8379 2013-12-14 Jan Hubicka <jh@suse.cz>
8381 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8383 2014-03-02 Jon Beniston <jon@beniston.com>
8389 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8390 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8391 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8392 (simple_return, *simple_return): New patterns
8393 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8394 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8396 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8398 * dwarf2out.c (gen_subprogram_die): Tidy.
8400 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8403 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8404 (*mov_t_msb_neg_negc): ... this new insn.
8406 2014-02-28 Jason Merrill <jason@redhat.com>
8409 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8412 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8415 * dwarf2out.c (decltype_auto_die): New static.
8416 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8417 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8418 (is_cxx_auto): Likewise.
8420 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8422 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8423 we are not using general regs only.
8425 2014-02-28 Richard Biener <rguenther@suse.de>
8428 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8429 previous fix and only allow to remove trivial pre-headers
8430 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8431 (remove_forwarder_block): Properly update the latch of a loop.
8433 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8436 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8437 (cselib_preserved_hash_table): New.
8438 (preserve_constants_and_equivs): Move preserved vals to it.
8439 (cselib_find_slot): Look it up first.
8440 (cselib_init): Initialize it.
8441 (cselib_finish): Release it.
8442 (dump_cselib_table): Dump it.
8444 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8447 * cselib.c (remove_useless_values): Skip to avoid quadratic
8448 behavior if the condition moved from...
8449 (cselib_process_insn): ... here holds.
8451 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8454 * var-tracking.c (vt_initialize): Apply the same condition to
8455 preserve the CFA base value.
8457 2014-02-28 Joey Ye <joey.ye@arm.com>
8460 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8461 if reload in progress or completed.
8463 2014-02-28 Tobias Burnus <burnus@net-b.de>
8466 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8469 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8471 * doc/tm.texi.in (Condition Code Status): Update documention for
8472 relative locations of cc0-setter and cc0-user.
8474 2014-02-27 Jeff Law <law@redhat.com>
8476 PR rtl-optimization/52714
8477 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8478 into two independent simple sets, if I3 is a jump, ensure the
8479 pattern we place into I3 is a (set (pc) ...).
8481 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8482 Jeff Law <law@redhat.com>
8484 PR rtl-optimization/49847
8485 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8486 are in different blocks.
8487 * doc/tm.texi (Condition Code Status): Update documention for
8488 relative locations of cc0-setter and cc0-user.
8490 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8493 * lra.c (lra_emit_add): Check SUBREG too.
8495 2014-02-27 Andreas Schwab <schwab@suse.de>
8497 * config/m68k/m68k.c (m68k_option_override): Disable
8498 -flive-range-shrinkage for classic m68k.
8499 (m68k_override_options_after_change): Likewise.
8501 2014-02-27 Marek Polacek <polacek@redhat.com>
8504 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8505 -Wmaybe-uninitialized.
8507 2014-02-27 Alan Modra <amodra@gmail.com>
8510 * reload1.c (emit_input_reload_insns): When reload_override_in,
8511 set old to rl->in_reg when rl->in_reg is a subreg.
8513 2014-02-26 Richard Biener <rguenther@suse.de>
8516 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8518 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8520 * common/config/i386/predicates.md (const1256_operand): Remove.
8521 (const2356_operand): New.
8522 (const_1_to_2_operand): Remove.
8523 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8524 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8525 (*avx512pf_gatherpf<mode>sf): Ditto.
8526 (avx512pf_gatherpf<mode>df): Ditto.
8527 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8528 (*avx512pf_gatherpf<mode>df): Ditto.
8529 (avx512pf_scatterpf<mode>sf): Ditto.
8530 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8531 (*avx512pf_scatterpf<mode>sf): Ditto.
8532 (avx512pf_scatterpf<mode>df): Ditto.
8533 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8534 (*avx512pf_scatterpf<mode>df): Ditto.
8535 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8537 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8539 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8540 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8541 (_mm512_mask_testn_epi64_mask): Move to ...
8542 * config/i386/avx512cdintrin.h: Here.
8543 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8544 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8545 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8546 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8547 TARGET_AVX512F from TARGET_AVX512CD.
8549 2014-02-26 Richard Biener <rguenther@suse.de>
8552 * ipa.c (walk_polymorphic_call_targets): Properly guard
8553 call to inline_update_overall_summary.
8555 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8558 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8559 and latches only if requested. Fix latch if it is removed.
8560 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8561 LOOPS_HAVE_PREHEADERS.
8563 2014-02-25 Andrew Pinski <apinski@cavium.com>
8565 * builtins.c (expand_builtin_thread_pointer): Create a new target
8566 when the target is NULL.
8568 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8570 PR rtl-optimization/60317
8571 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8572 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8573 * lra-assigns.c: Include params.h.
8574 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8575 other reload pseudos considerations.
8577 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8579 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8580 to use canonical form for nor<mode>3.
8582 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8585 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8588 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8590 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8591 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8592 (ix86_handle_option): Handle OPT_mprefetchwt1.
8593 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8594 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8596 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8597 OPTION_MASK_ISA_PREFETCHWT1.
8598 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8599 (PTA_PREFETCHWT1): New.
8600 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8601 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8602 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8603 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8604 (*prefetch_avx512pf_<mode>_: Change into ...
8605 (*prefetch_prefetchwt1_<mode>: This.
8606 * config/i386/i386.opt (mprefetchwt1): New.
8607 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8608 (_mm_prefetch): Handle intent to write.
8609 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8611 2014-02-25 Richard Biener <rguenther@suse.de>
8614 * emit-rtl.c (mem_attrs_htab): Remove.
8615 (mem_attrs_htab_hash): Likewise.
8616 (mem_attrs_htab_eq): Likewise.
8617 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8618 (init_emit_once): Do not allocate mem_attrs_htab.
8620 2014-02-25 Richard Biener <rguenther@suse.de>
8623 * lto-opts.c (lto_write_options): Output non-explicit conservative
8624 -fwrapv, -fno-trapv and -fno-strict-overflow.
8625 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8626 (run_gcc): And pass them through.
8628 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8630 * sel-sched.c (calculate_new_fences): New parameter ptime.
8631 Calculate it as a maximum over all fence cycles.
8632 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8633 Print the final schedule timing when sched_verbose.
8635 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8637 PR rtl-optimization/60292
8638 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8639 bit fot the fence instruction.
8641 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8643 * calls.h: Fix typo in comment.
8645 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8647 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8648 adjusting offsetable addresses.
8650 2014-02-24 Guozhi Wei <carrot@google.com>
8652 * sparseset.h (sparseset_pop): Fix the wrong index.
8654 2014-02-24 Walter Lee <walt@tilera.com>
8656 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8657 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8659 * common/config/tilegx/tilegx-common.c
8660 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8661 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8663 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8664 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8665 (tilegx_gimplify_va_arg_expr): Handle big endian.
8666 (tilegx_expand_unaligned_load): Ditto.
8667 (tilegx_expand_unaligned_store): Ditto.
8668 (TARGET_RETURN_IN_MSB): New.
8669 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8670 (TARGET_ENDIAN_DEFAULT): New.
8671 (TARGET_BIG_ENDIAN): Handle big endian.
8672 (BYTES_BIG_ENDIAN): Ditto.
8673 (WORDS_BIG_ENDIAN): Ditto.
8674 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8677 * config/tilegx/tilegx.md (extv): Handle big endian.
8679 (insn_st<n>): Ditto.
8680 (insn_st<n>_add<bitsuffix>): Ditto.
8681 (insn_stnt<n>): Ditto.
8682 (insn_stnt<n>_add<bitsuffix>):Ditto.
8683 (vec_interleave_highv8qi): Handle big endian.
8684 (vec_interleave_highv8qi_be): New.
8685 (vec_interleave_highv8qi_le): New.
8686 (insn_v1int_h): Handle big endian.
8687 (vec_interleave_lowv8qi): Handle big endian.
8688 (vec_interleave_lowv8qi_be): New.
8689 (vec_interleave_lowv8qi_le): New.
8690 (insn_v1int_l): Handle big endian.
8691 (vec_interleave_highv4hi): Handle big endian.
8692 (vec_interleave_highv4hi_be): New.
8693 (vec_interleave_highv4hi_le): New.
8694 (insn_v2int_h): Handle big endian.
8695 (vec_interleave_lowv4hi): Handle big endian.
8696 (vec_interleave_lowv4hi_be): New.
8697 (vec_interleave_lowv4hi_le): New.
8698 (insn_v2int_l): Handle big endian.
8699 (vec_interleave_highv2si): Handle big endian.
8700 (vec_interleave_highv2si_be): New.
8701 (vec_interleave_highv2si_le): New.
8702 (insn_v4int_h): Handle big endian.
8703 (vec_interleave_lowv2si): Handle big endian.
8704 (vec_interleave_lowv2si_be): New.
8705 (vec_interleave_lowv2si_le): New.
8706 (insn_v4int_l): Handle big endian.
8707 * config/tilegx/tilegx.opt (mbig-endian): New option.
8708 (mlittle-endian): New option.
8709 * doc/install.texi: Document tilegxbe-linux.
8710 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8712 2014-02-24 Martin Jambor <mjambor@suse.cz>
8715 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8716 there are no parameter descriptors.
8718 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8720 PR rtl-optimization/60268
8721 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8722 initialization to ...
8723 (sched_rgn_init): ... here.
8724 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8726 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8728 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8731 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8733 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8736 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8738 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8739 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8741 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8743 * config/microblaze/predicates.md: Add cmp_op predicate.
8744 * config/microblaze/microblaze.md: Add branch_compare instruction
8745 which uses cmp_op predicate and emits cmp insn before branch.
8746 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8747 to microblaze_expand_conditional_branch and consolidate logic.
8748 (microblaze_expand_conditional_branch): emit branch_compare
8749 insn instead of handling cmp op separate from branch insn.
8751 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8753 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8756 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8758 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8759 define_insn with define_expand and new define_insn
8760 *altivec_lve<VI_char>x_internal.
8761 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8762 and new define_insn *altivec_stve<VI_char>x_internal.
8763 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8765 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8767 (altivec_expand_stvex_be): New function.
8769 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8771 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8772 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8773 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8774 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8776 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8779 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8780 instead of emit_move_insn.
8782 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8784 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8786 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8787 gen_altivec_vsumsws.
8789 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8791 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8792 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8793 (altivec_lvxl_<mode>): New define_expand incorporating
8794 -maltivec=be semantics where needed.
8795 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8796 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8797 semantics where needed.
8798 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8799 (altivec_stvx_<mode>): New define_expand incorporating
8800 -maltivec=be semantics where needed.
8801 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8802 VM2 iterator instead of V4SI.
8803 (altivec_stvxl_<mode>): New define_expand incorporating
8804 -maltivec=be semantics where needed.
8805 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8806 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8807 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8808 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8809 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8810 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8811 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8812 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8813 ALTIVEC_BUILTIN_STVXL.
8814 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8815 (altivec_expand_stvx_be): Likewise.
8816 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8817 (altivec_expand_lvx_be): Likewise.
8818 (altivec_expand_stvx_be): Likewise.
8819 (altivec_expand_builtin): Add cases for
8820 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8821 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8822 (altivec_init_builtins): Add definitions for
8823 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8824 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
8826 2014-02-21 Catherine Moore <clm@codesourcery.com>
8828 * doc/invoke.texi (mvirt, mno-virt): Document.
8829 * config/mips/mips.opt (mvirt): New option.
8830 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
8832 2014-02-21 Richard Biener <rguenther@suse.de>
8834 PR tree-optimization/60276
8835 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
8836 (STMT_VINFO_MIN_NEG_DIST): New macro.
8837 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
8838 STMT_VINFO_MIN_NEG_DIST.
8839 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
8840 made for negative dependence distances still hold.
8842 2014-02-21 Richard Biener <rguenther@suse.de>
8845 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
8846 DECL_INITIAL for globals not in the current function context.
8848 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8850 PR tree-optimization/56490
8851 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
8852 * tree-ssa-uninit.c: Include params.h.
8853 (compute_control_dep_chain): Add num_calls argument, return false
8854 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
8855 num_calls to recursive call.
8856 (find_predicates): Change dep_chain into normal array,
8857 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
8858 variable and adjust compute_control_dep_chain caller.
8859 (find_def_preds): Likewise.
8861 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
8863 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
8864 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
8866 2014-02-21 Nick Clifton <nickc@redhat.com>
8868 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
8869 (pushhi1): Likewise.
8870 (popqi1): Add mode to pre_dec.
8873 2014-02-21 Jakub Jelinek <jakub@redhat.com>
8875 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
8876 mode for mask of V8SFmode permutation.
8878 2014-02-20 Richard Henderson <rth@redhat.com>
8881 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
8882 a new pseudo for OLDVAL.
8884 2014-02-20 Jakub Jelinek <jakub@redhat.com>
8887 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
8888 gen_reg_rtx if d->testing_p.
8889 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
8890 if d->testing_p and we will certainly return true.
8891 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
8894 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
8896 * emit-rtl.c (gen_reg_rtx): Assert that
8897 crtl->emit.regno_pointer_align_length is non-zero.
8899 2014-02-20 Richard Henderson <rth@redhat.com>
8902 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
8903 on failure the store back into EXPECT.
8905 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
8906 Sandra Loosemore <sandra@codesourcery.com>
8908 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
8909 * config/nios2/nios2.c (nios2_function_profiler): Add
8910 -fPIC (flag_pic == 2) support.
8911 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
8912 (nios2_large_offset_p): New function.
8913 (nios2_unspec_reloc_p): Move up position, update to use
8914 nios2_large_offset_p.
8915 (nios2_unspec_address): Remove function.
8916 (nios2_unspec_offset): New function.
8917 (nios2_large_got_address): New function.
8918 (nios2_got_address): Add large offset support.
8919 (nios2_legitimize_tls_address): Update usage of removed and new
8921 (nios2_symbol_binds_local_p): New function.
8922 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
8923 (nios2_legitimize_address): Update to use nios2_large_offset_p.
8924 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
8925 (nios2_print_operand): Merge H/L processing, add hiadj/lo
8926 processing for (const (unspec ...)).
8927 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
8929 2014-02-20 Richard Biener <rguenther@suse.de>
8931 * tree-cfg.c (replace_uses_by): Mark altered BBs before
8932 doing the substitution.
8933 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
8935 2014-02-20 Martin Jambor <mjambor@suse.cz>
8938 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
8939 info when checking whether lattices are bottom.
8941 2014-02-20 Richard Biener <rguenther@suse.de>
8944 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
8947 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
8950 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
8951 parameter specifying the scaling.
8952 (inline_call): Update.
8953 (want_inline_recursively): Guard division by zero.
8954 (recursive_inlining): Update.
8955 * ipa-inline.h (clone_inlined_nodes): Update.
8957 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8960 * config/i386/i386.c (classify_argument): Pass structures of size
8961 64 bytes or less in register.
8963 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
8964 Kirill Yukhin <kirill.yukhin@intel.com>
8966 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
8967 (_mm_rcp28_round_ss): Ditto.
8968 (_mm_rsqrt28_round_sd): Ditto.
8969 (_mm_rsqrt28_round_ss): Ditto.
8970 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
8971 (_mm_rcp14_round_ss): Ditto.
8972 (_mm_rsqrt14_round_sd): Ditto.
8973 (_mm_rsqrt14_round_ss): Ditto.
8974 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
8975 the first input operand, get rid of match_dup.
8976 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
8978 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8980 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
8981 operand as the first input operand, set type attribute.
8982 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8984 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
8985 operand as the first input operand, set type attribute.
8987 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8989 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
8992 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
8995 * config/i386/i386.c (construct_container): Remove TFmode check
8996 for X86_64_INTEGER_CLASS.
8998 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9001 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9002 only when -Wpsabi is enabled.
9004 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9007 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9008 passing arrays in registers are the same as for structs, so remove the
9009 special case for them.
9011 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9013 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9014 destination type, extract only the valid bits if the source type is not
9015 integral and has a different mode.
9017 2014-02-19 Richard Biener <rguenther@suse.de>
9020 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9023 2014-02-19 Richard Biener <rguenther@suse.de>
9026 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9027 (ipa_modify_call_arguments): Emit an argument load explicitely and
9028 preserve virtual SSA form there and for the replacement call.
9029 Do not update SSA form nor free dominance info.
9031 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9033 * ipa.c (function_and_variable_visibility): Also clear WEAK
9034 flag when disolving COMDAT_GROUP.
9036 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9038 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9039 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9041 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9042 do more sanity checks.
9043 (detect_type_change): Return true when giving up early.
9044 (compute_complex_assign_jump_func): Fix type parameter of
9045 ipa_set_ancestor_jf.
9046 (compute_complex_ancestor_jump_func): Likewise.
9047 (update_jump_functions_after_inlining): Fix updating of
9049 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9051 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9053 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9054 inline clones when edge disappears.
9056 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9059 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9060 Split 64-bit moves into 2 patterns. Do not allow the use of
9061 direct move for TDmode in little endian, since the decimal value
9062 has little endian bytes within a word, but the 64-bit pieces are
9063 ordered in a big endian fashion, and normal subreg's of TDmode are
9065 (mov<mode>_64bit_dm): Likewise.
9066 (movtd_64bit_nodm): Likewise.
9068 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9070 PR tree-optimization/60174
9071 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9072 statement of an SSA_NAME that occurs in an abnormal PHI node.
9074 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9077 * final.c (SEEN_BB): Remove.
9078 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9079 (final_scan_insn): Don't force_source_line on second
9080 NOTE_INSN_BASIC_BLOCK.
9082 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9085 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9086 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9087 (type_natural_mode): Warn ABI change when %zmm register is not
9088 available for AVX512F vector value passing.
9090 2014-02-18 Kai Tietz <ktietz@redhat.com>
9093 * config/i386/i386.c (ix86_expand_prologue): Use value in
9094 rax register as displacement when restoring %r10 or %rax.
9095 Fix wrong offset when restoring both registers.
9097 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9099 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9100 assertion with conditional return.
9102 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9103 Uros Bizjak <ubizjak@gmail.com>
9106 * config/i386/driver-i386.c (host_detect_local_cpu): If
9107 YMM state is not saved by the OS, also clear has_f16c. Move
9108 CPUID 0x80000001 handling before YMM state saving checking.
9110 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9112 PR rtl-optimization/58960
9113 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9114 factored out from ...
9115 (sched_init): ... here.
9116 (free_global_sched_pressure_data): New, factored out from ...
9117 (sched_finish): ... here.
9118 * sched-int.h (free_global_sched_pressure_data): Declare.
9119 * sched-rgn.c (nr_regions_initial): New static global.
9120 (haifa_find_rgns): Initialize it.
9121 (schedule_region): Disable sched-pressure for the newly
9124 2014-02-17 Richard Biener <rguenther@suse.de>
9126 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9127 release SSA defs of pattern stmts.
9129 2014-02-17 Richard Biener <rguenther@suse.de>
9131 * tree-inline.c (expand_call_inline): Release the virtual
9132 operand defined by the call we are about to inline.
9134 2014-02-17 Richard Biener <rguenther@suse.de>
9136 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9138 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9139 Ilya Tocar <ilya.tocar@intel.com>
9141 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9142 arguments order in builtin.
9143 (_mm512_permutexvar_epi64): Ditto.
9144 (_mm512_mask_permutexvar_epi64): Ditto
9145 (_mm512_maskz_permutexvar_epi32): Ditto
9146 (_mm512_permutexvar_epi32): Ditto
9147 (_mm512_mask_permutexvar_epi32): Ditto
9149 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9151 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9152 (p8_vmrgow): Likewise.
9154 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9156 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9159 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9162 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9163 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9164 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9165 using direct move instructions on ISA 2.07. Also adjust
9166 instruction length for 64-bit.
9167 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9168 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9170 2014-02-15 Alan Modra <amodra@gmail.com>
9174 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9175 find_replacement on parts of insn rtl that might be reloaded.
9177 2014-02-15 Richard Biener <rguenther@suse.de>
9179 PR tree-optimization/60183
9180 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9181 (tree_ssa_phiprop): Calculate and free post-dominators.
9183 2014-02-14 Jeff Law <law@redhat.com>
9185 PR rtl-optimization/60131
9186 * ree.c (get_extended_src_reg): New function.
9187 (combine_reaching_defs): Use it rather than assuming location of REG.
9188 (find_and_remove_re): Verify first operand of extension is
9189 a REG before adding the insns to the copy list.
9191 2014-02-14 Roland McGrath <mcgrathr@google.com>
9193 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9194 * configure: Regenerated.
9195 * config.in: Regenerated.
9196 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9197 instead of ASM_SHORT.
9199 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9200 Richard Earnshaw <rearnsha@arm.com>
9202 PR rtl-optimization/59535
9203 * lra-constraints.c (process_alt_operands): Encourage alternative
9204 when unassigned pseudo class is superset of the alternative class.
9205 (inherit_reload_reg): Don't inherit when optimizing for code size.
9206 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9207 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9208 modes not less than 4 for Thumb1.
9210 2014-02-14 Kyle McMartin <kyle@redhat.com>
9213 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9215 2014-02-14 Richard Biener <rguenther@suse.de>
9217 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9218 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9219 Do not strip INDIRECT_REFs.
9221 2014-02-14 Richard Biener <rguenther@suse.de>
9224 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9225 DECL_FUNCTION_SPECIFIC_TARGET.
9226 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9227 * tree-streamer-out.c (pack_ts_target_option): Remove.
9228 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9229 (write_ts_function_decl_tree_pointers): Do not stream
9230 DECL_FUNCTION_SPECIFIC_TARGET.
9231 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9232 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9233 (lto_input_ts_function_decl_tree_pointers): Do not stream
9234 DECL_FUNCTION_SPECIFIC_TARGET.
9236 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9238 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9239 (get_initial_def_for_induction, vectorizable_induction): Ignore
9240 debug stmts when looking for exit_phi.
9241 (vectorizable_live_operation): Fix up condition.
9243 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9245 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9246 nreverse() because it changes the content of original tree list.
9248 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9250 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9251 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9253 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9255 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9256 GNU coding standards.
9258 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9261 * dwarf2out.c (gen_subprogram_die): Don't call
9262 add_calling_convention_attribute if subr_die is old_die.
9264 2014-02-13 Sharad Singhai <singhai@google.com>
9266 * doc/optinfo.texi: Fix order of nodes.
9268 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9270 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9271 operands[2], not operands[3].
9273 2014-02-13 Richard Biener <rguenther@suse.de>
9276 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9277 mention the possibility of an in-tree build.
9278 (CLooG): Update recommended version to 0.18.1, mention the
9279 possibility of an in-tree build and clarify that the ISL
9280 bundled with CLooG does not work.
9282 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9285 * expr.c (compress_float_constant): If x is a hard register,
9286 extend into a pseudo and then move to x.
9288 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9290 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9291 caused by bad second argument to warning_at() with -mhotpatch and
9292 nested functions (e.g. with gfortran).
9294 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9296 * opts.c (option_name): Remove "enabled by default" rider.
9298 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9300 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9302 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9303 Uros Bizjak <ubizjak@gmail.com>
9306 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9307 * configure: Regenerated.
9309 2014-02-12 Richard Biener <rguenther@suse.de>
9311 * vec.c (vec_prefix::calculate_allocation): Move as
9312 inline variant to vec.h.
9313 (vec_prefix::calculate_allocation_1): New out-of-line version.
9314 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9315 (vec_prefix::m_has_auto_buf): Rename to ...
9316 (vec_prefix::m_using_auto_storage): ... this.
9317 (vec_prefix::calculate_allocation): Inline the easy cases
9318 and dispatch to calculate_allocation_1 which doesn't need the
9320 (va_heap::reserve): Use gcc_checking_assert.
9321 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9322 m_using_auto_storage.
9323 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9325 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9326 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9327 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9329 2014-02-12 Richard Biener <rguenther@suse.de>
9331 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9332 when we found a dependence.
9334 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9336 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9338 (maybe_fold_stmt): ... into this new function.
9339 * omp-low.c (lower_omp): Update comment.
9341 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9344 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9347 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9349 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9350 identifiers in comments.
9351 (cortexa53_extra_costs): Likewise.
9352 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9353 (cortexa7_extra_costs): Likewise.
9354 (cortexa12_extra_costs): Likewise.
9355 (cortexa15_extra_costs): Likewise.
9356 (v7m_extra_costs): Likewise.
9358 2014-02-12 Richard Biener <rguenther@suse.de>
9361 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9362 of posix_memalign being successful.
9363 (lower_stmt): Restrict lowering of posix_memalign to when
9364 -ftree-bit-ccp is enabled.
9366 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9368 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9370 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9372 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9374 PR rtl-optimization/60116
9375 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9376 other_insn once the combination has been validated.
9378 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9381 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9383 * ipa-devirt.c: Include demangle.h
9384 (odr_violation_reported): New static variable.
9385 (add_type_duplicate): Update odr_violations.
9386 (maybe_record_node): Add completep parameter; update it.
9387 (record_target_from_binfo): Add COMPLETEP parameter;
9388 update it as needed.
9389 (possible_polymorphic_call_targets_1): Likewise.
9390 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9391 rename FINAL to COMPLETE.
9392 (record_targets_from_bases): Sanity check we found the binfo;
9393 fix COMPLETEP updating.
9394 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9395 parameter, fix computing of COMPLETEP.
9396 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9397 at LTO time do demangling.
9398 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9399 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9401 (gimple_get_virt_method_for_binfo): Likewise.
9402 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9403 gimple_get_virt_method_for_vtable): Update prototypes.
9405 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9408 * genautomata.c (add_presence_absence): Fix typo with
9409 {final_}presence_list.
9411 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9414 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9415 for VSX/Altivec vectors that land in GPR registers.
9417 2014-02-11 Richard Henderson <rth@redhat.com>
9418 Jakub Jelinek <jakub@redhat.com>
9421 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9422 around drhs if type conversion to lacc->type is not useless.
9424 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9426 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9428 (cortex-a57.cortex-a53): Likewise.
9429 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9431 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9433 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9436 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9438 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9439 add_options_for_arm_vfp3.
9441 2014-02-11 Jeff Law <law@redhat.com>
9444 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9445 object with an undesirable mode.
9447 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9450 * config/i386/sol2-9.h: New file.
9451 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9452 *-*-solaris2.9*): Use it.
9454 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9456 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9457 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9459 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9461 * config/microblaze/microblaze.c: Extend mcpu version format
9463 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9465 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9467 2014-02-10 Richard Henderson <rth@redhat.com>
9470 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9471 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9472 ms-abi vs -mno-accumulate-outgoing-args.
9473 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9474 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9477 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9480 * cfgexpand.c (expand_asm_operands): Attach source location to
9481 ASM_INPUT rtx objects.
9482 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9484 2014-02-10 Nick Clifton <nickc@redhat.com>
9486 * config/mn10300/mn10300.c (popcount): New function.
9487 (mn10300_expand_prologue): Include saved registers in stack usage
9490 2014-02-10 Jeff Law <law@redhat.com>
9493 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9494 when changing the SET_DEST of a prior insn to avoid an input reload.
9496 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9498 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9499 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9500 -mcall-openbsd, or -mcall-linux.
9501 (CC1_ENDIAN_BIG_SPEC): Remove.
9502 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9503 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9504 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9505 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9506 and %cc1_endian_default.
9507 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9509 2014-02-10 Richard Biener <rguenther@suse.de>
9511 PR tree-optimization/60115
9512 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9513 MEM_REF handling. Properly verify that the accesses are not
9514 out of the objects bound.
9516 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9518 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9521 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9523 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9524 proper constants and fix formatting.
9525 (possible_polymorphic_call_targets): Fix formatting.
9527 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9528 Ilya Tocar <ilya.tocar@intel.com>
9530 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9531 (_mm512_loadu_epi32): Renamed into...
9532 (_mm512_loadu_si512): This.
9533 (_mm512_storeu_epi32): Renamed into...
9534 (_mm512_storeu_si512): This.
9535 (_mm512_maskz_ceil_ps): Removed.
9536 (_mm512_maskz_ceil_pd): Ditto.
9537 (_mm512_maskz_floor_ps): Ditto.
9538 (_mm512_maskz_floor_pd): Ditto.
9539 (_mm512_floor_round_ps): Ditto.
9540 (_mm512_floor_round_pd): Ditto.
9541 (_mm512_ceil_round_ps): Ditto.
9542 (_mm512_ceil_round_pd): Ditto.
9543 (_mm512_mask_floor_round_ps): Ditto.
9544 (_mm512_mask_floor_round_pd): Ditto.
9545 (_mm512_mask_ceil_round_ps): Ditto.
9546 (_mm512_mask_ceil_round_pd): Ditto.
9547 (_mm512_maskz_floor_round_ps): Ditto.
9548 (_mm512_maskz_floor_round_pd): Ditto.
9549 (_mm512_maskz_ceil_round_ps): Ditto.
9550 (_mm512_maskz_ceil_round_pd): Ditto.
9551 (_mm512_expand_pd): Ditto.
9552 (_mm512_expand_ps): Ditto.
9553 * config/i386/i386.c (ix86_builtins): Remove
9554 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9555 (bdesc_args): Ditto.
9556 * config/i386/predicates.md (const1256_operand): New.
9557 (const_1_to_2_operand): Ditto.
9558 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9559 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9560 (*avx512pf_gatherpf<mode>sf): Ditto.
9561 (avx512pf_gatherpf<mode>df): Ditto.
9562 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9563 (*avx512pf_gatherpf<mode>df): Ditto.
9564 (avx512pf_scatterpf<mode>sf): Ditto.
9565 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9566 (*avx512pf_scatterpf<mode>sf): Ditto.
9567 (avx512pf_scatterpf<mode>df): Ditto.
9568 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9569 (*avx512pf_scatterpf<mode>df): Ditto.
9570 (avx512f_expand<mode>): Removed.
9571 (<shift_insn><mode>3<mask_name>): Change predicate type.
9573 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9575 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9576 not at the end of datarefs vector use ordered_remove to avoid
9577 reordering datarefs vector.
9580 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9581 mark local addressable non-static vars as GOVD_PRIVATE
9582 instead of GOVD_LOCAL.
9583 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9584 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9588 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9589 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9590 assume_aligned or alloc_align attributes.
9591 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9592 arguments. Handle also assume_aligned and alloc_align attributes.
9593 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9594 calls to functions with assume_aligned or alloc_align attributes.
9595 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9597 2014-02-08 Terry Guo <terry.guo@arm.com>
9599 * doc/invoke.texi: Document ARM -march=armv7e-m.
9601 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9603 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9604 flag on __cilkrts_rethrow builtin.
9607 * ipa-cp.c (determine_versionability): Fail at -O0
9608 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9609 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9612 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9615 * tree-inline.c (copy_forbidden): Fail for
9616 __attribute__((optimize (0))) functions.
9618 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9620 * varpool.c: Include pointer-set.h.
9621 (varpool_remove_unreferenced_decls): Variables in other partitions
9622 will not be output; be however careful to not lose information
9625 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9627 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9628 lookup in the vtable constructor.
9630 2014-02-07 Jeff Law <law@redhat.com>
9633 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9634 define_insn_and_split.
9636 * ipa-inline.c (inline_small_functions): Fix typos.
9638 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9640 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9641 (s390_can_use_return_insn): Declare.
9642 * config/s390/s390.h (EPILOGUE_USES): Define.
9643 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9645 (s390_chunkify_start): Handle return JUMP_LABELs.
9646 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9647 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9648 (s390_can_use_return_insn): New functions.
9649 (s390_fix_long_loop_prediction): Handle conditional returns.
9650 (TARGET_SET_UP_BY_PROLOGUE): Define.
9651 * config/s390/s390.md (ANY_RETURN): New code iterator.
9652 (*creturn, *csimple_return, return, simple_return): New patterns.
9654 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9656 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9657 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9658 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9659 REG_CFA_RESTORE list when deciding not to restore a register.
9661 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9663 * config/s390/s390.c: Include tree-pass.h and context.h.
9664 (s390_early_mach): New function, split out from...
9665 (s390_emit_prologue): ...here.
9666 (pass_data_s390_early_mach): New pass structure.
9667 (pass_s390_early_mach): New class.
9668 (s390_option_override): Create and register early_mach pass.
9669 Move to end of file.
9671 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9673 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9674 to match for the exit block.
9676 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9678 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9679 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9680 Reject misaligned operands.
9682 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9684 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9686 2014-02-07 Richard Biener <rguenther@suse.de>
9689 * gimple-low.c (lower_builtin_posix_memalign): New function.
9690 (lower_stmt): Call it to lower posix_memalign in a way
9691 to make alignment info accessible.
9693 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9696 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9697 __builtin_setjmp_receiver.
9699 2014-02-07 Richard Biener <rguenther@suse.de>
9702 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9703 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9704 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9705 Handle BUILT_IN_POSIX_MEMALIGN.
9706 (find_func_clobbers): Likewise.
9707 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9708 (call_may_clobber_ref_p_1): Likewise.
9710 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9713 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9716 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9719 * lto-cgraph.c (lto_output_node): Use
9720 symtab_get_symbol_partitioning_class.
9721 (lto_output_varpool_node): likewise.
9722 (symtab_get_symbol_partitioning_class): Move here from
9724 * cgraph.h (symbol_partitioning_class): Likewise.
9725 (symtab_get_symbol_partitioning_class): Declare.
9727 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9729 * ggc.h (ggc_internal_cleared_alloc): New macro.
9730 * vec.h (vec_safe_copy): Handle memory stats.
9731 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9732 * target-globals.c (save_target_globals): Likewise.
9734 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9737 * expr.c (emit_move_resolve_push): Export; be bit more selective
9738 on when to clear alias set.
9739 * expr.h (emit_move_resolve_push): Declare.
9740 * function.h (struct function): Add tail_call_marked.
9741 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9742 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9743 * config/i386/i386.md (TImode move expander): De not call
9745 (FP push expanders): Preserve memory attributes.
9746 * config/i386/sse.md (push<mode>1): Remove.
9747 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9748 (ix86_expand_push): Remove.
9749 * config/i386/mmx.md (push<mode>1): Remove.
9751 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9753 PR rtl-optimization/60030
9754 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9755 lopart with paradoxical subreg before shifting it up by hprec.
9757 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9759 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9760 Remove extra newline at end of file.
9761 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9762 (arm_issue_rate): Handle cortexa57.
9763 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9764 (cortex-a57.cortex-a53): Likewise.
9766 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9769 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9770 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9772 (arm_expand_prologue): Adjust all callers.
9773 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9774 info, registers also at the lowest numbered registers side. Use
9775 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9779 * var-tracking.c (adjust_mems): Before adding a SET to
9780 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9782 2014-02-06 Alan Modra <amodra@gmail.com>
9785 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9786 change SDmode to DDmode when lra_in_progress.
9788 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9791 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9792 free_data_ref on the dr first, and before goto again also set dr
9793 to the next dr. For simd_lane_access, free old datarefs[i] before
9794 overwriting it. For get_vectype_for_scalar_type failure, don't
9795 free_data_ref if simd_lane_access.
9797 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9800 * tree.h (opts_for_fn): New inline function.
9801 (opt_for_fn): Define.
9802 * config/i386/i386.c (ix86_function_regparm): Use
9803 opt_for_fn (decl, optimize) instead of optimize.
9805 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9807 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9808 for SYMBOL_REF in large memory model.
9810 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9812 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9814 (cortex-a57): Likewise.
9815 (cortex-a57.cortex-a53): Likewise.
9817 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9818 Kugan Vivekanandarajah <kuganv@linaro.org>
9820 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9822 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9824 2014-02-06 Richard Biener <rguenther@suse.de>
9826 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
9827 set_loop_copy and initialize_original_copy_tables.
9829 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
9831 * config/aarch64/aarch64-simd.md
9832 (aarch64_ashr_simddi): Change QI to SI.
9834 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9835 Jakub Jelinek <jakub@redhat.com>
9838 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
9841 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9843 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
9844 CODE_FOR_altivec_vpku[hw]um to
9845 CODE_FOR_altivec_vpku[hw]um_direct.
9846 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
9847 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
9848 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
9849 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
9851 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9853 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
9854 generation for -maltivec=be.
9855 (altivec_vsumsws): Simplify redundant test.
9857 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9859 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
9860 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
9861 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
9862 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
9863 gen_altivec_vpkuwum.
9864 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
9866 (altivec_vpks<VI_char>ss): Likewise.
9867 (altivec_vpks<VI_char>us): Likewise.
9868 (altivec_vpku<VI_char>us): Likewise.
9869 (altivec_vpku<VI_char>um): Likewise.
9870 (altivec_vpku<VI_char>um_direct): New (copy of
9871 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
9873 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
9874 target is little endian and -maltivec=be is not specified.
9875 (*altivec_vupkhs<VU_char>_direct): New (copy of
9876 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
9877 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
9878 target is little endian and -maltivec=be is not specified.
9879 (*altivec_vupkls<VU_char>_direct): New (copy of
9880 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
9881 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
9882 little endian and -maltivec=be is not specified.
9883 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
9884 little endian and -maltivec=be is not specified.
9886 2014-02-05 Richard Henderson <rth@redhat.com>
9889 * combine-stack-adj.c: Revert r206943.
9890 * sched-int.h (struct deps_desc): Add last_args_size.
9891 * sched-deps.c (init_deps): Initialize it.
9892 (sched_analyze_insn): Add OUTPUT dependencies between insns that
9893 contain REG_ARGS_SIZE notes.
9895 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
9897 * lto-cgraph.c (asm_nodes_output): Make global.
9898 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
9899 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
9900 (driver_handle_option): Handle OPT_fwpa.
9902 2014-02-05 Jakub Jelinek <jakub@redhat.com>
9905 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
9906 a comment typo and formatting issue. If odr_hash hasn't been
9907 created, return vNULL and set *completep to false.
9910 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
9911 bb with no successors.
9913 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
9916 * doc/invoke.texi (-march): Clarify documentation for ARM.
9920 2014-02-05 Richard Biener <rguenther@suse.de>
9922 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
9923 when not vectorizing because of too many alias checks.
9924 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
9925 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
9927 2014-02-05 Nick Clifton <nickc@redhat.com>
9929 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
9930 accept extended registers in any mode when compiling for the MN10300.
9932 2014-02-05 Yury Gribov <y.gribov@samsung.com>
9934 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
9935 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
9936 sanitization attributes.
9937 (can_inline_edge_p): Likewise.
9938 (sanitize_attrs_match_for_inline_p): New function.
9940 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9942 * ipa-prop.c (detect_type_change): Shor circuit testing of
9943 type changes on THIS pointer.
9945 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
9948 * config/pa/pa.c (legitimize_tls_address): Return original address
9949 if not passed a SYMBOL_REF rtx.
9950 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
9952 (pa_emit_move_sequence): Simplify TLS source operands.
9953 (pa_legitimate_constant_p): Reject all TLS constants.
9954 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
9955 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
9957 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9959 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
9960 groups when we know they are controlled by LTO.
9961 * varasm.c (default_binds_local_p_1): If object is in other partition,
9962 it will be resolved locally.
9964 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9966 * config/host-linux.c (linux_gt_pch_use_address): Don't
9967 use SSIZE_MAX because it is not always defined.
9969 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
9972 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
9973 threshold for pseudo splitting.
9974 (update_ebb_live_info): Process call argument hard registers and
9975 hard registers from insn definition too.
9976 (max_small_class_regs_num): New constant.
9977 (inherit_in_ebb): Update live hard regs through EBBs. Update
9978 reloads_num only for small register classes. Don't split for
9981 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
9984 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
9987 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
9989 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
9992 2014-02-04 Marek Polacek <polacek@redhat.com>
9994 * gdbinit.in (pel): Define.
9996 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
9998 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10001 2014-02-04 Richard Biener <rguenther@suse.de>
10004 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10005 in function context local.
10006 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10007 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10008 similar to LTO_imported_decl_ref.
10010 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10012 PR tree-optimization/60002
10013 * cgraphclones.c (build_function_decl_skip_args): Clear
10014 DECL_LANG_SPECIFIC.
10016 PR tree-optimization/60023
10017 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10018 false to gsi_replace.
10019 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10020 has been in some EH region and vec_stmt could throw, add
10021 vec_stmt into the same EH region.
10022 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10023 has no lhs, ignore it.
10024 * internal-fn.c (expand_MASK_LOAD): Likewise.
10027 * tree-inline.c (copy_forbidden): Fail for
10028 __attribute__((optimize (0))) functions.
10031 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10032 is set, copy one less argument.
10033 (expand_simd_clones): Don't subtract clone_info->inbranch
10034 from simd_clone_struct_alloc argument.
10036 PR rtl-optimization/57915
10037 * recog.c (simplify_while_replacing): If all unary/binary/relational
10038 operation arguments are constant, attempt to simplify those.
10040 PR middle-end/59261
10041 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10042 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10044 2014-02-04 Richard Biener <rguenther@suse.de>
10046 PR tree-optimization/60012
10047 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10048 TBAA disambiguation to all DDRs.
10050 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10053 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10054 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10056 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10059 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10061 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10063 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10064 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10066 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10069 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10070 to figure out targets of polymorphic calls with known decl.
10071 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10072 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10073 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10074 (get_polymorphic_call_info): ... here.
10075 (get_polymorphic_call_info_from_invariant): New function.
10077 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10079 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10080 lookup via vtable pointer; check for type consistency
10081 and turn inconsitent facts into UNREACHABLE.
10082 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10083 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10084 type inconsistent querries; return UNREACHABLE instead.
10086 2014-02-03 Richard Henderson <rth@twiddle.net>
10089 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10090 already processed this node.
10091 (normalize_one_pred_1): Pass along mark_set.
10092 (normalize_one_pred): Create and destroy a pointer_set_t.
10093 (normalize_one_pred_chain): Likewise.
10095 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10097 PR gcov-profile/58602
10098 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10100 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10103 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10104 -fno-devirtualize; try to devirtualize by the knowledge of
10105 virtual table pointer given by aggregate propagation.
10106 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10107 (ipa_print_node_jump_functions): Dump also offset that
10108 is relevant for polymorphic calls.
10109 (determine_known_aggregate_parts): Add arg_type parameter; use it
10110 instead of determining the type from pointer type.
10111 (ipa_compute_jump_functions_for_edge): Update call of
10112 determine_known_aggregate_parts.
10113 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10114 (gimple_get_virt_method_for_binfo): ... here; simplify using
10115 vtable_pointer_value_to_vtable.
10116 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10117 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10118 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10119 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10121 (vtable_pointer_value_to_binfo): ... here.
10122 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10124 2014-02-03 Teresa Johnson <tejohnson@google.com>
10126 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10127 redef of outer loop index variable.
10129 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10133 * doc/extend.texi (Function Attributes): Typo.
10135 2014-02-03 Cong Hou <congh@google.com>
10137 PR tree-optimization/60000
10138 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10139 if the vectorized statement is a store. A store statement can only
10140 appear at the end of pattern statements.
10142 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10144 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10145 (ix86_option_override_internal): Default long double to 64-bit for
10146 32-bit Bionic and to 128-bit for 64-bit Bionic.
10148 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10149 TARGET_LONG_DOUBLE_128 is true.
10150 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10152 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10153 (mlong-double-64): Negate -mlong-double-128.
10154 (mlong-double-128): New option.
10156 * config/i386/i386-c.c (ix86_target_macros): Define
10157 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10159 * doc/invoke.texi: Document -mlong-double-128.
10161 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10163 PR rtl-optimization/60024
10164 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10166 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10168 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10170 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10172 PR rtl-optimization/57662
10173 * sel-sched.c (code_motion_path_driver): Do not mark already not
10174 existing blocks in the visiting bitmap.
10176 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10178 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10179 on the insn being emitted.
10181 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10182 Will Deacon <will.deacon@arm.com>
10184 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10186 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10188 * config/arm/arm-tables.opt: Regenerate.
10190 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10192 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10193 for vector types other than V16QImode.
10194 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10195 define_expand, and call altivec_expand_vec_perm_le when producing
10196 code with little endian element order.
10197 (*altivec_vperm_<mode>_internal): New insn having previous
10198 behavior of altivec_vperm_<mode>.
10199 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10200 altivec_expand_vec_perm_le when producing code with little endian
10202 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10203 behavior of altivec_vperm_<mode>_uns.
10205 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10207 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10208 (altivec_vsumsws): Add handling for -maltivec=be with a little
10210 (altivec_vsumsws_direct): New.
10211 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10212 gen_altivec_vsumsws.
10214 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10216 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10217 vtable_pointer_value_to_binfo): New functions.
10218 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10219 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10221 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10223 * config/nios2/nios2.md (load_got_register): Initialize GOT
10224 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10225 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10227 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10229 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10230 preserverd by passthrough, do not propagate the type.
10232 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10234 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10235 (mips_atomic_assign_expand_fenv): New function.
10236 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10238 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10240 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10241 (__builtin_mips_set_fcsr): Likewise.
10242 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10243 MIPS_USI_FTYPE_VOID.
10244 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10245 (mips16_expand_set_fcsr): Likewise.
10246 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10247 (mips16_set_fcsr_stub): Likewise.
10248 (mips16_get_fcsr_one_only_stub): New class.
10249 (mips16_set_fcsr_one_only_stub): Likewise.
10250 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10251 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10252 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10253 (hard_float): New availability predicate.
10254 (mips_builtins): Add get_fcsr and set_fcsr.
10255 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10256 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10257 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10258 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10259 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10262 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10264 * config/mips/mips.c (mips_one_only_stub): New class.
10265 (mips_need_mips16_rdhwr_p): Replace with...
10266 (mips16_rdhwr_stub): ...this new variable.
10267 (mips16_stub_call_address): New function.
10268 (mips16_rdhwr_one_only_stub): New class.
10269 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10270 (mips_output_mips16_rdhwr): Delete.
10271 (mips_finish_stub): New function.
10272 (mips_code_end): Use it to handle rdhwr stubs.
10274 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10277 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10278 when calculating size of integer atomic types.
10280 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10282 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10284 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10286 PR tree-optimization/60003
10287 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10288 * profile.c (branch_prob): Use gimple_call_builtin_p
10289 to check for BUILT_IN_SETJMP_RECEIVER.
10290 * tree-inline.c (copy_bb): Call notice_special_calls.
10292 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10295 * lra-constraints.c (process_alt_operands): Update reload_sum only
10298 2014-01-31 Richard Henderson <rth@redhat.com>
10300 PR middle-end/60004
10301 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10302 until after else_eh is processed.
10304 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10306 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10307 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10308 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10309 in smmintrin.h, remove them.
10310 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10311 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10312 * config/i386/i386.md (ROUND_SAE): Fix value.
10313 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10314 (const48_operand): New.
10315 * config/i386/subst.md (round), (round_expand): Use
10316 const_4_or_8_to_11_operand.
10317 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10319 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10321 * config/i386/constraints.md (Yk): Swap meaning with k.
10322 * config/i386/i386.md (movhi_internal): Change Yk to k.
10323 (movqi_internal): Ditto.
10324 (*k<logic><mode>): Ditto.
10327 (kandn<mode>): Ditto.
10328 (*<code>hi_1): Ditto.
10329 (*<code>qi_1): Ditto.
10330 (kxnor<mode>): Ditto.
10331 (kortestzhi): Ditto.
10332 (kortestchi): Ditto.
10334 (*one_cmplhi2_1): Ditto.
10335 (*one_cmplqi2_1): Ditto.
10336 * config/i386/sse.md (): Change k to Yk.
10337 (avx512f_load<mode>_mask): Ditto.
10338 (avx512f_blendm<mode>): Ditto.
10339 (avx512f_store<mode>_mask): Ditto.
10340 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10341 (avx512f_storedqu<mode>_mask): Ditto.
10342 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10344 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10345 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10346 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10347 (avx512f_maskcmp<mode>3): Ditto.
10348 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10349 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10350 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10351 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10352 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10353 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10354 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10355 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10356 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10357 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10358 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10359 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10360 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10361 (vec_extract_lo_<mode>_maskm): Ditto.
10362 (vec_extract_hi_<mode>_maskm): Ditto.
10363 (avx512f_vternlog<mode>_mask): Ditto.
10364 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10365 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10366 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10367 (avx512f_<code>v8div16qi2_mask): Ditto.
10368 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10369 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10370 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10371 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10372 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10373 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10374 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10375 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10376 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10377 (avx512cd_maskb_vec_dupv8di): Ditto.
10378 (avx512cd_maskw_vec_dupv16si): Ditto.
10379 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10380 (avx512f_vpermi2var<mode>3_mask): Ditto.
10381 (avx512f_vpermi2var<mode>3_mask): Ditto.
10382 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10383 (*avx512f_gathersi<mode>): Ditto.
10384 (*avx512f_gathersi<mode>_2): Ditto.
10385 (*avx512f_gatherdi<mode>): Ditto.
10386 (*avx512f_gatherdi<mode>_2): Ditto.
10387 (*avx512f_scattersi<mode>): Ditto.
10388 (*avx512f_scatterdi<mode>): Ditto.
10389 (avx512f_compress<mode>_mask): Ditto.
10390 (avx512f_compressstore<mode>_mask): Ditto.
10391 (avx512f_expand<mode>_mask): Ditto.
10392 * config/i386/subst.md (mask): Change k to Yk.
10393 (mask_scalar_merge): Ditto.
10396 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10398 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10400 2014-01-31 Richard Biener <rguenther@suse.de>
10402 PR middle-end/59990
10403 * builtins.c (fold_builtin_memory_op): Make sure to not
10404 use a floating-point mode or a boolean or enumeral type for
10405 the copy operation.
10407 2014-01-30 DJ Delorie <dj@redhat.com>
10409 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10410 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10411 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10412 whenever main() has an epilogue.
10414 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10416 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10417 unused variable "field".
10418 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10419 (vsx_mergeh_<mode>): Likewise.
10420 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10421 (altivec_vmrghh): Likewise.
10422 (altivec_vmrghw): Likewise.
10423 (altivec_vmrglb): Likewise.
10424 (altivec_vmrglh): Likewise.
10425 (altivec_vmrglw): Likewise.
10426 (altivec_vspltb): Add missing uses.
10427 (altivec_vsplth): Likewise.
10428 (altivec_vspltw): Likewise.
10429 (altivec_vspltsf): Likewise.
10431 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10434 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10435 frame related instructions.
10437 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10439 PR rtl-optimization/59959
10440 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10441 any reload of register whose subreg is invalid.
10443 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10445 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10446 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10447 Add missing return type - void.
10449 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10451 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10452 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10453 remove element index adjustment for endian (now handled in vsx.md
10455 (altivec_expand_vec_perm_const): Use
10456 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10457 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10458 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10459 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10460 define_expand and a new define_insn *altivec_vspltb_internal;
10461 adjust for -maltivec=be on a little endian target.
10462 (altivec_vspltb_direct): New.
10463 (altivec_vsplth): Divide into a define_expand and a new
10464 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10465 little endian target.
10466 (altivec_vsplth_direct): New.
10467 (altivec_vspltw): Divide into a define_expand and a new
10468 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10469 little endian target.
10470 (altivec_vspltw_direct): New.
10471 (altivec_vspltsf): Divide into a define_expand and a new
10472 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10473 a little endian target.
10475 2014-01-30 Richard Biener <rguenther@suse.de>
10477 PR tree-optimization/59993
10478 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10479 can propagate form the earlier stmt and avoid the transform
10480 when the intermediate result is needed.
10482 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10484 * README.Portability: Fix typo.
10486 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10488 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10489 comparison_operator with ordered_comparison_operator.
10491 2014-01-30 Nick Clifton <nickc@redhat.com>
10493 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10494 Rename to mn10300_store_multiple_regs.
10495 * config/mn10300/mn10300.c: Likewise.
10496 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10497 store_multiple_regs.
10498 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10499 Call mn10300_store_multiple_regs.
10501 2014-01-30 Nick Clifton <nickc@redhat.com>
10502 DJ Delorie <dj@redhat.com>
10504 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10505 %fp 2 to keep registers after it properly word-aligned.
10506 (rl78_alloc_physical_registers_umul): Handle the case where both
10507 input operands are the same.
10509 2014-01-30 Richard Biener <rguenther@suse.de>
10511 PR tree-optimization/59903
10512 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10515 2014-01-30 Jason Merrill <jason@redhat.com>
10518 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10521 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10523 2014-01-30 Richard Biener <rguenther@suse.de>
10525 PR tree-optimization/59951
10526 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10528 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10531 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10532 SFmode to DFmode case.
10534 2014-01-29 DJ Delorie <dj@redhat.com>
10536 * config/msp430/msp430.opt (-minrt): New.
10537 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10539 (ENDFILE_SPEC): Likewise.
10541 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10543 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10544 (estimate_function_body_sizes): Use it.
10546 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10549 * dwarf2out.c (is_cxx_auto): New.
10550 (is_base_type): Use it.
10551 (gen_type_die_with_usage): Likewise.
10553 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10555 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10556 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10557 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10558 -maltivec=be with LE targets.
10559 (vsx_mergeh_<mode>): Likewise.
10560 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10561 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10562 (altivec_vmrghb): Replace with define_expand and new
10563 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10564 (altivec_vmrghb_direct): New define_insn.
10565 (altivec_vmrghh): Replace with define_expand and new
10566 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10567 (altivec_vmrghh_direct): New define_insn.
10568 (altivec_vmrghw): Replace with define_expand and new
10569 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10570 (altivec_vmrghw_direct): New define_insn.
10571 (*altivec_vmrghsf): Adjust for endianness.
10572 (altivec_vmrglb): Replace with define_expand and new
10573 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10574 (altivec_vmrglb_direct): New define_insn.
10575 (altivec_vmrglh): Replace with define_expand and new
10576 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10577 (altivec_vmrglh_direct): New define_insn.
10578 (altivec_vmrglw): Replace with define_expand and new
10579 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10580 (altivec_vmrglw_direct): New define_insn.
10581 (*altivec_vmrglsf): Adjust for endianness.
10582 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10583 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10584 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10585 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10586 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10587 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10588 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10589 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10591 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10593 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10594 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10597 2014-01-29 Richard Biener <rguenther@suse.de>
10599 PR tree-optimization/58742
10600 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10601 associate_pointerplus_align.
10602 (associate_pointerplus_diff): New function.
10603 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10604 and associate_pointerplus_diff.
10606 2014-01-29 Richard Biener <rguenther@suse.de>
10608 * lto-streamer.h (LTO_major_version): Bump to 3.
10609 (LTO_minor_version): Reset to 0.
10611 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10613 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10614 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10615 (arm_file_start): Generate correct asm header for armv7ve.
10616 * config/arm/bpabi.h: Add multilib support for armv7ve.
10617 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10618 and cortex-a15 to armv7ve.
10619 * config/arm/t-aprofile: Add multilib support for armv7ve.
10620 * doc/invoke.texi: Document -march=armv7ve.
10622 2014-01-29 Richard Biener <rguenther@suse.de>
10624 PR tree-optimization/58742
10625 * tree-ssa-forwprop.c (associate_plusminus): Return true
10626 if we changed sth, defer EH cleanup to ...
10627 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10628 (simplify_mult): New function.
10630 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10632 PR middle-end/59917
10633 PR tree-optimization/59920
10634 * tree.c (build_common_builtin_nodes): Remove
10635 __builtin_setjmp_dispatcher initialization.
10636 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10637 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10638 instead of gsi_after_labels + manually skipping debug stmts.
10639 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10640 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10641 * tree-inline.c (copy_edges_for_bb): Remove
10642 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10643 argument. Ignore computed_goto_p stmts. Don't call
10644 make_abnormal_goto_edges. If a call might need abnormal edges
10645 for non-local gotos, see if it already has an edge to
10646 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10647 with true argument, don't do anything then, otherwise add
10648 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10649 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10651 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10652 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10653 (lower_stmt): Don't set data->calls_builtin_setjmp.
10654 (lower_builtin_setjmp): Adjust comment.
10655 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10656 * tree-cfg.c (found_computed_goto): Remove.
10657 (factor_computed_gotos): Remove.
10658 (make_goto_expr_edges): Return bool, true for computed gotos.
10659 Don't call make_abnormal_goto_edges.
10660 (build_gimple_cfg): Don't set found_computed_goto, don't call
10661 factor_computed_gotos.
10662 (computed_goto_p): No longer static.
10663 (make_blocks): Don't set found_computed_goto.
10664 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10665 (make_edges): If make_goto_expr_edges returns true, push bb
10666 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10667 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10668 vector. Record mapping between bbs and OpenMP regions if there
10669 are any, adjust make_gimple_omp_edges caller. Call
10670 handle_abnormal_edges.
10671 (make_abnormal_goto_edges): Remove.
10672 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10673 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10674 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10675 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10676 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10677 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10678 filling *region also set *region_idx to (*region)->entry->index.
10681 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10682 For REGs set ORIGINAL_REGNO.
10684 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10686 * doc/md.texi: Mention that a target shouldn't implement
10687 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10690 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10692 PR tree-optimization/59594
10693 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10694 a copy of the datarefs vector rather than the vector itself.
10696 2014-01-28 Jason Merrill <jason@redhat.com>
10699 * dwarf2out.c (auto_die): New static.
10700 (gen_type_die_with_usage): Handle C++1y 'auto'.
10701 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10704 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10707 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10708 (SPEC_X32): Likewise.
10709 (SPEC_64): Likewise.
10710 * config/i386/i386.c (ix86_option_override_internal): Turn off
10711 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10713 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10714 * config/i386/i386.h (TARGET_16BIT): New macro.
10715 (TARGET_16BIT_P): Likewise.
10716 * config/i386/i386.opt: Add m16.
10717 * doc/invoke.texi: Document -m16.
10719 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10721 PR preprocessor/59935
10722 * input.c (location_get_source_line): Bail out on when line number
10723 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10725 2014-01-28 Richard Biener <rguenther@suse.de>
10727 PR tree-optimization/58742
10728 * tree-ssa-forwprop.c (associate_plusminus): Handle
10729 pointer subtraction of the form (T)(P + A) - (T)P.
10731 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10733 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10736 2014-01-28 Richard Biener <rguenther@suse.de>
10739 2014-01-28 Richard Biener <rguenther@suse.de>
10741 PR rtl-optimization/45364
10742 PR rtl-optimization/59890
10743 * var-tracking.c (local_get_addr_clear_given_value): Handle
10744 already cleared slot.
10745 (val_reset): Handle not allocated local_get_addr_cache.
10746 (vt_find_locations): Use post-order on the inverted CFG.
10748 2014-01-28 Richard Biener <rguenther@suse.de>
10750 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10752 2014-01-28 Richard Biener <rguenther@suse.de>
10754 PR rtl-optimization/45364
10755 PR rtl-optimization/59890
10756 * var-tracking.c (local_get_addr_clear_given_value): Handle
10757 already cleared slot.
10758 (val_reset): Handle not allocated local_get_addr_cache.
10759 (vt_find_locations): Use post-order on the inverted CFG.
10761 2014-01-28 Alan Modra <amodra@gmail.com>
10763 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10764 * configure.ac <recursive call for build != host>: Define
10765 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10766 and LD_FOR_BUILD too.
10767 * configure: Regenerate.
10769 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10771 * config/i386/i386.c (get_builtin_code_for_version): Separate
10772 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10773 Broadwell from Haswell.
10775 2014-01-27 Steve Ellcey <sellcey@mips.com>
10777 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10778 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10779 * config/mips/mips.c (mips_option_override): Change setting
10781 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10782 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10783 Change from Mask to Var.
10785 2014-01-27 Jeff Law <law@redhat.com>
10787 * ipa-inline.c (inline_small_functions): Fix typo.
10789 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10791 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10792 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10793 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10794 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10795 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10796 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10797 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10798 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10799 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10800 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10801 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10802 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10803 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10804 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10805 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10806 (_mm512_storeu_epi64): Ditto.
10807 (_mm512_cmpge_epi32_mask): Ditto.
10808 (_mm512_cmpge_epu32_mask): Ditto.
10809 (_mm512_cmpge_epi64_mask): Ditto.
10810 (_mm512_cmpge_epu64_mask): Ditto.
10811 (_mm512_cmple_epi32_mask): Ditto.
10812 (_mm512_cmple_epu32_mask): Ditto.
10813 (_mm512_cmple_epi64_mask): Ditto.
10814 (_mm512_cmple_epu64_mask): Ditto.
10815 (_mm512_cmplt_epi32_mask): Ditto.
10816 (_mm512_cmplt_epu32_mask): Ditto.
10817 (_mm512_cmplt_epi64_mask): Ditto.
10818 (_mm512_cmplt_epu64_mask): Ditto.
10819 (_mm512_cmpneq_epi32_mask): Ditto.
10820 (_mm512_cmpneq_epu32_mask): Ditto.
10821 (_mm512_cmpneq_epi64_mask): Ditto.
10822 (_mm512_cmpneq_epu64_mask): Ditto.
10823 (_mm512_expand_pd): Ditto.
10824 (_mm512_expand_ps): Ditto.
10825 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
10826 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
10827 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
10828 * config/i386/i386.c (ix86_builtins): Add
10829 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
10830 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
10831 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
10832 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
10833 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
10834 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
10835 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
10836 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
10837 IX86_BUILTIN_PMOVUSQW512_MEM.
10838 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
10839 __builtin_ia32_pmovsqd512mem_mask,
10840 __builtin_ia32_pmovqd512mem_mask,
10841 __builtin_ia32_pmovusqw512mem_mask,
10842 __builtin_ia32_pmovsqw512mem_mask,
10843 __builtin_ia32_pmovqw512mem_mask,
10844 __builtin_ia32_pmovusdw512mem_mask,
10845 __builtin_ia32_pmovsdw512mem_mask,
10846 __builtin_ia32_pmovdw512mem_mask,
10847 __builtin_ia32_pmovqb512mem_mask,
10848 __builtin_ia32_pmovusqb512mem_mask,
10849 __builtin_ia32_pmovsqb512mem_mask,
10850 __builtin_ia32_pmovusdb512mem_mask,
10851 __builtin_ia32_pmovsdb512mem_mask,
10852 __builtin_ia32_pmovdb512mem_mask.
10853 (bdesc_args): Add __builtin_ia32_expanddf512,
10854 __builtin_ia32_expandsf512.
10855 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
10856 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
10857 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
10858 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
10859 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
10860 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
10861 (avx512f_<code>v8div16qi2_mask_store): This.
10862 (avx512f_expand<mode>): New.
10864 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
10866 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
10868 (_mm512_mask_prefetch_i64gather_pd): Ditto.
10869 (_mm512_prefetch_i32scatter_pd): Ditto.
10870 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
10871 (_mm512_prefetch_i64scatter_pd): Ditto.
10872 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
10873 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
10874 (_mm512_mask_prefetch_i64gather_ps): Ditto.
10875 (_mm512_prefetch_i32scatter_ps): Ditto.
10876 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
10877 (_mm512_prefetch_i64scatter_ps): Ditto.
10878 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
10879 * config/i386/i386-builtin-types.def: Define
10880 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
10881 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
10882 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
10883 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
10884 IX86_BUILTIN_SCATTERPFQPD.
10885 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
10886 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
10887 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
10888 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
10889 __builtin_ia32_scatterpfqps.
10890 (ix86_expand_builtin): Expand new built-ins.
10891 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
10892 fix memory access data type.
10893 (*avx512pf_gatherpf<mode>_mask): Ditto.
10894 (*avx512pf_gatherpf<mode>): Ditto.
10895 (avx512pf_scatterpf<mode>): Ditto.
10896 (*avx512pf_scatterpf<mode>_mask): Ditto.
10897 (*avx512pf_scatterpf<mode>): Ditto.
10898 (GATHER_SCATTER_SF_MEM_MODE): New.
10899 (avx512pf_gatherpf<mode>df): Ditto.
10900 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10901 (*avx512pf_scatterpf<mode>df): Ditto.
10903 2014-01-27 Jakub Jelinek <jakub@redhat.com>
10906 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
10907 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
10910 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10912 * common/config/arm/arm-common.c
10913 (arm_rewrite_mcpu): Handle multiple names.
10915 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
10917 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
10919 * gimple-builder.h (create_gimple_tmp): Delete.
10921 2014-01-27 Christian Bruel <christian.bruel@st.com>
10923 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
10926 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
10928 * config/pa/pa.md (call): Generate indirect long calls to non-local
10929 functions when outputing 32-bit code.
10930 (call_value): Likewise except for special call to buggy powf function.
10932 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
10933 portable runtime and PIC indirect calls.
10934 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
10935 and PIC call sequences. Use ldo instead of blr to set return register
10936 in PIC call sequence.
10938 2014-01-25 Walter Lee <walt@tilera.com>
10940 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
10941 avoid clobbering a live register.
10943 2014-01-25 Walter Lee <walt@tilera.com>
10945 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
10946 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
10947 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
10948 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
10950 2014-01-25 Walter Lee <walt@tilera.com>
10952 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
10953 arguments on even registers.
10954 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
10956 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
10957 (BIGGEST_ALIGNMENT): Ditto.
10958 (BIGGEST_FIELD_ALIGNMENT): Ditto.
10960 2014-01-25 Walter Lee <walt@tilera.com>
10962 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
10963 insns before bundling.
10964 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
10966 2014-01-25 Walter Lee <walt@tilera.com>
10968 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
10969 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
10970 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
10972 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10974 * config/mips/constraints.md (kl): Delete.
10975 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
10976 define expands, using...
10977 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
10978 instructions for MIPS16.
10979 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
10980 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
10982 2014-01-25 Walter Lee <walt@tilera.com>
10984 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
10988 2014-01-25 Walter Lee <walt@tilera.com>
10990 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
10991 (TARGET_EXPAND_TO_RTL_HOOK): Define.
10993 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
10995 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
10998 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11000 * print-rtl.c (in_call_function_usage): New var.
11001 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11002 EXPR_LIST mode as mode and not as reg note name.
11004 PR middle-end/59561
11005 * cfgloopmanip.c (copy_loop_info): If
11006 loop->warned_aggressive_loop_optimizations, make sure
11007 the flag is set in target loop too.
11009 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11011 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11013 * builtins.def: Likewise.
11014 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11015 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11016 * ira.c (ira_setup_eliminable_regset): Likewise.
11017 * omp-low.c (gate_expand_omp): Likewise.
11018 (execute_lower_omp): Likewise.
11019 (diagnose_sb_0): Likewise.
11020 (gate_diagnose_omp_blocks): Likewise.
11021 (simd_clone_clauses_extract): Likewise.
11024 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11026 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11027 correction for little endian...
11028 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11031 2014-01-24 Jeff Law <law@redhat.com>
11033 PR tree-optimization/59919
11034 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11035 for non-returning calls.
11037 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11039 * common/config/aarch64/aarch64-common.c
11040 (aarch64_rewrite_mcpu): Handle multiple names.
11041 * config/aarch64/aarch64.h
11042 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11044 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11046 * input.c (add_file_to_cache_tab): Handle the case where fopen
11049 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11052 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11053 from push operand if code of push isn't PRE_DEC.
11055 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11058 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11059 -mquad-memory-atomic. Update -mquad-memory documentation to say
11060 it is only used for non-atomic loads/stores.
11062 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11063 -mquad-memory or -mquad-memory-atomic switches.
11065 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11066 -mquad-memory-atomic to ISA 2.07 support.
11068 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11069 to separate support of normal quad word memory operations (ldq, stq)
11070 from the atomic quad word memory operations.
11072 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11073 support to separate non-atomic quad word operations from atomic
11074 quad word operations. Disable non-atomic quad word operations in
11075 little endian mode so that we don't have to swap words after the
11076 load and before the store.
11077 (quad_load_store_p): Add comment about atomic quad word support.
11078 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11079 options printed with -mdebug=reg.
11081 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11082 -mquad-memory-atomic as the test for whether we have quad word
11083 atomic instructions.
11084 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11085 or -mp8-vector are used, allow byte/half-word atomic operations.
11087 * config/rs6000/sync.md (load_lockedti): Insure that the address
11088 is a proper indexed or indirect address for the lqarx instruction.
11089 On little endian systems, swap the hi/lo registers after the lqarx
11091 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11092 insure the address is valid for the lqarx instruction.
11093 (store_conditionalti): Insure that the address is a proper indexed
11094 or indirect address for the stqcrx. instruction. On little endian
11095 systems, swap the hi/lo registers before doing the stqcrx.
11097 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11098 insure the address is valid for the stqcrx. instruction.
11100 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11101 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11102 type of quad memory support is available.
11104 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11106 PR regression/59915
11107 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11108 there is a danger of looping.
11110 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11112 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11113 force flag_ira_loop_pressure if set via command line.
11115 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11117 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11118 (ashr_simd): New builtin handling DI mode.
11119 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11120 (aarch64_sshr_simddi): New match pattern.
11121 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11122 (vshrd_n_s64): Likewise.
11123 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11125 2014-01-23 Nick Clifton <nickc@redhat.com>
11127 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11128 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11129 favour of mcu specific scripts.
11130 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11133 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11134 Alex Velenko <Alex.Velenko@arm.com>
11136 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11137 (vaddv_s16): Likewise.
11138 (vaddv_s32): Likewise.
11139 (vaddv_u8): Likewise.
11140 (vaddv_u16): Likewise.
11141 (vaddv_u32): Likewise.
11142 (vaddvq_s8): Likewise.
11143 (vaddvq_s16): Likewise.
11144 (vaddvq_s32): Likewise.
11145 (vaddvq_s64): Likewise.
11146 (vaddvq_u8): Likewise.
11147 (vaddvq_u16): Likewise.
11148 (vaddvq_u32): Likewise.
11149 (vaddvq_u64): Likewise.
11150 (vaddv_f32): Likewise.
11151 (vaddvq_f32): Likewise.
11152 (vaddvq_f64): Likewise.
11153 (vmaxv_f32): Likewise.
11154 (vmaxv_s8): Likewise.
11155 (vmaxv_s16): Likewise.
11156 (vmaxv_s32): Likewise.
11157 (vmaxv_u8): Likewise.
11158 (vmaxv_u16): Likewise.
11159 (vmaxv_u32): Likewise.
11160 (vmaxvq_f32): Likewise.
11161 (vmaxvq_f64): Likewise.
11162 (vmaxvq_s8): Likewise.
11163 (vmaxvq_s16): Likewise.
11164 (vmaxvq_s32): Likewise.
11165 (vmaxvq_u8): Likewise.
11166 (vmaxvq_u16): Likewise.
11167 (vmaxvq_u32): Likewise.
11168 (vmaxnmv_f32): Likewise.
11169 (vmaxnmvq_f32): Likewise.
11170 (vmaxnmvq_f64): Likewise.
11171 (vminv_f32): Likewise.
11172 (vminv_s8): Likewise.
11173 (vminv_s16): Likewise.
11174 (vminv_s32): Likewise.
11175 (vminv_u8): Likewise.
11176 (vminv_u16): Likewise.
11177 (vminv_u32): Likewise.
11178 (vminvq_f32): Likewise.
11179 (vminvq_f64): Likewise.
11180 (vminvq_s8): Likewise.
11181 (vminvq_s16): Likewise.
11182 (vminvq_s32): Likewise.
11183 (vminvq_u8): Likewise.
11184 (vminvq_u16): Likewise.
11185 (vminvq_u32): Likewise.
11186 (vminnmv_f32): Likewise.
11187 (vminnmvq_f32): Likewise.
11188 (vminnmvq_f64): Likewise.
11190 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11192 * config/aarch64/aarch64-simd.md
11193 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11194 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11195 (*aarch64_mul3_elt<mode>): Likewise.
11196 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11197 (*aarch64_mul3_elt_to_64v2df): Likewise.
11198 (*aarch64_mla_elt<mode>): Likewise.
11199 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11200 (*aarch64_mls_elt<mode>): Likewise.
11201 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11202 (*aarch64_fma4_elt<mode>): Likewise.
11203 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11204 (*aarch64_fma4_elt_to_64v2df): Likewise.
11205 (*aarch64_fnma4_elt<mode>): Likewise.
11206 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11207 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11208 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11209 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11210 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11211 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11212 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11213 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11214 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11216 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11218 * config/aarch64/aarch64-simd.md
11219 (aarch64_be_checked_get_lane<mode>): New define_expand.
11220 * config/aarch64/aarch64-simd-builtins.def
11221 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11222 New builtin definition.
11223 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11224 Use new safe be builtin.
11226 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11228 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11230 (aarch64_be_st1<mode>): Likewise.
11231 (aarch_ld1<VALL:mode>): Define_expand modified.
11232 (aarch_st1<VALL:mode>): Likewise.
11233 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11234 (UNSPEC_ST1): Likewise.
11236 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11238 * config/microblaze/microblaze.md: Add trap insn and attribute
11240 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11242 PR preprocessor/58580
11243 * input.h (location_get_source_line): Take an additional line_size
11245 (void diagnostics_file_cache_fini): Declare new function.
11246 * input.c (struct fcache): New type.
11247 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11248 New static constants.
11249 (diagnostic_file_cache_init, total_lines_num)
11250 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11251 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11252 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11253 (get_next_line, read_next_line, goto_next_line, read_line_num):
11254 New static function definitions.
11255 (diagnostic_file_cache_fini): New function.
11256 (location_get_source_line): Take an additional output line_len
11257 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11259 * diagnostic.c (diagnostic_finish): Call
11260 diagnostic_file_cache_fini.
11261 (adjust_line): Take an additional input parameter for the length
11262 of the line, rather than calculating it with strlen.
11263 (diagnostic_show_locus): Adjust the use of
11264 location_get_source_line and adjust_line with respect to their new
11265 signature. While displaying a line now, do not stop at the first
11266 null byte. Rather, display the zero byte as a space and keep
11267 going until we reach the size of the line.
11268 * Makefile.in: Add vec.o to OBJS-libcommon
11270 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11271 Ilya Tocar <ilya.tocar@intel.com>
11273 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11274 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11275 (__builtin_ia32_kmov16): Ditto.
11276 * config/i386/i386.md (UNSPEC_KMOV): New.
11279 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11281 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11282 (_mm512_storeu_si512): Ditto.
11284 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11287 * rtl.h (get_referenced_operands): Declare.
11288 * recog.c (get_referenced_operands): New function.
11289 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11290 operands have been referenced when recording LO_SUM references.
11292 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11294 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11296 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11298 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11299 Enable for generic and recent AMD targets.
11301 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11303 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11304 ARG_SIZE note when adjustment was eliminated.
11306 2014-01-22 Jeff Law <law@redhat.com>
11308 PR tree-optimization/59597
11309 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11310 in file. Accept new argument REGISTERING and use it to modify
11311 dump output appropriately.
11312 (register_jump_thread): Corresponding changes.
11313 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11314 thread paths involving joiner blocks. Add code to dump cancelled
11315 jump threading paths.
11317 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11319 PR rtl-optimization/59477
11320 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11321 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11323 2014-01-22 Tom Tromey <tromey@redhat.com>
11325 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11327 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11329 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11331 PR rtl-optimization/59896
11332 * lra-constraints.c (process_alt_operands): Check unused note for
11333 matched operands of insn with no output reloads.
11335 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11337 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11338 (mips_move_from_gpr_cost): Likewise.
11340 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11342 PR rtl-optimization/59858
11343 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11344 ira_class_hard_regs_num.
11345 (process_alt_operands): Increase reject for dying matched operand.
11347 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11350 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11351 smaller than size, perform several stores or loads and stores
11352 at dst + count - size to store or copy all of size bytes, rather
11353 than just last modesize bytes.
11355 2014-01-20 DJ Delorie <dj@redhat.com>
11357 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11358 that CLOBBERs are REGs before propogating their values.
11360 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11362 PR middle-end/59789
11363 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11364 (cgraph_inline_failed_type): New function.
11365 * cgraph.h (DEFCIFCODE): Add type.
11366 (cgraph_inline_failed_type_t): New enum.
11367 (cgraph_inline_failed_type): New prototype.
11368 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11369 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11370 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11371 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11372 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11373 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11374 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11375 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11376 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11377 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11378 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11379 OPTIMIZATION_MISMATCH.
11380 * tree-inline.c (expand_call_inline): Emit errors during
11381 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11383 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11386 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11387 mode attribute in insn output.
11389 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11391 * output.h (output_constant): Delete.
11392 * varasm.c (output_constant): Make private.
11394 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11396 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11398 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11400 PR middle-end/59860
11401 * tree.h (fold_builtin_strcat): New prototype.
11402 * builtins.c (fold_builtin_strcat): No longer static. Add len
11403 argument, if non-NULL, don't call c_strlen. Optimize
11404 directly into __builtin_memcpy instead of __builtin_strcpy.
11405 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11406 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11408 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11410 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11411 for SImode_address_operand operands, having only a REG argument.
11413 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11415 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11416 loader name using mbig-endian.
11417 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11419 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11421 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11422 (-mtune): Likewise.
11425 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11427 * config/aarch64/aarch64-protos.h
11428 (aarch64_cannot_change_mode_class_ptr): Declare.
11429 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11430 aarch64_cannot_change_mode_class_ptr): New.
11431 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11432 backend hook aarch64_cannot_change_mode_class.
11434 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11436 * common/config/aarch64/aarch64-common.c
11437 (aarch64_handle_option): Don't handle any option order logic here.
11438 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11439 selected_cpu, warn on architecture version mismatch.
11440 (aarch64_override_options): Fix parsing order for option strings.
11442 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11443 Iain Sandoe <iain@codesourcery.com>
11446 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11447 warning. Amend comment to reflect current functionality.
11449 2014-01-20 Richard Biener <rguenther@suse.de>
11451 PR middle-end/59860
11452 * builtins.c (fold_builtin_strcat): Remove case better handled
11453 by tree-ssa-strlen.c.
11455 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11457 * config/aarch64/aarch64.opt
11458 (mcpu, march, mtune): Make case-insensitive.
11460 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11463 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11464 if operands[1] is a REG or ZERO_EXTEND of a REG.
11466 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11468 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11470 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11472 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11473 long non-pic millicode calls.
11475 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11477 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11479 2014-01-19 Kito Cheng <kito@0xlab.org>
11481 * builtins.c (expand_movstr): Check movstr expand done or fail.
11483 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11484 H.J. Lu <hongjiu.lu@intel.com>
11487 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11488 to DImode for zero-extended addresses.
11490 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11492 PR rtl-optimization/57763
11493 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11494 on the new indirect jump_insn and increment LABEL_NUSES (label).
11496 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11500 * config.gcc (x86_archs): New variable.
11501 (x86_64_archs): Likewise.
11502 (x86_cpus): Likewise.
11503 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11504 --with-arch/--with-cpu= options.
11505 Support --with-arch=/--with-cpu={nehalem,westmere,
11506 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11508 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11510 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11511 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11513 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11515 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11517 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11520 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11521 clear cpp_get_options (parse_in)->warn_unused_macros for
11522 ix86_target_macros_internal with cpp_define.
11524 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11526 * jump.c (delete_related_insns): Keep (use (insn))s.
11527 * reorg.c (redundant_insn): Check for barriers too.
11529 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11531 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11533 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11535 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11536 call to $$dyncall when TARGET_LONG_CALLS is true.
11538 2014-01-17 Jeff Law <law@redhat.com>
11540 * ree.c (combine_set_extension): Temporarily disable test for
11541 changing number of hard registers.
11543 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11545 PR middle-end/58125
11546 * ipa-inline-analysis.c (inline_free_summary):
11547 Do not free summary of aliases.
11549 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11551 PR middle-end/59706
11552 * gimplify.c (gimplify_expr): Use create_tmp_var
11553 instead of create_tmp_var_raw. If cond doesn't have
11554 integral type, don't add the IFN_ANNOTATE builtin at all.
11556 2014-01-17 Martin Jambor <mjambor@suse.cz>
11559 * ipa-cp.c (prev_edge_clone): New variable.
11560 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11561 Also resize prev_edge_clone vector.
11562 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11563 (ipcp_edge_removal_hook): New function.
11564 (ipcp_driver): Register ipcp_edge_removal_hook.
11566 2014-01-17 Andrew Pinski <apinski@cavium.com>
11567 Steve Ellcey <sellcey@mips.com>
11570 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11573 2014-01-17 Jeff Law <law@redhat.com>
11575 PR middle-end/57904
11576 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11577 so that pass_ccp runs first.
11579 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11581 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11582 (ix86_adjust_cost): Use !TARGET_XXX.
11583 (do_reorder_for_imul): Likewise.
11584 (swap_top_of_ready_list): Likewise.
11585 (ix86_sched_reorder): Likewise.
11587 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11589 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11590 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11591 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11592 (intel_memset): New. Duplicate slm_memset.
11593 (intel_cost): New. Duplicate slm_cost.
11594 (m_INTEL): New macro.
11595 (processor_target_table): Add "intel".
11596 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11597 with PROCESSOR_INTEL for "intel".
11598 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11599 PROCESSOR_SILVERMONT.
11600 (ix86_issue_rate): Likewise.
11601 (ix86_adjust_cost): Likewise.
11602 (ia32_multipass_dfa_lookahead): Likewise.
11603 (swap_top_of_ready_list): Likewise.
11604 (ix86_sched_reorder): Likewise.
11605 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11606 instead of TARGET_OPT_AGU.
11607 * config/i386/i386.h (TARGET_INTEL): New.
11608 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11609 (processor_type): Add PROCESSOR_INTEL.
11610 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11611 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11613 2014-01-17 Marek Polacek <polacek@redhat.com>
11616 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11619 2014-01-17 Richard Biener <rguenther@suse.de>
11621 PR tree-optimization/46590
11622 * opts.c (default_options_table): Add entries for
11623 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11624 all enabled at -O1 but not for -Og.
11625 * common.opt (fbranch-count-reg): Remove Init(1).
11626 (fmove-loop-invariants): Likewise.
11627 (ftree-pta): Likewise.
11629 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11631 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11632 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11633 decls to at least the GCC 4.8 used alignments.
11636 * tree-nested.c (convert_nonlocal_reference_stmt,
11637 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11638 of GIMPLE_BIND stmts, adjust associated decls.
11640 2014-01-17 Richard Biener <rguenther@suse.de>
11642 PR tree-optimization/46590
11643 * vec.h (vec<>::bseach): New member function implementing
11644 binary search according to C89 bsearch.
11645 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11646 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11647 bitmap pointer again. Make accesses_in_loop a flat array.
11648 (mem_ref_obstack): New global.
11649 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11650 (mark_ref_stored): Likewise.
11651 (ref_indep_loop_p_2): Likewise.
11652 (set_ref_stored_in_loop): New helper function.
11653 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11654 (memref_free): Adjust.
11655 (record_mem_ref_loc): Simplify.
11656 (gather_mem_refs_stmt): Adjust.
11657 (sort_locs_in_loop_postorder_cmp): New function.
11658 (analyze_memory_references): Sort accesses_in_loop after
11659 loop postorder number.
11660 (find_ref_loc_in_loop_cmp): New function.
11661 (for_all_locs_in_loop): Find relevant cluster of locs in
11662 accesses_in_loop and iterate without recursion.
11663 (execute_sm): Avoid uninit warning.
11664 (struct ref_always_accessed): Simplify.
11665 (ref_always_accessed::operator ()): Likewise.
11666 (ref_always_accessed_p): Likewise.
11667 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11668 loop postorder numbers here.
11669 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11672 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11675 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11676 on decls for which assemble_alias has been called.
11678 2014-01-17 Nick Clifton <nickc@redhat.com>
11680 * config/msp430/msp430.opt: (mcpu): New option.
11681 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11682 (msp430_option_override): Parse target_cpu. If the MCU name
11683 matches a generic string, clear target_mcu.
11684 (msp430_attr): Allow numeric interrupt values up to 63.
11685 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11686 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11688 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11690 * config/msp430/msp430.md (popm): Use %J rather than %I.
11691 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11692 (addhi_cy_i): Use immediate_operand for operand 2.
11693 * doc/invoke.texi: Document -mcpu option.
11695 2014-01-17 Richard Biener <rguenther@suse.de>
11697 PR rtl-optimization/38518
11698 * df.h (df_analyze_loop): Declare.
11699 * df-core.c: Include cfgloop.h.
11700 (df_analyze_1): Split out main part of df_analyze.
11701 (df_analyze): Adjust.
11702 (loop_inverted_post_order_compute): New function.
11703 (loop_post_order_compute): Likewise.
11704 (df_analyze_loop): New function avoiding whole-function
11705 postorder computes.
11706 * loop-invariant.c (find_defs): Use df_analyze_loop.
11707 (find_invariants): Adjust.
11708 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11710 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11712 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11713 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11715 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11717 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11718 traversal when removing references.
11720 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11723 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11725 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11727 PR middle-end/56791
11728 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11729 pushing a reload for an autoinc when we had previously reloaded an
11730 inner part of the address.
11732 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11734 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11736 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11737 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11738 when not giving up or versioning for alias only because of
11740 (vect_analyze_data_ref_dependences): Set to true.
11741 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11743 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11744 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11747 PR middle-end/58344
11748 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11751 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11752 operand 0 predicate for gathers, use a new pseudo as subtarget.
11754 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11756 PR middle-end/59609
11757 * lra-constraints.c (process_alt_operands): Add printing debug info.
11758 Check absence of input/output reloads for matched operands too.
11760 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11762 PR rtl-optimization/59835
11763 * ira.c (ira_init_register_move_cost): Increase cost for
11766 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11768 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11770 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11773 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11774 non-register objects. Use gen_(high/low)part more consistently.
11777 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11780 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11781 endian support, remove tests for WORDS_BIG_ENDIAN.
11782 (p8_mfvsrd_3_<mode>): Likewise.
11783 (reload_gpr_from_vsx<mode>): Likewise.
11784 (reload_gpr_from_vsxsf): Likewise.
11785 (p8_mfvsrd_4_disf): Likewise.
11787 2014-01-16 Richard Biener <rguenther@suse.de>
11789 PR rtl-optimization/46590
11790 * lcm.c (compute_antinout_edge): Use postorder iteration.
11791 (compute_laterin): Use inverted postorder iteration.
11793 2014-01-16 Nick Clifton <nickc@redhat.com>
11795 PR middle-end/28865
11796 * varasm.c (output_constant): Return the number of bytes actually
11798 (output_constructor_array_range): Update the field size with the
11799 number of bytes emitted by output_constant.
11800 (output_constructor_regular_field): Likewise. Also do not
11801 complain if the total number of bytes emitted is now greater
11802 than the expected fieldpos.
11803 * output.h (output_constant): Update prototype and descriptive comment.
11805 2014-01-16 Marek Polacek <polacek@redhat.com>
11807 PR middle-end/59827
11808 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11809 it is error_mark_node.
11811 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11813 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11814 VALID_AVX256_REG_OR_OI_MODE.
11816 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11818 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11819 current procedure should be profiled.
11821 2014-01-15 Andrew Pinski <apinski@cavium.com>
11823 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11824 of moving from/to the STACK_REG register class.
11826 2014-01-15 Richard Henderson <rth@redhat.com>
11829 * reginfo.c (global_regs_decl): Globalize.
11830 * rtl.h (global_regs_decl): Declare.
11831 * ira.c (do_reload): Diagnose frame_pointer_needed and it
11832 reserved via global_regs.
11834 2014-01-15 Teresa Johnson <tejohnson@google.com>
11836 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
11838 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11840 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
11841 and vmulosh rather than call gen_vec_widen_smult_*.
11842 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
11843 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
11844 (vec_widen_smult_even_v16qi): Likewise.
11845 (vec_widen_umult_even_v8hi): Likewise.
11846 (vec_widen_smult_even_v8hi): Likewise.
11847 (vec_widen_umult_odd_v16qi): Likewise.
11848 (vec_widen_smult_odd_v16qi): Likewise.
11849 (vec_widen_umult_odd_v8hi): Likewise.
11850 (vec_widen_smult_odd_v8hi): Likewise.
11851 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
11852 vmuloub rather than call gen_vec_widen_umult_*.
11853 (vec_widen_umult_lo_v16qi): Likewise.
11854 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
11855 vmulosb rather than call gen_vec_widen_smult_*.
11856 (vec_widen_smult_lo_v16qi): Likewise.
11857 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
11858 rather than call gen_vec_widen_umult_*.
11859 (vec_widen_umult_lo_v8hi): Likewise.
11860 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
11861 rather than call gen_vec_widen_smult_*.
11862 (vec_widen_smult_lo_v8hi): Likewise.
11864 2014-01-15 Jeff Law <law@redhat.com>
11866 PR tree-optimization/59747
11867 * ree.c (find_and_remove_re): Properly handle case where a second
11868 eliminated extension requires widening a copy created for elimination
11869 of a prior extension.
11870 (combine_set_extension): Ensure that the number of hard regs needed
11871 for a destination register does not change when we widen it.
11873 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
11875 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
11876 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
11877 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
11878 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
11879 (avr-*-rtems*): Likewise.
11880 (bfin*-rtems*): Likewise.
11881 (moxie-*-rtems*): Likewise.
11882 (h8300-*-rtems*): Likewise.
11883 (i[34567]86-*-rtems*): Likewise.
11884 (lm32-*-rtems*): Likewise.
11885 (m32r-*-rtems*): Likewise.
11886 (m68k-*-rtems*): Likewise.
11887 (microblaze*-*-rtems*): Likewise.
11888 (mips*-*-rtems*): Likewise.
11889 (powerpc-*-rtems*): Likewise.
11890 (sh-*-rtems*): Likewise.
11891 (sparc-*-rtems*): Likewise.
11892 (sparc64-*-rtems*): Likewise.
11893 (v850-*-rtems*): Likewise.
11894 (m32c-*-rtems*): Likewise.
11896 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
11898 PR rtl-optimization/59511
11899 * ira.c (ira_init_register_move_cost): Use memory costs for some
11900 cases of register move cost calculations.
11901 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
11902 instead of BB frequency.
11903 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
11904 * lra-assigns.c (find_hard_regno_for): Ditto.
11906 2014-01-15 Richard Biener <rguenther@suse.de>
11908 PR tree-optimization/59822
11909 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
11910 (vectorizable_load): Use it to hoist defs of uses of invariant
11911 loads out of the loop.
11913 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11914 Kugan Vivekanandarajah <kuganv@linaro.org>
11917 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
11920 2014-01-15 Richard Biener <rguenther@suse.de>
11922 PR rtl-optimization/59802
11923 * lcm.c (compute_available): Use inverted postorder to seed
11924 the initial worklist.
11926 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11929 * config/s390/s390.c (s390_preferred_reload_class): Don't return
11930 ADDR_REGS for invalid symrefs in non-PIC code.
11932 2014-01-15 Jakub Jelinek <jakub@redhat.com>
11935 * builtins.c (determine_block_size): Initialize *probable_max_size
11936 even if len_rtx is CONST_INT.
11938 2014-01-14 Andrew Pinski <apinski@cavium.com>
11940 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
11941 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
11942 (cortexa53_tunings): Likewise.
11943 (aarch64_sched_issue_rate): New function.
11944 (TARGET_SCHED_ISSUE_RATE): Define.
11946 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11948 * ira-costs.c (find_costs_and_classes): Add missed
11949 ira_init_register_move_cost_if_necessary.
11951 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
11954 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
11956 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
11959 * config/i386/i386.c (type_natural_mode): Add a bool parameter
11960 to indicate if type is used for function return value. Warn ABI
11961 change if the vector mode isn't available for function return value.
11962 (ix86_function_arg_advance): Pass false to type_natural_mode.
11963 (ix86_function_arg): Likewise.
11964 (ix86_gimplify_va_arg): Likewise.
11965 (function_arg_32): Don't warn ABI change.
11966 (ix86_function_value): Pass true to type_natural_mode.
11967 (ix86_return_in_memory): Likewise.
11968 (ix86_struct_value_rtx): Removed.
11969 (TARGET_STRUCT_VALUE_RTX): Likewise.
11971 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11973 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
11974 converting a conditional jump into a conditional return.
11976 2014-01-14 Richard Biener <rguenther@suse.de>
11978 PR tree-optimization/58921
11979 PR tree-optimization/59006
11980 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
11981 hoisting invariant stmts.
11982 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
11983 invariant loads on the preheader edge if possible.
11985 2014-01-14 Joey Ye <joey.ye@arm.com>
11987 * doc/plugin.texi (Building GCC plugins): Update to C++.
11989 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
11991 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
11992 (_mm_rcp28_round_ss): Ditto.
11993 (_mm_rsqrt28_round_sd): Ditto.
11994 (_mm_rsqrt28_round_ss): Ditto.
11995 (_mm_rcp28_sd): Ditto.
11996 (_mm_rcp28_ss): Ditto.
11997 (_mm_rsqrt28_sd): Ditto.
11998 (_mm_rsqrt28_ss): Ditto.
11999 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12000 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12001 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12002 (IX86_BUILTIN_RCP28SD): Ditto.
12003 (IX86_BUILTIN_RCP28SS): Ditto.
12004 (IX86_BUILTIN_RSQRT28SD): Ditto.
12005 (IX86_BUILTIN_RSQRT28SS): Ditto.
12006 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12007 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12008 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12009 (ix86_expand_special_args_builtin): Expand new FTYPE.
12010 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12011 (srcp14<mode>): Make insn unary.
12012 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12013 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12014 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12015 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12016 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12017 Fix rounding: make it SAE only.
12018 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12020 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12022 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12023 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12024 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12025 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12026 (round_saeonly_mask_scalar_operand4): Ditto.
12027 (round_saeonly_mask_scalar_op3): Ditto.
12028 (round_saeonly_mask_scalar_op4): Ditto.
12030 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12032 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12033 Implement -maltivec=be for vec_insert and vec_extract.
12035 2014-01-10 DJ Delorie <dj@redhat.com>
12037 * config/msp430/msp430.md (call_internal): Don't allow memory
12038 references with SP as the base register.
12039 (call_value_internal): Likewise.
12040 * config/msp430/constraints.md (Yc): New. For memory references
12041 that don't use SP as a base register.
12043 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12044 "an integer without a # prefix"
12045 * config/msp430/msp430.md (epilogue_helper): Use it.
12047 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12050 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12051 AVX512F gather builtins.
12052 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12053 on gather decls with INTEGER_TYPE masktype.
12054 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12055 directly into the builtin rather than hoisting it before loop.
12057 PR tree-optimization/59387
12058 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12059 (scev_const_prop): If folded_casts and type has undefined overflow,
12060 use force_gimple_operand instead of force_gimple_operand_gsi and
12061 for each added stmt if it is assign with
12062 arith_code_with_undefined_signed_overflow, call
12063 rewrite_to_defined_overflow.
12064 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12065 gimple-fold.h instead.
12066 (arith_code_with_undefined_signed_overflow,
12067 rewrite_to_defined_overflow): Moved to ...
12068 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12069 rewrite_to_defined_overflow): ... here. No longer static.
12070 Include gimplify-me.h.
12071 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12072 rewrite_to_defined_overflow): New prototypes.
12074 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12076 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12078 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12080 * builtins.c (get_object_alignment_2): Minor tweak.
12081 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12083 2014-01-13 Christian Bruel <christian.bruel@st.com>
12085 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12086 optimized non constant lengths.
12088 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12091 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12092 load as __atomic_load_N if possible.
12094 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12096 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12098 (rs6000_expand_builtin): Adjust call.
12100 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12103 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12104 * config/rs6000/rs6000.c: Include target-globals.h.
12105 (rs6000_set_current_function): Instead of doing target_reinit
12106 unconditionally, use save_target_globals_default_opts and
12107 restore_target_globals.
12109 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12111 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12112 (rs6000_expand_builtin): Handle mffs and mtfsf.
12113 (rs6000_init_builtins): Define mffs and mtfsf.
12114 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12115 (rs6000_mffs): New pattern.
12116 (rs6000_mtfsf): New pattern.
12118 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12120 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12121 Start narrowing with START. Apply candidate-use pair
12122 and check overall cost in narrowing.
12123 (iv_ca_prune): Pass new argument.
12125 2014-01-10 Jeff Law <law@redhat.com>
12127 PR middle-end/59743
12128 * ree.c (combine_reaching_defs): Ensure the defining statement
12129 occurs before the extension when optimizing extensions with
12130 different source and destination hard registers.
12132 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12135 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12136 vtables into the type inheritance graph.
12138 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12140 PR rtl-optimization/59754
12141 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12142 modes in the REGNO != REGNO case.
12144 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12146 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12148 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12150 PR tree-optimization/59745
12151 * tree-predcom.c (tree_predictive_commoning_loop): Call
12152 free_affine_expand_cache if giving up because components is NULL.
12154 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12155 GC in payload of target_globals struct instead of allocating them on
12156 the heap and the larger structs separately using GC.
12157 * target-globals.h (struct target_globals): Make regs, hard_regs,
12158 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12159 of GTY((skip)) and change type to void *.
12160 (reset_target_globals): Cast loads from those fields to corresponding
12163 2014-01-10 Steve Ellcey <sellcey@mips.com>
12166 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12167 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12168 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12170 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12173 * aarch64-modes.def (CC_Zmode): New flags mode.
12174 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12175 represents an equality.
12176 (aarch64_get_condition_code): Handle CC_Zmode.
12177 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12179 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12181 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12182 extraction in good case.
12184 2014-01-10 Richard Biener <rguenther@suse.de>
12186 PR tree-optimization/59374
12187 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12188 checking after SLP discovery. Mark stmts not participating
12189 in any SLP instance properly.
12191 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12193 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12194 when handling a SET rtx.
12196 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12198 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12199 (cortex-a57): Likewise.
12200 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12202 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12204 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12205 non-iwmmxt builtins.
12207 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12211 * ipa-devirt.c record_target_from_binfo): Take as argument
12212 stack of binfos and lookup matching one for virtual inheritance.
12213 (possible_polymorphic_call_targets_1): Update.
12215 2014-01-10 Huacai Chen <chenhc@lemote.com>
12217 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12218 kernel strings for Loongson-2E/2F/3A.
12220 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12222 PR middle-end/59670
12223 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12224 is_gimple_call before calling gimple_call_internal_p.
12226 2014-01-09 Steve Ellcey <sellcey@mips.com>
12228 * Makefile.in (TREE_FLOW_H): Remove.
12229 (TREE_SSA_H): Add file names from tree-flow.h.
12230 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12231 * tree.h: Remove tree-flow.h reference.
12232 * hash-table.h: Remove tree-flow.h reference.
12233 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12234 reference with tree-ssa-loop.h.
12236 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12238 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12239 default element-order behavior for -maltivec.
12240 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12241 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12242 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12243 when targeting big endian, at least for now.
12244 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12246 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12248 PR middle-end/47735
12249 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12250 var satisfies use_register_for_decl, just take into account type
12251 alignment, rather than decl alignment.
12253 PR tree-optimization/59622
12254 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12255 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12256 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12257 Don't devirtualize for inplace at all. For targets.length () == 1,
12258 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12260 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12262 * config/i386/i386.md (cpu): Remove the unused btver1.
12264 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12266 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12268 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12271 * tree-core.h (struct target_globals): New forward declaration.
12272 (struct tree_target_option): Add globals field.
12273 * tree.h (TREE_TARGET_GLOBALS): Define.
12274 (prepare_target_option_nodes_for_pch): New prototype.
12275 * target-globals.h (struct target_globals): Define even if
12276 !SWITCHABLE_TARGET.
12277 * tree.c (prepare_target_option_node_for_pch,
12278 prepare_target_option_nodes_for_pch): New functions.
12279 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12280 * config/i386/i386.c: Include target-globals.h.
12281 (ix86_set_current_function): Instead of doing target_reinit
12282 unconditionally, use save_target_globals_default_opts and
12283 restore_target_globals.
12285 2014-01-09 Richard Biener <rguenther@suse.de>
12287 PR tree-optimization/59715
12288 * tree-cfg.h (split_critical_edges): Declare.
12289 * tree-cfg.c (split_critical_edges): Export.
12290 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12292 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12294 * cfgexpand.c (expand_stack_vars): Optionally disable
12295 asan stack protection.
12296 (expand_used_vars): Likewise.
12297 (partition_stack_vars): Likewise.
12298 * asan.c (asan_emit_stack_protection): Optionally disable
12299 after return stack usage.
12300 (instrument_derefs): Optionally disable memory access instrumentation.
12301 (instrument_builtin_call): Likewise.
12302 (instrument_strlen_call): Likewise.
12303 (asan_protect_global): Optionally disable global variables protection.
12304 * doc/invoke.texi: Added doc for new options.
12305 * params.def: Added new options.
12306 * params.h: Likewise.
12308 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12310 PR rtl-optimization/59724
12311 * ifcvt.c (cond_exec_process_if_block): Don't call
12312 flow_find_head_matching_sequence with 0 longest_match.
12313 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12314 non-active insns if !stop_after.
12315 (try_head_merge_bb): Revert 2014-01-07 changes.
12317 2014-01-08 Jeff Law <law@redhat.com>
12319 * ree.c (get_sub_rtx): New function, extracted from...
12320 (merge_def_and_ext): Here.
12321 (combine_reaching_defs): Use get_sub_rtx.
12323 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12325 * cgraph.h (varpool_variable_node): Do not choke on null node.
12327 2014-01-08 Catherine Moore <clm@codesourcery.com>
12329 * config/mips/mips.md (simple_return): Attempt to use JRC
12331 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12333 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12335 PR rtl-optimization/59137
12336 * reorg.c (steal_delay_list_from_target): Call update_block for
12338 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12340 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12342 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12343 two duplicate entries.
12345 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12348 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12350 * config/mips/mips.c (mips_truncated_op_cost): New function.
12351 (mips_rtx_costs): Adjust test for BADDU.
12352 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12354 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12356 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12357 (*baddu_si): ...this new pattern.
12359 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12362 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12364 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12366 PR middle-end/57748
12367 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12369 (expand_expr, expand_normal): Adjust.
12370 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12371 inner_reference_p. Use inner_reference_p to expand inner references.
12372 (store_expr): Adjust.
12373 * cfgexpand.c (expand_call_stmt): Adjust.
12375 2014-01-08 Rong Xu <xur@google.com>
12377 * gcov-io.c (gcov_var): Move from gcov-io.h.
12378 (gcov_position): Ditto.
12379 (gcov_is_error): Ditto.
12380 (gcov_rewrite): Ditto.
12381 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12382 only part to libgcc/libgcov.h.
12384 2014-01-08 Marek Polacek <polacek@redhat.com>
12386 PR middle-end/59669
12387 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12389 2014-01-08 Marek Polacek <polacek@redhat.com>
12392 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12394 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12396 PR rtl-optimization/59649
12397 * stor-layout.c (get_mode_bounds): For BImode return
12398 0 and STORE_FLAG_VALUE.
12400 2014-01-08 Richard Biener <rguenther@suse.de>
12402 PR middle-end/59630
12403 * gimple.h (is_gimple_builtin_call): Remove.
12404 (gimple_builtin_call_types_compatible_p): New.
12405 (gimple_call_builtin_p): New overload.
12406 * gimple.c (is_gimple_builtin_call): Remove.
12407 (validate_call): Rename to ...
12408 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12409 check return types.
12410 (validate_type): New static function.
12411 (gimple_call_builtin_p): New overload and adjust.
12412 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12413 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12414 (gimple_fold_stmt_to_constant_1): Likewise.
12415 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12417 2014-01-08 Richard Biener <rguenther@suse.de>
12419 PR middle-end/59471
12420 * gimplify.c (gimplify_expr): Gimplify register-register type
12421 VIEW_CONVERT_EXPRs to separate stmts.
12423 2014-01-07 Jeff Law <law@redhat.com>
12425 PR middle-end/53623
12426 * ree.c (combine_set_extension): Handle case where source
12427 and destination registers in an extension insn are different.
12428 (combine_reaching_defs): Allow source and destination registers
12429 in extension to be different under limited circumstances.
12430 (add_removable_extension): Remove restriction that the
12431 source and destination registers in the extension are the same.
12432 (find_and_remove_re): Emit a copy from the extension's
12433 destination to its source after the defining insn if
12434 the source and destination registers are different.
12436 PR middle-end/59285
12437 * ifcvt.c (merge_if_block): If we are merging a block with more than
12438 one successor with a block with no successors, remove any BARRIER
12439 after the second block.
12441 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12443 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12445 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12448 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12449 for 14-bit register offsets when INT14_OK_STRICT is false.
12451 2014-01-07 Roland Stigge <stigge@antcom.de>
12452 Michael Meissner <meissner@linux.vnet.ibm.com>
12455 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12456 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12458 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12460 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12463 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12465 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12466 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12467 rtx is const0_rtx or not.
12469 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12472 * target-globals.c (save_target_globals): Remove this_fn_optab
12474 * toplev.c: Include optabs.h.
12475 (target_reinit): Temporarily restore the global options if another
12476 set of options are in force.
12478 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12480 PR rtl-optimization/58668
12481 * cfgcleanup.c (flow_find_cross_jump): Don't count
12482 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12483 to determine what is counted.
12484 (flow_find_head_matching_sequence): Use active_insn_p to determine
12486 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12488 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12489 determine what is counted.
12491 PR tree-optimization/59643
12492 * tree-predcom.c (split_data_refs_to_components): If one dr is
12493 read and one write, determine_offset fails and the write isn't
12494 in the bad component, just put the read into the bad component.
12496 2014-01-07 Mike Stump <mikestump@comcast.net>
12497 Jakub Jelinek <jakub@redhat.com>
12500 * tree-core.h (struct tree_optimization_option): Change optabs
12501 type from unsigned char * to void *.
12502 * optabs.c (init_tree_optimization_optabs): Adjust
12503 TREE_OPTIMIZATION_OPTABS initialization.
12505 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12508 * config/i386/i386.h (struct machine_function): Add
12509 no_drap_save_restore field.
12510 * config/i386/i386.c (ix86_save_reg): Use
12511 !cfun->machine->no_drap_save_restore instead of
12512 crtl->stack_realign_needed.
12513 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12514 this function clears frame_pointer_needed. Set
12515 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12516 and DRAP reg is needed.
12518 2014-01-06 Marek Polacek <polacek@redhat.com>
12521 * doc/implement-c.texi: Mention that other integer types are
12522 permitted as bit-field types in strictly conforming mode.
12524 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12526 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12527 is newly allocated.
12529 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12531 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12533 2014-01-06 Martin Jambor <mjambor@suse.cz>
12536 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12538 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12540 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12544 * var-tracking.c (add_stores): Preserve the value of the source even if
12545 we don't record the store.
12547 2014-01-06 Terry Guo <terry.guo@arm.com>
12549 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12551 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12554 * config/darwin.c (darwin_function_section): Adjust return values to
12555 correspond to optimisation changes made in r206070.
12557 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12559 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12560 from prefetch_block tune setting.
12561 (nocona_cost): Correct size of prefetch block to 64.
12563 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12565 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12566 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12567 used to save the static chain register in the computation of the offset
12568 from which the FP registers need to be restored.
12570 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12572 PR tree-optimization/59519
12573 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12574 ICE if get_current_def (current_new_name) is already non-NULL, as long
12575 as it is a phi result of some other phi in *new_exit_bb that has
12578 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12579 or vmovdqu* for misaligned_operand.
12580 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12581 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12582 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12583 aligned_mem for AVX512F masked aligned load and store builtins and for
12584 non-temporal moves.
12586 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12588 PR tree-optimization/59651
12589 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12590 Address range for negative step should be added by TYPE_SIZE_UNIT.
12592 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12594 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12595 overlapping registers also for registers other than the stack pointer.
12597 2014-01-03 Marek Polacek <polacek@redhat.com>
12600 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12603 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12606 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12609 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12610 (push splitter): Use <P:MODE_SIZE> instead of
12611 GET_MODE_SIZE (<P:MODE>mode).
12612 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12613 (mov -1, reg peephole2): Likewise.
12614 * config/i386/sse.md (*mov<mode>_internal,
12615 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12616 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12617 *<code><mode>3, *andnot<mode>3<mask_name>,
12618 <mask_codefor><code><mode>3<mask_name>): Likewise.
12619 * config/i386/subst.md (mask_mode512bit_condition,
12620 sd_mask_mode512bit_condition): Likewise.
12622 2014-01-02 Xinliang David Li <davidxl@google.com>
12624 PR tree-optimization/59303
12625 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12626 (dump_predicates): Better output format.
12627 (pred_equal_p): New function.
12628 (is_neq_relop_p): Ditto.
12629 (is_neq_zero_form_p): Ditto.
12630 (pred_expr_equal_p): Ditto.
12631 (pred_neg_p): Ditto.
12632 (simplify_pred): Ditto.
12633 (simplify_preds_2): Ditto.
12634 (simplify_preds_3): Ditto.
12635 (simplify_preds_4): Ditto.
12636 (simplify_preds): Ditto.
12637 (push_pred): Ditto.
12638 (push_to_worklist): Ditto.
12639 (get_pred_info_from_cmp): Ditto.
12640 (is_degenerated_phi): Ditto.
12641 (normalize_one_pred_1): Ditto.
12642 (normalize_one_pred): Ditto.
12643 (normalize_one_pred_chain): Ditto.
12644 (normalize_preds): Ditto.
12645 (normalize_cond_1): Remove function.
12646 (normalize_cond): Ditto.
12647 (is_gcond_subset_of): Ditto.
12648 (is_subset_of_any): Ditto.
12649 (is_or_set_subset_of): Ditto.
12650 (is_and_set_subset_of): Ditto.
12651 (is_norm_cond_subset_of): Ditto.
12652 (pred_chain_length_cmp): Ditto.
12653 (convert_control_dep_chain_into_preds): Type change.
12654 (find_predicates): Ditto.
12655 (find_def_preds): Ditto.
12656 (destroy_predicates_vecs): Ditto.
12657 (find_matching_predicates_in_rest_chains): Ditto.
12658 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12659 (is_pred_expr_subset): Ditto.
12660 (is_pred_chain_subset_of): Ditto.
12661 (is_included_in): Ditto.
12662 (is_superset_of): Ditto.
12664 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12666 Update copyright years.
12668 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12670 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12671 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12672 config/arc/arc.md, config/arc/arc.opt,
12673 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12674 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12675 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12676 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12677 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12678 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12679 vtable-verify.c, vtable-verify.h: Use the standard form for the
12682 2014-01-02 Tobias Burnus <burnus@net-b.de>
12684 * gcc.c (process_command): Update copyright notice dates.
12685 * gcov-dump.c: Ditto.
12687 * doc/cpp.texi: Bump @copying's copyright year.
12688 * doc/cppinternals.texi: Ditto.
12689 * doc/gcc.texi: Ditto.
12690 * doc/gccint.texi: Ditto.
12691 * doc/gcov.texi: Ditto.
12692 * doc/install.texi: Ditto.
12693 * doc/invoke.texi: Ditto.
12695 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12697 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12699 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12701 * config/i386/sse.md (*mov<mode>_internal): Guard
12702 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12704 PR rtl-optimization/59647
12705 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12706 new_rtx into UNSIGNED_FLOAT rtxes.
12708 Copyright (C) 2014 Free Software Foundation, Inc.
12710 Copying and distribution of this file, with or without modification,
12711 are permitted in any medium without royalty provided the copyright
12712 notice and this notice are preserved.