1 2014-05-28 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
4 * wide-int.cc: Do not include longlong.h when compiling with clang.
6 2014-05-28 Richard Biener <rguenther@suse.de>
8 * tree-ssa-propagate.c (add_control_edge): Print less vertical space.
9 * tree-vrp.c (extract_range_from_ssa_name): Also copy VR_UNDEFINED.
10 (vrp_visit_assignment_or_call): Print less vertical space.
11 (vrp_visit_stmt): Likewise.
12 (vrp_visit_phi_node): Likewise. For a PHI argument with
13 VR_VARYING range consider recording it as copy.
15 2014-05-28 Richard Biener <rguenther@suse.de>
18 2014-05-28 Richard Biener <rguenther@suse.de>
20 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
22 2014-05-28 Bernd Edlinger <bernd.edlinger@hotmail.de>
24 * expr.c (expand_assignment): Fold the bitpos in the to_rtx if
25 sufficiently aligned and an offset is used at the same time.
26 (expand_expr_real_1): Likewise.
28 2014-05-28 Richard Biener <rguenther@suse.de>
31 * fold-const.c (fold_comparison): When folding
32 X +- C1 CMP Y +- C2 to X CMP Y +- C2 +- C1 also ensure
33 the sign of the remaining constant operand stays the same.
35 2014-05-28 Kaushik Phatak <kaushik.phatak@kpit.com>
37 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
38 __RL78_64BIT_DOUBLES__ or __RL78_32BIT_DOUBLES__.
39 (ASM_SPEC): Pass -m64bit-doubles or -m32bit-doubles on
41 (DOUBLE_TYPE_SIZE): Use 64 bit if TARGET_64BIT_DOUBLES
43 * gcc/config/rl78/rl78.opt (m64bit-doubles): New option.
44 (m32bit-doubles) Likewise.
45 * gcc/config/rl78/t-rl78: Add 64-bit-double multilib.
46 * doc/invoke.texi: Document -m32bit-doubles and -m64bit-doubles
49 2014-05-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
51 * configure.ac ($gcc_cv_ld_clearcap): New test.
52 * configure: Regenerate.
53 * config.in: Regenerate.
54 * config/sol2.opt (mclear-hwcap): New option.
55 * config/sol2.h (LINK_CLEARCAP_SPEC): Define.
56 * config/sol2-clearcap.map: Moved here from
57 testsuite/gcc.target/i386/clearcap.map.
58 * config/sol2-clearcapv2.map: Move here from
59 gcc.target/i386/clearcapv2.map.
60 * config/t-sol2 (install): Depend on install-clearcap-map.
61 (install-clearcap-map): New target.
62 * doc/invoke.texi (Option Summary, Solaris 2 Options): Document
65 2014-05-28 Richard Biener <rguenther@suse.de>
67 * hwint.h (*_HALF_WIDE_INT*): Move to ...
68 * wide-int.cc (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT):
69 ... here and remove the rest.
70 * hwint.h (HOST_WIDE_INT_PRINT_*): Define in terms of PRI*64.
72 2014-05-28 Richard Biener <rguenther@suse.de>
74 PR tree-optimization/61335
75 * tree-vrp.c (vrp_visit_phi_node): If the compare of old and
76 new range fails, drop to varying.
78 2014-05-28 Tom de Vries <tom@codesourcery.com>
80 * lra-int.h (struct lra_reg): Add field actual_call_used_reg_set.
81 * lra.c (initialize_lra_reg_info_element): Add init of
82 actual_call_used_reg_set field.
83 (lra): Call lra_create_live_ranges before lra_inheritance for
85 * lra-assigns.c (lra_assign): Allow call_used_regs to cross calls for
87 * lra-constraints.c (need_for_call_save_p): Use actual_call_used_reg_set
88 instead of call_used_reg_set for -fuse-caller-save.
89 * lra-lives.c (process_bb_lives): Calculate actual_call_used_reg_set.
91 2014-05-28 Radovan Obradovic <robradovic@mips.com>
92 Tom de Vries <tom@codesourcery.com>
94 * doc/invoke.texi (@item Optimization Options): Add -fuse-caller-save to
96 (@item -fuse-caller-save): New item.
98 2014-05-28 Radovan Obradovic <robradovic@mips.com>
99 Tom de Vries <tom@codesourcery.com>
101 * opts.c (default_options_table): Add OPT_LEVELS_2_PLUS entry with
102 OPT_fuse_caller_save.
104 2014-05-28 Radovan Obradovic <robradovic@mips.com>
105 Tom de Vries <tom@codesourcery.com>
107 * df-scan.c (df_get_call_refs): Use get_call_reg_set_usage.
108 * caller-save.c (setup_save_areas, save_call_clobbered_regs): Use
109 get_call_reg_set_usage.
110 * resource.c (mark_set_resources, mark_target_live_regs): Use
111 get_call_reg_set_usage.
112 * ira-int.h (struct ira_allocno): Add crossed_calls_clobbered_regs
114 (ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS): Define.
115 * ira-lives.c (process_bb_node_lives): Use get_call_reg_set_usage.
116 Calculate ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
117 * ira-build.c (ira_create_allocno): Init
118 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
119 (create_cap_allocno, propagate_allocno_info)
120 (propagate_some_info_from_allocno)
121 (copy_info_to_removed_store_destinations): Handle
122 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS.
123 * ira-costs.c (ira_tune_allocno_costs): Use
124 ALLOCNO_CROSSED_CALLS_CLOBBERED_REGS to adjust costs.
126 2014-05-28 Radovan Obradovic <robradovic@mips.com>
127 Tom de Vries <tom@codesourcery.com>
129 * cgraph.h (struct cgraph_rtl_info): Add function_used_regs
130 and function_used_regs_valid fields.
131 * final.c: Move include of hard-reg-set.h to before rtl.h to declare
132 find_all_hard_reg_sets.
133 (collect_fn_hard_reg_usage, get_call_fndecl, get_call_cgraph_rtl_info)
134 (get_call_reg_set_usage): New function.
135 (rest_of_handle_final): Use collect_fn_hard_reg_usage.
136 * regs.h (get_call_reg_set_usage): Declare.
138 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
141 * config/dbx.h (License): Add Runtime Library Exception.
142 * config/newlib-stdint.h (License): Same.
143 * config/rtems.h (License): Same
144 * config/initfini-array.h (License): Same
145 * config/v850/v850.h (License): Same.
146 * config/v850/v850-opts.h (License): Same
147 * config/v850/rtems.h (License): Same.
149 2014-05-28 Georg-Johann Lay <avr@gjlay.de>
152 * doc/extend.texi (Local Labels): Note that label differences are
153 not supported for AVR.
155 2014-05-27 Andrew Pinski <apinski@cavium.com>
157 * config/aarch64/aarch64.md (stack_protect_set_<mode>):
158 Use <w> for the register in assembly template.
159 (stack_protect_test): Use the mode of operands[0] for the
161 (stack_protect_test_<mode>): Use <w> for the register
162 in assembly template.
164 2014-05-27 DJ Delorie <dj@redhat.com>
166 * config/rx/rx.c (add_vector_labels): New.
167 (rx_output_function_prologue): Call it.
168 (rx_handle_func_attribute): Don't require empty arguments.
169 (rx_handle_vector_attribute): New.
170 (rx_attribute_table): Add "vector" attribute.
171 * doc/extend.texi (interrupt, vector): Document new/changed
172 RX-specific attributes.
174 * config/rx/rx.c (rx_adjust_insn_length): Skip for non-insns.
176 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
178 * double-int.c (div_and_round_double) <ROUND_DIV_EXPR>: Use the proper
179 predicate to detect a negative quotient.
181 2014-05-27 Eric Botcazou <ebotcazou@adacore.com>
183 * fold-const.c (fold_comparison): Clean up and extend X +- C1 CMP C2
184 to X CMP C2 -+ C1 transformation to EQ_EXPR/NE_EXPR.
185 Add X - Y CMP 0 to X CMP Y transformation.
186 (fold_binary_loc) <EQ_EXPR/NE_EXPR>: Remove same transformations.
188 2014-05-27 Segher Boessenkool <segher@kernel.crashing.org>
190 * stmt.c (dump_case_nodes): Don't convert values to HOST_WIDE_INT
193 2014-05-27 Steve Ellcey <sellcey@mips.com>
195 * config/mips/mips.c: Add include of cgraph.h.
197 2014-05-27 Richard Biener <rguenther@suse.de>
199 * system.h (__STDC_FORMAT_MACROS): Define as very first thing.
201 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
204 * config/arm/arm.h (License): Add note to COPYING.RUNTIME.
205 * config/arm/arm-cores.def (License): Same.
206 * config/arm/arm-opts.h (License): Same.
207 * config/arm/aout.h (License): Same.
208 * config/arm/bpabi.h (License): Same.
209 * config/arm/elf.h (License): Same.
210 * config/arm/linux-elf.h (License): Same.
211 * config/arm/linux-gas.h (License): Same.
212 * config/arm/netbsd-elf.h (License): Same.
213 * config/arm/uclinux-eabi.h (License): Same.
214 * config/arm/uclinux-elf.h (License): Same.
215 * config/arm/vxworks.h (License): Same.
217 2014-05-27 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
219 * config/arm/neon.md (neon_bswap<mode>): New pattern.
220 * config/arm/arm.c (neon_itype): Add NEON_BSWAP.
221 (arm_init_neon_builtins): Handle NEON_BSWAP.
222 Define required type nodes.
223 (arm_expand_neon_builtin): Handle NEON_BSWAP.
224 (arm_builtin_vectorized_function): Handle BUILTIN_BSWAP builtins.
225 * config/arm/arm_neon_builtins.def (bswap): Define builtins.
226 * config/arm/iterators.md (VDQHSD): New mode iterator.
228 2014-05-27 Richard Biener <rguenther@suse.de>
230 * tree-vrp.c (vrp_evaluate_conditional_warnv_with_ops_using_ranges):
231 Try using literal operands when comparing value-ranges failed.
233 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
235 * ira.c (commutative_operand): Adjust for change to recog_data.
236 [Missing from previous commit.]
238 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
240 * system.h (TEST_BIT): New macro.
241 * recog.h (alternative_mask): New type.
242 (ALL_ALTERNATIVES, ALTERNATIVE_BIT): New macros.
243 (recog_data_d): Replace alternative_enabled_p array with
244 enabled_alternatives.
245 (target_recog): New structure.
246 (default_target_recog, this_target_recog): Declare.
247 (get_enabled_alternatives, recog_init): Likewise.
248 * recog.c (default_target_recog, this_target_recog): New variables.
249 (get_enabled_alternatives): New function.
250 (extract_insn): Use it.
251 (recog_init): New function.
252 (preprocess_constraints, constrain_operands): Adjust for change to
254 * postreload.c (reload_cse_simplify_operands): Likewise.
255 * reload.c (find_reloads): Likewise.
256 * ira-costs.c (record_reg_classes): Likewise.
257 * ira-lives.c (single_reg_class): Likewise. Fix bug in which
258 all alternatives after a disabled one would be skipped.
259 (ira_implicitly_set_insn_hard_regs): Likewise.
260 * ira.c (ira_setup_alts): Adjust for change to recog_data.
261 * lra-int.h (lra_insn_recog_data): Replace alternative_enabled_p
262 with enabled_alternatives.
263 * lra.c (free_insn_recog_data): Update accordingly.
264 (lra_update_insn_recog_data): Likewise.
265 (lra_set_insn_recog_data): Likewise. Use get_enabled_alternatives.
266 * lra-constraints.c (process_alt_operands): Likewise. Handle
267 only_alternative as part of the enabled mask.
268 * target-globals.h (this_target_recog): Declare.
269 (target_globals): Add a recog field.
270 (restore_target_globals): Restore this_target_recog.
271 * target-globals.c: Include recog.h.
272 (default_target_globals): Initialize recog field.
273 (save_target_globals): Likewise.
274 * reginfo.c (reinit_regs): Call recog_init.
275 * toplev.c (backend_init_target): Likewise.
277 2014-05-27 Richard Sandiford <rdsandiford@googlemail.com>
279 * gencodes.c (main): Make LAST_INSN_CODE higher than any insn code,
280 rather than any named insn's code.
282 2014-05-27 Georg-Johann Lay <avr@gjlay.de>
285 * config/arm/arm-opts.h (License): Add GCC Runtime Library Exception.
286 * config/arm/arm-cores.def (License): Same.
288 2014-05-26 Jan Hubicka <hubicka@ucw.cz>
290 * tree.h (decl_comdat_group): Declare.
291 * cgraph.h (symtab_in_same_comdat_p): Move offline to ...
292 * tree.c (decl_comdat_group): Here.
294 2014-05-26 Richard Sandiford <r.sandiford@uk.ibm.com>
296 PR rtl-optimization/61222
297 * combine.c (simplify_shift_const_1): When moving a PLUS outside
298 the shift, truncate the PLUS operand to the result mode.
300 2014-05-26 Uros Bizjak <ubizjak@gmail.com>
303 * config/i386/i386.c (ix86_rtx_costs)
304 <case CONST_INT, case CONST, case LABEL_REF, case SYMBOL_REF>:
307 2014-05-26 Martin Jambor <mjambor@suse.cz>
309 * ira.c (split_live_ranges_for_shrink_wrap): Remove bailout on
312 2014-05-26 Richard Biener <rguenther@suse.de>
314 * wide-int.h (wi::int_traits <long>, wi::int_traits <unsigned long>,
315 wi::int_traits <long long>, wi::int_traits <unsigned long long>):
316 Provide specializations.
317 (wi::int_traits <HOST_WIDE_INT>,
318 wi::int_traits <unsigned HOST_WIDE_INT>): Remove specializations.
320 2014-05-26 Alan Modra <amodra@gmail.com>
323 * config/rs6000/rs6000.c (rs6000_emit_set_const): Remove unneeded
324 params and return a bool. Remove dead code. Update comment.
325 Assert we have a const_int source. Remove bogus code from
326 32-bit HWI days. Move !TARGET_POWERPC64 handling, and correct
327 handling of constants > 2G and reg_equal note, from..
328 (rs6000_emit_set_long_const): ..here. Remove unneeded param and
329 return value. Update comment. If we can, use a new pseudo
330 for intermediate calculations.
331 * config/rs6000/rs6000-protos.h (rs6000_emit_set_const): Update
333 * config/rs6000/rs6000.md (movsi_internal1_single+1): Update
334 call to rs6000_emit_set_const in splitter.
335 (movdi_internal64+2, +3): Likewise.
337 2014-05-26 Richard Biener <rguenther@suse.de>
339 * system.h: Define __STDC_FORMAT_MACROS before
340 including inttypes.h.
341 * hwint.h (HOST_WIDEST_INT, HOST_BITS_PER_WIDEST_INT,
342 HOST_WIDEST_INT_PRINT, HOST_WIDEST_INT_PRINT_DEC,
343 HOST_WIDEST_INT_PRINT_DEC_C, HOST_WIDEST_INT_PRINT_UNSIGNED,
344 HOST_WIDEST_INT_PRINT_HEX, HOST_WIDEST_INT_PRINT_DOUBLE_HEX,
345 HOST_WIDEST_INT_C): Remove.
346 (PRId64, PRIi64, PRIo64, PRIu64, PRIx64, PRIX64): Define
347 if C99 inttypes.h is not available.
348 * coretypes.h (gcov_type, gcov_type_unsigned): Use [u]int64_t.
349 * gcov-io.h (gcov_type, gcov_type_unsigned): Likewise.
350 * gcov-io.c (gcov_histo_index): Drop non-64bit hwi case.
351 * cfgloop.h (struct niter_desc): Use uint64_t for niter field.
352 * bitmap.c (struct bitmap_descriptor_d): Use uint64_t for counters.
353 (struct output_info): Likewise.
354 (print_statistics): Adjust.
355 (dump_bitmap_statistics): Likewise.
356 * bt-load.c (migrate_btr_defs): Print with PRId64.
357 * cfg.c (dump_edge_info, dump_bb_info): Likewise.
358 (MAX_SAFE_MULTIPLIER): Adjust.
359 * cfghooks.c (dump_bb_for_graph): Print with PRId64.
360 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee,
361 dump_cgraph_node): Likewise.
362 * final.c (dump_basic_block_info): Likewise.
363 * gcov-dump.c (tag_counters, tag_summary, dump_working_sets): Likewise.
364 * gcov.c (format_gcov): Likewise.
365 * ipa-cp.c (good_cloning_opportunity_p): Likewise. Use int64_t
367 (get_clone_agg_value): Use HOST_WIDE_INT for offset.
368 * ipa-inline.c (compute_max_insns): Use int64_t for calcuation.
369 (inline_small_functions, dump_overall_stats, dump_inline_stats):
370 Use PRId64 for dumping.
371 * ipa-profile.c (dump_histogram, ipa_profile): Likewise.
372 * ira-color.c (struct allocno_hard_regs): Use int64_t for cost.
373 (add_allocno_hard_regs): Adjust.
374 * loop-doloop.c (doloop_modify): Print using PRId64.
375 * loop-iv.c (inverse): Compute in uint64_t.
376 (determine_max_iter, iv_number_of_iterations): Likewise.
377 * loop-unroll.c (decide_peel_completely, decide_peel_simple):
379 * lto-streamer-out.c (write_symbol): Use uint64_t.
380 * mcf.c (CAP_INFINITY): Use int64_t maximum.
381 (dump_fixup_edge, create_fixup_graph, cancel_negative_cycle,
382 find_max_flow, adjust_cfg_counts): Use int64_t and dump with PRId64.
383 * modulo-sched.c (const_iteration_count): Use int64_t.
384 (sms_schedule): Dump using PRId64.
385 * predict.c (dump_prediction): Likewise.
386 * pretty-print.h (pp_widest_integer): Remove.
387 * profile.c (get_working_sets, is_edge_inconsistent,
388 is_inconsistent, read_profile_edge_counts): Dump using PRId64.
389 * tree-pretty-print.c (pp_double_int): Remove case handling
390 HOST_BITS_PER_DOUBLE_INT == HOST_BITS_PER_WIDEST_INT.
391 * tree-ssa-math-opts.c (struct symbolic_number): Use uint64_t
393 (pass_optimize_bswap::execute): Remove restriction on hosts.
394 * tree-streamer-in.c (streamer_alloc_tree): Use HOST_WIDE_INT.
395 * tree-streamer-out.c (streamer_write_tree_header): Likewise.
396 * tree.c (widest_int_cst_value): Remove.
397 * tree.h (widest_int_cst_value): Likewise.
398 * value-prof.c (dump_histogram_value): Print using PRId64.
399 * gengtype.c (main): Also inject int64_t.
400 * ggc-page.c (struct max_alignment): Use int64_t.
401 * alloc-pool.c (struct allocation_object_def): Likewise.
402 * ira-conflicts.c (build_conflict_bit_table): Use uint64_t
404 * doc/tm.texi.in: Remove reference to HOST_WIDEST_INT.
405 * doc/tm.texi: Regenerated.
406 * gengtype-lex.l (IWORD): Handle [u]int64_t.
407 * config/sh/sh.c (expand_cbranchdi4): Use gcov_type.
408 * config/mmix/mmix-protos.h (mmix_intval, mmix_shiftable_wyde_value,
409 mmix_output_register_setting): Use [u]int64_t in prototypes.
410 * config/mmix/mmix.c (mmix_print_operand, mmix_output_register_setting,
411 mmix_shiftable_wyde_value, mmix_output_shiftvalue_op_from_str,
412 mmix_output_octa, mmix_output_shifted_value): Adjust.
413 (mmix_intval): Adjust. Remove unreachable case.
414 * config/mmix/mmix.md (*nonlocal_goto_receiver_expanded): Use int64_t.
416 2014-05-26 Richard Biener <rguenther@suse.de>
418 * configure.ac: Drop __int64 type check. Insist that we
419 found uint64_t and int64_t.
420 * hwint.h (HOST_BITS_PER___INT64): Remove.
421 (HOST_BITS_PER_WIDE_INT): Define to 64 and remove __int64 case.
422 (HOST_WIDE_INT_PRINT_*): Remove 32bit case.
423 (HOST_WIDEST_INT*): Define to HOST_WIDE_INT*.
424 (HOST_WIDEST_FAST_INT): Remove __int64 case.
425 * vmsdbg.h (struct _DST_SRC_COMMAND): Use int64_t
426 for dst_q_src_df_rms_cdt.
427 * configure: Regenerate.
428 * config.in: Likewise.
430 2014-05-26 Michael Tautschnig <mt@debian.org>
433 * doc/extend.texi (X86 Built-in Functions): Fix parameter lists of
434 __builtin_ia32_vfrczs[sd] and __builtin_ia32_mpsadbw256.
436 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
438 PR rtl-optimization/61278
439 * shrink-wrap.c (move_insn_for_shrink_wrap): Check df_live.
441 2014-05-26 Zhenqiang Chen <zhenqiang.chen@linaro.org>
443 PR rtl-optimization/61220
444 Part of PR rtl-optimization/61225
445 * shrink-wrap.c (move_insn_for_shrink_wrap): Skip SP and FP adjustment
446 insn; skip split_edge for a block with only one successor.
448 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
450 * symtab.c (symtab_nonoverwritable_alias): Copy READONLY flag
453 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
455 * ipa-visibility.c (can_replace_by_local_alias_in_vtable): New function.
456 (update_vtable_references): New function.
457 (function_and_variable_visibility): Rewrite also vtable initializers.
458 * varpool.c (cgraph_variable_initializer_availability): Remove assert.
460 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
462 * ggc.h (ggc_grow): New function.
463 * ggc-none.c (ggc_grow): New function.
464 * ggc-page.c (ggc_grow): Likewise.
466 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
468 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
469 address_taken_from_non_vtable_p, comdat_can_be_unshared_p_1,
470 comdat_can_be_unshared_p, cgraph_externally_visible_p,
471 varpool_externally_visible_p, can_replace_by_local_alias,
472 update_visibility_by_resolution_info, function_and_variable_visibility,
473 pass_data_ipa_function_and_variable_visibility,
474 make_pass_ipa_function_and_variable_visibility,
475 whole_program_function_and_variable_visibility,
476 pass_data_ipa_whole_program_visibility,
477 make_pass_ipa_whole_program_visibility): Move to ipa-visibility.c
478 * cgraph.h (cgraph_local_node_p): Declare.
479 * ipa-visibility.c: New file.
480 * Makefile.in (OBJS): Add ipa-visiblity.o
482 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
484 * gimple-fold.c (can_refer_decl_in_current_unit_p): Be sure
485 that var decl is available.
487 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
489 * tree-core.h (tree_decl_with_vis): Replace comdat_group by
491 * tree.c (copy_node_stat): Be sure to not copy symtab_node pointer.
492 (find_decls_types_r): Do not walk COMDAT_GROUP.
493 * tree.h (DECL_COMDAT_GROUP): Revamp to use decl_comdat_group.
494 * varasm.c (make_decl_one_only): Use set_comdat_group;
495 create node if needed.
496 * ipa-inline-transform.c (save_inline_function_body): Update
497 way we decl->symtab mapping.
498 * symtab.c (symtab_hash, hash_node, eq_node
499 symtab_insert_node_to_hashtable): Remove.
500 (symtab_register_node): Update.
501 (symtab_unregister_node): Update.
502 (symtab_get_node): Reimplement as inline function.
503 (symtab_add_to_same_comdat_group): Update.
504 (symtab_dissolve_same_comdat_group_list): Update.
505 (dump_symtab_base): Update.
506 (verify_symtab_base): Update.
507 (symtab_make_decl_local): Update.
508 (fixup_same_cpp_alias_visibility): Update.
509 (symtab_nonoverwritable_alias): Update.
510 * cgraphclones.c (set_new_clone_decl_and_node_flags): Update.
511 * ipa.c (update_visibility_by_resolution_info): UPdate.
512 * bb-reorder.c: Include cgraph.h
513 * lto-streamer-out.c (DFS_write_tree_body, hash_tree): Do not deal
515 * ipa-comdats.c (set_comdat_group, ipa_comdats): Update.
516 * cgraph.c (cgraph_get_create_node): Update.
517 * cgraph.h (struct symtab_node): Add get_comdat_group, set_comdat_group
519 (symtab_get_node): Make inline.
520 (symtab_insert_node_to_hashtable): Remove.
521 (symtab_can_be_discarded): Update.
522 (decl_comdat_group): New function.
523 * tree-streamer-in.c (lto_input_ts_decl_with_vis_tree_pointers):
525 * lto-cgraph.c (lto_output_node, lto_output_varpool_node): Stream out
527 (read_comdat_group): New function.
528 (input_node, input_varpool_node): Use it.
529 * trans-mem.c (ipa_tm_create_version_alias): Update code creating
531 * mips.c (mips_start_unique_function): Likewise.
532 (ix86_code_end): Likewise.
533 (rs6000_code_end): Likweise.
534 * tree-streamer-out.c (DECL_COMDAT_GROUP): Do not stream comdat group.
536 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
538 * gengtype-state.c (fatal_reading_state): Bring offline.
539 * optabs.c (widening_optab_handler): Bring offline.
540 * optabs.h (widening_optab_handler): Likewise.
541 * final.c (get_attr_length_1): Likewise.
543 2014-05-23 Jan Hubicka <hubicka@ucw.cz>
545 * sched-int.h (sd_iterator_cond): Manually tail recurse.
547 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
549 * config/rs6000/440.md (ppc440-integer): Include shift without dot.
550 (ppc440-compare): Include shift with dot.
551 * config/rs6000/e300c2c3.md (ppce300c3_iu): Include shift without dot.
552 * config/rs6000/e5500.md (e5500_sfx2): Include constant shift
554 * config/rs6000/e6500.md (e6500_sfx): Exclude constant shift
556 (e6500_sfx2): Include it.
557 * config/rs6000/rs6000.md ( *zero_extend<mode>di2_internal1,
558 *zero_extend<mode>di2_internal2, *zero_extend<mode>di2_internal3,
559 *zero_extendsidi2_lfiwzx, andsi3_mc, andsi3_nomc,
560 andsi3_internal0_nomc, extzvsi_internal, extzvdi_internal,
561 *extzvdi_internal1, *extzvdi_internal2, rotlsi3, *rotlsi3_64,
562 *rotlsi3_internal4, *rotlsi3_internal7le, *rotlsi3_internal7be,
563 *rotlsi3_internal10le, *rotlsi3_internal10be, rlwinm,
564 *lshiftrt_internal1le, *lshiftrt_internal1be,
565 *lshiftrt_internal4le, *lshiftrt_internal4be, rotldi3,
566 *rotldi3_internal4, *rotldi3_internal7le, *rotldi3_internal7be,
567 *rotldi3_internal10le, *rotldi3_internal10be,
568 *rotldi3_internal13le, *rotldi3_internal13be, *ashldi3_internal4,
569 ashldi3_internal5, *ashldi3_internal6, *ashldi3_internal7,
570 ashldi3_internal8, *ashldi3_internal9, anddi3_mc, anddi3_nomc,
571 *anddi3_internal2_mc, *anddi3_internal3_mc, and 4 anonymous
572 define_insns): Use type "shift" in the appropriate alternatives.
574 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
576 * config/rs6000/rs6000.md (type): Add "logical". Delete
578 (dot): Adjust comment.
579 (andsi3_mc, *andsi3_internal2_mc, *andsi3_internal3_mc,
580 *andsi3_internal4, *andsi3_internal5_mc, *boolsi3_internal2,
581 *boolsi3_internal3, *boolccsi3_internal2, *boolccsi3_internal3,
582 anddi3_mc, *anddi3_internal2_mc, *anddi3_internal3_mc,
583 *booldi3_internal2, *booldi3_internal3, *boolcdi3_internal2,
584 *boolcdi3_internal3, *boolccdi3_internal2, *boolccdi3_internal3,
585 *mov<mode>_internal2, and 10 anonymous define_insns): Use "logical".
586 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
588 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
589 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
590 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
591 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
592 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
593 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
594 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
595 * config/rs6000/8540.md (ppc8540_su): Adjust.
596 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
597 cell-cmp-microcoded): Adjust.
598 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
599 * config/rs6000/e500mc.md (e500mc_su): Adjust.
600 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
601 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
602 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
603 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
604 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
605 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
606 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
608 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
609 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare):
610 Adjust. Adjust comment.
611 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
612 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
614 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
616 * config/rs6000/rs6000.md (type): Add "add".
617 (*add<mode>3_internal1, addsi3_high, *add<mode>3_internal2,
618 *add<mode>3_internal3, *neg<mode>2_internal, and 5 anonymous
619 define_insns): Use it.
620 * config/rs6000/rs6000.c (rs6000_adjust_cost): Adjust.
622 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
623 * config/rs6000/440.md (ppc440-integer, ppc440-compare): Adjust.
624 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
625 * config/rs6000/601.md (ppc601-integer): Adjust.
626 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
627 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
628 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
629 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
630 * config/rs6000/8540.md (ppc8540_su): Adjust.
631 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
632 cell-cmp-microcoded): Adjust.
633 * config/rs6000/e300c2c3.md (ppce300c3_cmp, ppce300c3_iu): Adjust.
634 * config/rs6000/e500mc.md (e500mc_su): Adjust.
635 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2): Adjust.
636 * config/rs6000/e5500.md (e5500_sfx, e5500_sfx2): Adjust.
637 * config/rs6000/e6500.md (e6500_sfx, e6500_sfx2): Adjust.
638 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
639 * config/rs6000/power4.md (power4-integer, power4-cmp): Adjust.
640 * config/rs6000/power5.md (power5-integer, power5-cmp): Adjust.
641 * config/rs6000/power6.md (power6-integer, power6-fast-compare):
643 * config/rs6000/power7.md (power7-integer, power7-cmp): Adjust.
644 * config/rs6000/power8.md (power8-1cyc, power8-fast-compare): Adjust.
645 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
646 * config/rs6000/titan.md (titan_fxu_adder, titan_fxu_alu): Adjust.
648 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
650 * config/rs6000/rs6000.md (type): Delete "var_shift_rotate",
651 "delayed_compare", "var_delayed_compare".
652 (var_shift): New attribute.
653 (cell_micro): Adjust.
654 (*andsi3_internal2_mc, *andsi3_internal3_mc, *andsi3_internal4,
655 *andsi3_internal5_mc, *extzvsi_internal1, *extzvsi_internal2,
656 rotlsi3, *rotlsi3_64, *rotlsi3_internal2, *rotlsi3_internal3,
657 *rotlsi3_internal4, *rotlsi3_internal5, *rotlsi3_internal6,
658 *rotlsi3_internal8le, *rotlsi3_internal8be, *rotlsi3_internal9le,
659 *rotlsi3_internal9be, *rotlsi3_internal10le, *rotlsi3_internal10be,
660 *rotlsi3_internal11le, *rotlsi3_internal11be, *rotlsi3_internal12le,
661 *rotlsi3_internal12be, ashlsi3, *ashlsi3_64, lshrsi3, *lshrsi3_64,
662 *lshiftrt_internal2le, *lshiftrt_internal2be, *lshiftrt_internal3le,
663 *lshiftrt_internal3be, *lshiftrt_internal5le, *lshiftrt_internal5be,
664 *lshiftrt_internal5le, *lshiftrt_internal5be, ashrsi3, *ashrsi3_64,
665 rotldi3, *rotldi3_internal2, *rotldi3_internal3, *rotldi3_internal4,
666 *rotldi3_internal5, *rotldi3_internal6, *rotldi3_internal7le,
667 *rotldi3_internal7be, *rotldi3_internal8le, *rotldi3_internal8be,
668 *rotldi3_internal9le, *rotldi3_internal9be, *rotldi3_internal10le,
669 *rotldi3_internal10be, *rotldi3_internal11le, *rotldi3_internal11be,
670 *rotldi3_internal12le, *rotldi3_internal12be, *rotldi3_internal13le,
671 *rotldi3_internal13be, *rotldi3_internal14le, *rotldi3_internal14be,
672 *rotldi3_internal15le, *rotldi3_internal15be, *ashldi3_internal1,
673 *ashldi3_internal2, *ashldi3_internal3, *lshrdi3_internal1,
674 *lshrdi3_internal2, *lshrdi3_internal3, *ashrdi3_internal1,
675 *ashrdi3_internal2, *ashrdi3_internal3, *anddi3_internal2_mc,
676 *anddi3_internal3_mc, as well as 11 anonymous define_insns): Adjust.
677 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
678 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
680 * config/rs6000/40x.md (ppc403-integer, ppc403-compare): Adjust.
681 * config/rs6000/440.md (ppc440-integer): Adjust.
682 * config/rs6000/476.md (ppc476-simple-integer, ppc476-compare): Adjust.
683 * config/rs6000/601.md (ppc601-integer, ppc601-compare): Adjust.
684 * config/rs6000/603.md (ppc603-integer, ppc603-compare): Adjust.
685 * config/rs6000/6xx.md (ppc604-integer, ppc604-compare): Adjust.
686 * config/rs6000/7450.md (ppc7450-integer, ppc7450-compare): Adjust.
687 * config/rs6000/7xx.md (ppc750-integer, ppc750-compare): Adjust.
688 * config/rs6000/8540.md (ppc8540_su): Adjust.
689 * config/rs6000/cell.md (cell-integer, cell-fast-cmp,
690 cell-cmp-microcoded): Adjust.
691 * config/rs6000/e300c2c3.md (ppce300c3_cmp): Adjust.
692 * config/rs6000/e500mc.md (e500mc_su): Adjust.
693 * config/rs6000/e500mc64.md (e500mc64_su, e500mc64_su2,
694 e500mc64_delayed): Adjust.
695 * config/rs6000/e5500.md (e5500_sfx, e5500_delayed): Adjust.
696 * config/rs6000/e6500.md (e6500_sfx, e6500_delayed): Adjust.
697 * config/rs6000/mpc.md (mpccore-integer, mpccore-compare): Adjust.
698 * config/rs6000/power4.md (power4-integer, power4-compare): Adjust.
699 * config/rs6000/power5.md (power5-integer, power5-compare): Adjust.
700 * config/rs6000/power6.md (power6-shift, power6-var-rotate,
701 power6-delayed-compare, power6-var-delayed-compare): Adjust.
702 * config/rs6000/power7.md (power7-integer, power7-compare): Adjust.
703 * config/rs6000/power8.md (power8-1cyc, power8-compare): Adjust.
705 * config/rs6000/rs64.md (rs64a-integer, rs64a-compare): Adjust.
706 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
708 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
710 * config/rs6000/rs6000.md (type): Delete "idiv", "ldiv". Add "div".
711 (bits): New mode_attr.
712 (idiv_ldiv): Delete mode_attr.
713 (udiv<mode>3, *div<mode>3, div<div_extend>_<mode>): Adjust.
714 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
715 rs6000_adjust_priority, is_nonpipeline_insn,
716 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
718 * config/rs6000/40x.md (ppc403-idiv): Adjust.
719 * config/rs6000/440.md (ppc440-idiv): Adjust.
720 * config/rs6000/476.md (ppc476-idiv): Adjust.
721 * config/rs6000/601.md (ppc601-idiv): Adjust.
722 * config/rs6000/603.md (ppc603-idiv): Adjust.
723 * config/rs6000/6xx.md (ppc604-idiv, ppc620-idiv, ppc630-idiv,
724 ppc620-ldiv): Adjust.
725 * config/rs6000/7450.md (ppc7450-idiv): Adjust.
726 * config/rs6000/7xx.md (ppc750-idiv): Adjust.
727 * config/rs6000/8540.md (ppc8540_divide): Adjust.
728 * config/rs6000/a2.md (ppca2-idiv, ppca2-ldiv): Adjust.
729 * config/rs6000/cell.md (cell-idiv, cell-ldiv): Adjust.
730 * config/rs6000/e300c2c3.md (ppce300c3_divide): Adjust.
731 * config/rs6000/e500mc.md (e500mc_divide): Adjust.
732 * config/rs6000/e500mc64.md (e500mc64_divide): Adjust.
733 * config/rs6000/e5500.md (e5500_divide, e5500_divide_d): Adjust.
734 * config/rs6000/e6500.md (e6500_divide, e6500_divide_d): Adjust.
735 * config/rs6000/mpc.md (mpccore-idiv): Adjust.
736 * config/rs6000/power4.md (power4-idiv, power4-ldiv): Adjust.
737 * config/rs6000/power5.md (power5-idiv, power5-ldiv): Adjust.
738 * config/rs6000/power6.md (power6-idiv, power6-ldiv): Adjust.
739 * config/rs6000/power7.md (power7-idiv, power7-ldiv): Adjust.
740 * config/rs6000/power8.md (power8-idiv, power8-ldiv): Adjust.
741 * config/rs6000/rs64.md (rs64a-idiv, rs64a-ldiv): Adjust.
742 * config/rs6000/titan.md (titan_fxu_div): Adjust.
744 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
746 * config/rs6000/rs6000.md (type): Delete "insert_word",
747 "insert_dword". Add "insert".
748 (size): Update comment.
749 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
750 insn_must_be_first_in_group): Adjust.
751 (insvsi_internal, *insvsi_internal1, *insvsi_internal2,
752 *insvsi_internal3, *insvsi_internal4, *insvsi_internal5,
753 *insvsi_internal6, insvdi_internal): Adjust.
755 * config/rs6000/40x.md (ppc403-integer): Adjust.
756 * config/rs6000/440.md (ppc440-integer): Adjust.
757 * config/rs6000/476.md (ppc476-simple-integer): Adjust.
758 * config/rs6000/601.md (ppc601-integer): Adjust.
759 * config/rs6000/603.md (ppc603-integer): Adjust.
760 * config/rs6000/6xx.md (ppc604-integer): Adjust.
761 * config/rs6000/7450.md (ppc7450-integer): Adjust.
762 * config/rs6000/7xx.md (ppc750-integer): Adjust.
763 * config/rs6000/8540.md (ppc8540_su): Adjust.
764 * config/rs6000/cell.md (cell-integer, cell-insert): Adjust.
765 * config/rs6000/e300c2c3.md (ppce300c3_iu): Adjust.
766 * config/rs6000/e500mc.md (e500mc_su): Adjust.
767 * config/rs6000/e500mc64.md (e500mc64_su): Adjust.
768 * config/rs6000/e5500.md (e5500_sfx): Adjust.
769 * config/rs6000/e6500.md (e6500_sfx): Adjust.
770 * config/rs6000/mpc.md (mpccore-integer): Adjust.
771 * config/rs6000/power4.md (power4-integer, power4-insert): Adjust.
772 * config/rs6000/power5.md (power5-integer, power5-insert): Adjust.
773 * config/rs6000/power6.md (power6-insert, power6-insert-dword): Adjust.
774 * config/rs6000/power7.md (power7-integer): Adjust.
775 * config/rs6000/power8.md (power8-1cyc): Adjust.
776 * config/rs6000/rs64.md (rs64a-integer): Adjust.
777 * config/rs6000/titan.md (titan_fxu_shift_and_rotate): Adjust.
779 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
781 * config/rs6000/rs6000.md (type): Add "mul". Delete "imul",
782 "imul2", "imul3", "lmul", "imul_compare", "lmul_compare".
783 (size): New attribute.
784 (dot): New attribute.
785 (cell_micro): Adjust.
786 (mulsi3, *mulsi3_internal1, *mulsi3_internal2, mulsidi3,
787 umulsidi3, smulsi3_highpart, umulsi3_highpart, muldi3,
788 *muldi3_internal1, *muldi3_internal2, smuldi3_highpart,
789 umuldi3_highpart): Adjust.
790 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_cracked_insn,
791 rs6000_adjust_priority, is_nonpipeline_insn,
792 insn_must_be_first_in_group, insn_must_be_last_in_group): Adjust.
794 * config/rs6000/40x.md (ppc403-imul, ppc405-imul, ppc405-imul2,
795 ppc405-imul3): Adjust.
796 * config/rs6000/440.md (ppc440-imul, ppc440-imul2): Adjust.
797 * config/rs6000/476.md (ppc476-imul): Adjust.
798 * config/rs6000/601.md (ppc601-imul): Adjust.
799 * config/rs6000/603.md (ppc603-imul, ppc603-imul2): Adjust.
800 * config/rs6000/6xx.md (ppc604-imul, ppc604e-imul, ppc620-imul,
801 ppc620-imul2, ppc620-imul3, ppc620-lmul): Adjust.
802 * config/rs6000/7450.md (ppc7450-imul, ppc7450-imul2): Adjust.
803 * config/rs6000/7xx.md (ppc750-imul, ppc750-imul2, ppc750-imul3):
805 * config/rs6000/8540.md (ppc8540_multiply): Adjust.
806 * config/rs6000/a2.md (ppca2-imul, ppca2-lmul): Adjust.
807 * config/rs6000/cell.md (cell-lmul, cell-lmul-cmp, cell-imul23,
809 * config/rs6000/e300c2c3.md (ppce300c3_multiply): Adjust.
810 * config/rs6000/e500mc.md (e500mc_multiply): Adjust.
811 * config/rs6000/e500mc64.md (e500mc64_multiply): Adjust.
812 * config/rs6000/e5500.md (e5500_multiply, e5500_multiply_i): Adjust.
813 * config/rs6000/e6500.md (e6500_multiply, e6500_multiply_i): Adjust.
814 * config/rs6000/mpc.md (mpccore-imul): Adjust.
815 * config/rs6000/power4.md (power4-lmul-cmp, power4-imul-cmp,
816 power4-lmul, power4-imul, power4-imul3): Adjust.
817 * config/rs6000/power5.md (power5-lmul-cmp, power5-imul-cmp,
818 power5-lmul, power5-imul, power5-imul3): Adjust.
819 * config/rs6000/power6.md (power6-lmul-cmp, power6-imul-cmp,
820 power6-lmul, power6-imul, power6-imul3): Adjust.
821 * config/rs6000/power7.md (power7-mul, power7-mul-compare): Adjust.
822 * config/rs6000/power8.md (power8-mul, power8-mul-compare): Adjust.
824 * config/rs6000/rs64.md (rs64a-imul, rs64a-imul2, rs64a-imul3,
826 * config/rs6000/titan.md (titan_imul): Adjust.
828 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
830 * config/rs6000/rs6000.md (type): Add new value "halfmul".
831 (*macchwc, *macchw, *macchwuc, *macchwu, *machhwc, *machhw,
832 *machhwuc, *machhwu, *maclhwc, *maclhw, *maclhwuc, *maclhwu,
833 *nmacchwc, *nmacchw, *nmachhwc, *nmachhw, *nmaclhwc, *nmaclhw,
834 *mulchwc, *mulchw, *mulchwuc, *mulchwu, *mulhhwc, *mulhhw,
835 *mulhhwuc, *mulhhwu, *mullhwc, *mullhw, *mullhwuc, *mullhwu): Use it.
836 * config/rs6000/40x.md (ppc405-imul3): Add type halfmul.
837 * config/rs6000/440.md (ppc440-imul2): Add type halfmul.
838 * config/rs6000/476.md (ppc476-imul): Add type halfmul.
839 * config/rs6000/titan.md: Delete nonsensical comment.
840 (titan_imul): Add type imul3.
841 (titan_mulhw): Remove type imul3; add type halfmul.
843 2014-05-23 Segher Boessenkool <segher@kernel.crashing.org>
845 * config/rs6000/rs6000.md (type): Reorder, reformat.
847 2014-05-23 Martin Jambor <mjambor@suse.cz>
849 PR tree-optimization/53787
850 * params.def (PARAM_IPA_MAX_AA_STEPS): New param.
851 * ipa-prop.h (ipa_node_params): Rename uses_analysis_done to
852 analysis_done, update all uses.
853 * ipa-prop.c: Include domwalk.h
854 (param_analysis_info): Removed.
855 (param_aa_status): New type.
856 (ipa_bb_info): Likewise.
857 (func_body_info): Likewise.
858 (ipa_get_bb_info): New function.
859 (aa_overwalked): Likewise.
860 (find_dominating_aa_status): Likewise.
861 (parm_bb_aa_status_for_bb): Likewise.
862 (parm_preserved_before_stmt_p): Changed to use new param AA info.
863 (load_from_unmodified_param): Accept func_body_info as a parameter
864 instead of parms_ainfo.
865 (parm_ref_data_preserved_p): Changed to use new param AA info.
866 (parm_ref_data_pass_through_p): Likewise.
867 (ipa_load_from_parm_agg_1): Likewise. Update callers.
868 (compute_complex_assign_jump_func): Changed to use new param AA info.
869 (compute_complex_ancestor_jump_func): Likewise.
870 (ipa_compute_jump_functions_for_edge): Likewise.
871 (ipa_compute_jump_functions): Removed.
872 (ipa_compute_jump_functions_for_bb): New function.
873 (ipa_analyze_indirect_call_uses): Likewise, moved variable
875 (ipa_analyze_virtual_call_uses): Accept func_body_info instead of node
876 and info, moved variable declarations down.
877 (ipa_analyze_call_uses): Accept and pass on func_body_info instead of
879 (ipa_analyze_stmt_uses): Likewise.
880 (ipa_analyze_params_uses): Removed.
881 (ipa_analyze_params_uses_in_bb): New function.
882 (ipa_analyze_controlled_uses): Likewise.
883 (free_ipa_bb_info): Likewise.
884 (analysis_dom_walker): New class.
885 (ipa_analyze_node): Handle node-specific forbidden analysis,
886 initialize and free func_body_info, use dominator walker.
887 (ipcp_modif_dom_walker): New class.
888 (ipcp_transform_function): Create and free func_body_info, use
889 ipcp_modif_dom_walker, moved a lot of functionality there.
891 2014-05-23 Marek Polacek <polacek@redhat.com>
892 Jakub Jelinek <jakub@redhat.com>
894 * builtins.def: Change SANITIZE_FLOAT_DIVIDE to SANITIZE_NONDEFAULT.
895 * gcc.c (sanitize_spec_function): Likewise.
896 * convert.c (convert_to_integer): Include "ubsan.h". Add
897 floating-point to integer instrumentation.
898 * doc/invoke.texi: Document -fsanitize=float-cast-overflow.
899 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_CAST and
901 * opts.c (common_handle_option): Handle -fsanitize=float-cast-overflow.
902 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW,
903 BUILT_IN_UBSAN_HANDLE_FLOAT_CAST_OVERFLOW_ABORT): Add.
904 * ubsan.c: Include "realmpfr.h" and "dfp.h".
905 (get_ubsan_type_info_for_type): Handle REAL_TYPEs.
906 (ubsan_type_descriptor): Set tkind to 0xffff for types other than
907 float/double/long double.
908 (ubsan_instrument_float_cast): New function.
909 * ubsan.h (ubsan_instrument_float_cast): Declare.
911 2014-05-23 Jiong Wang <jiong.wang@arm.com>
913 * config/aarch64/predicates.md (aarch64_call_insn_operand): New
915 * config/aarch64/constraints.md ("Ucs", "Usf"): New constraints.
916 * config/aarch64/aarch64.md (*sibcall_insn, *sibcall_value_insn):
917 Adjust for tailcalling through registers.
918 * config/aarch64/aarch64.h (enum reg_class): New caller save
920 (REG_CLASS_NAMES): Likewise.
921 (REG_CLASS_CONTENTS): Likewise.
922 * config/aarch64/aarch64.c (aarch64_function_ok_for_sibcall):
923 Allow tailcalling without decls.
925 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
927 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
928 Rewrite check for ORT_PARALLEL and ORT_COMBINED_PARALLEL.
930 * omp-low.c (expand_omp_for_static_chunk): Rename variable si to
931 gsi, and variables v_* to v*.
933 2014-05-23 Eric Botcazou <ebotcazou@adacore.com>
935 * varasm.c (output_constructor_bitfield): Fix thinkos in latest change.
937 2014-05-23 Thomas Schwinge <thomas@codesourcery.com>
939 * gimple.h (enum gf_mask): Add and use GF_OMP_FOR_SIMD.
940 * omp-low.c: Update accordingly.
942 * gimple.h (enum gf_mask): Rewrite "<< 0" shift expressions used
943 for GF_OMP_FOR_KIND_MASK, GF_OMP_FOR_KIND_FOR,
944 GF_OMP_FOR_KIND_DISTRIBUTE, GF_OMP_FOR_KIND_SIMD,
945 GF_OMP_FOR_KIND_CILKSIMD, GF_OMP_TARGET_KIND_MASK,
946 GF_OMP_TARGET_KIND_REGION, GF_OMP_TARGET_KIND_DATA,
947 GF_OMP_TARGET_KIND_UPDATE.
949 * gimplify.c (omp_notice_variable) <case OMP_CLAUSE_DEFAULT_NONE>:
950 Explicitly enumerate the expected region types.
952 2014-05-23 Paul Eggert <eggert@cs.ucla.edu>
955 * doc/extend.texi (Function Attributes): Fix __attribute__ ((malloc))
956 documentation; the old documentation didn't clearly state the
957 constraints on the contents of the pointed-to storage.
959 2014-05-23 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
961 Fix bootstrap error on ia64
962 * config/ia64/ia64.c (ia64_first_cycle_multipass_dfa_lookahead_guard):
963 Return default value.
965 2014-05-23 Thomas Preud'homme <thomas.preudhomme@arm.com>
967 PR tree-optimization/54733
968 * tree-ssa-math-opts.c (nop_stats): New "bswap_stats" structure.
970 (find_bswap_or_nop_load): New.
971 (find_bswap_1): Renamed to ...
972 (find_bswap_or_nop_1): This. Also add support for memory source.
973 (find_bswap): Renamed to ...
974 (find_bswap_or_nop): This. Also add support for memory source and
975 detection of bitwise operations equivalent to load in host endianness.
976 (execute_optimize_bswap): Likewise. Also move its leading comment back
977 in place and split statement transformation into ...
978 (bswap_replace): This.
980 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
982 PR rtl-optimization/61215
983 * lra-elelimination.c (lra_eliminate_regs_1): Don't use
984 simplify_gen_subreg until final substitution.
986 2014-05-23 Alan Modra <amodra@gmail.com>
989 * config/rs6000/rs6000.c (mem_operand_gpr): Handle SImode.
990 * config/rs6000/rs6000.md (extendsidi2_lfiwax, extendsidi2_nocell):
991 Use "Y" constraint rather than "m".
993 2014-05-23 Kugan Vivekanandarajah <kuganv@linaro.org>
995 * config/aarch64/aarch64.c (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New
997 * config/aarch64/aarch64-protos.h (aarch64_atomic_assign_expand_fenv):
998 New function declaration.
999 * config/aarch64/aarch64-builtins.c (aarch64_builtins) : Add
1000 AARCH64_BUILTIN_GET_FPCR, AARCH64_BUILTIN_SET_FPCR.
1001 AARCH64_BUILTIN_GET_FPSR and AARCH64_BUILTIN_SET_FPSR.
1002 (aarch64_init_builtins) : Initialize builtins
1003 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1004 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1005 (aarch64_expand_builtin) : Expand builtins __builtins_aarch64_set_fpcr
1006 __builtins_aarch64_get_fpcr, __builtins_aarch64_get_fpsr,
1007 and __builtins_aarch64_set_fpsr.
1008 (aarch64_atomic_assign_expand_fenv): New function.
1009 * config/aarch64/aarch64.md (set_fpcr): New pattern.
1010 (get_fpcr) : Likewise.
1011 (set_fpsr) : Likewise.
1012 (get_fpsr) : Likewise.
1013 (unspecv): Add UNSPECV_GET_FPCR and UNSPECV_SET_FPCR, UNSPECV_GET_FPSR
1014 and UNSPECV_SET_FPSR.
1015 * doc/extend.texi (AARCH64 Built-in Functions) : Document
1016 __builtins_aarch64_set_fpcr, __builtins_aarch64_get_fpcr.
1017 __builtins_aarch64_set_fpsr and __builtins_aarch64_get_fpsr.
1019 2014-05-22 Vladimir Makarov <vmakarov@redhat.com>
1021 PR rtl-optimization/60969
1022 * ira-costs.c (record_reg_classes): Process NO_REGS for matching
1023 constraints. Set up mem cost for NO_REGS case.
1025 2014-05-22 Thomas Schwinge <thomas@codesourcery.com>
1027 * builtin-types.def: Simplify examples for DEF_FUNCTION_TYPE_*.
1029 2012-05-22 Bernd Schmidt <bernds@codesourcery.com>
1031 * config/darwin.c: Include "lto-section-names.h".
1032 (LTO_SEGMENT_NAME): Don't define.
1033 * config/i386/winnt.c: Include "lto-section-names.h".
1034 * lto-streamer.c: Include "lto-section-names.h".
1035 * lto-streamer.h (LTO_SECTION_NAME_PREFIX): Don't define.
1036 * lto-wrapper.c: Include "lto-section-names.h".
1037 (LTO_SECTION_NAME_PREFIX): Don't define.
1038 * lto-section-names.h: New file.
1039 * cgraphunit.c: Include "lto-section-names.h".
1041 2014-05-22 Peter Bergner <bergner@vnet.ibm.com>
1043 * config/rs6000/htm.md (ttest): Use correct shift value to get CR0.
1045 2014-05-22 Richard Earnshaw <rearnsha@arm.com>
1048 * arm.md (arm_cmpdi_unsigned): Fix length calculation for Thumb2.
1050 2014-05-22 Nick Clifton <nickc@redhat.com>
1052 * config/msp430/msp430.h (ASM_SPEC): Add spaces after inserted options.
1054 2014-05-22 Eric Botcazou <ebotcazou@adacore.com>
1056 * tree-ssa-forwprop.c (associate_plusminus): Extend (T)(P + A) - (T)P
1057 -> (T)A transformation to integer types.
1059 2014-05-22 Teresa Johnson <tejohnson@google.com>
1061 * gcov-io.c (gcov_position): Use gcov_nonruntime_assert.
1062 (gcov_is_error): Remove gcc_assert from IN_LIBGCOV code.
1063 (gcov_rewrite): Use gcov_nonruntime_assert.
1065 (gcov_write_words): Ditto.
1066 (gcov_write_length): Ditto.
1067 (gcov_read_words): Use gcov_nonruntime_assert, and remove
1068 gcc_assert from IN_LIBGCOV code.
1069 (gcov_read_summary): Use gcov_error to flag profile corruption.
1070 (gcov_sync): Use gcov_nonruntime_assert.
1071 (gcov_seek): Remove gcc_assert from IN_LIBGCOV code.
1072 (gcov_histo_index): Use gcov_nonruntime_assert.
1073 (static void gcov_histogram_merge): Ditto.
1074 (compute_working_sets): Ditto.
1075 * gcov-io.h (gcov_nonruntime_assert): Define.
1076 (gcov_error): Define for !IN_LIBGCOV
1078 2014-05-22 Richard Biener <rguenther@suse.de>
1080 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
1081 BUILT_IN_REALLOC like BUILT_IN_STRDUP.
1082 (call_may_clobber_ref_p_1): Handle BUILT_IN_REALLOC as allocation
1083 and deallocation site.
1084 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
1085 Handle BUILT_IN_REALLOC similar to BUILT_IN_STRDUP with also
1086 passing through the incoming points-to set.
1087 (handle_lhs_call): Use flags argument instead of recomputing it.
1088 (find_func_aliases_for_call): Call handle_lhs_call with proper
1091 2014-05-22 Jakub Jelinek <jakub@redhat.com>
1093 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Make sure
1094 all padding bits in REAL_VALUE_TYPE are cleared.
1096 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1098 Cleanup and improve multipass_dfa_lookahead_guard
1099 * config/i386/i386.c (core2i7_first_cycle_multipass_filter_ready_try,)
1100 (core2i7_first_cycle_multipass_begin,)
1101 (core2i7_first_cycle_multipass_issue,)
1102 (core2i7_first_cycle_multipass_backtrack): Update signature.
1103 * config/ia64/ia64.c
1104 (ia64_first_cycle_multipass_dfa_lookahead_guard_spec): Remove.
1105 (ia64_first_cycle_multipass_dfa_lookahead_guard): Update signature.
1106 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove
1108 (ia64_first_cycle_multipass_dfa_lookahead_guard): Merge logic from
1109 ia64_first_cycle_multipass_dfa_lookahead_guard_spec. Update return
1111 * config/rs6000/rs6000.c (rs6000_use_sched_lookahead_guard): Update
1113 * doc/tm.texi: Regenerate.
1115 (TARGET_SCHED_FIRST_CYCLE_MULTIPASS_DFA_LOOKAHEAD_GUARD_SPEC): Remove.
1116 * haifa-sched.c (ready_try): Make signed to allow negative values.
1117 (rebug_ready_list_1): Update.
1118 (choose_ready): Simplify.
1119 (sched_extend_ready_list): Update.
1121 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1123 Remove IA64 speculation tweaking flags
1124 * config/ia64/ia64.c (ia64_set_sched_flags): Delete handling of
1125 speculation tuning flags.
1126 (msched-prefer-non-data-spec-insns,)
1127 (msched-prefer-non-control-spec-insns): Obsolete options.
1128 * haifa-sched.c (choose_ready): Remove handling of
1129 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1130 * sched-int.h (enum SPEC_SCHED_FLAGS): Remove PREFER_NON_CONTROL_SPEC
1131 and PREFER_NON_DATA_SPEC.
1132 * sel-sched.c (process_spec_exprs): Remove handling of
1133 PREFER_NON_CONTROL_SPEC and PREFER_NON_DATA_SPEC.
1135 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1137 Improve scheduling debug output
1138 * haifa-sched.c (debug_ready_list): Remove unnecessary prototype.
1139 (advance_one_cycle): Update.
1140 (schedule_insn, queue_to_ready): Add debug printouts.
1141 (debug_ready_list_1): New static function.
1142 (debug_ready_list): Update.
1143 (max_issue): Add debug printouts.
1144 (dump_insn_stream): New static function.
1145 (schedule_block): Use it. Also better indent printouts.
1147 2014-05-22 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
1149 Fix sched_insn debug counter
1150 * haifa-sched.c (schedule_insn): Update.
1151 (struct haifa_saved_data): Add nonscheduled_insns_begin.
1152 (save_backtrack_point, restore_backtrack_point): Update.
1153 (first_nonscheduled_insn): New static function.
1154 (queue_to_ready, choose_ready): Use it.
1155 (schedule_block): Init nonscheduled_insns_begin.
1156 (sched_emit_insn): Update.
1159 2014-05-22 Kugan Vivekanandarajah <kuganv@linaro.org>
1161 * config/aarch64/aarch64.c (aarch64_regno_regclass) : Change CORE_REGS
1163 (aarch64_secondary_reload) : LikeWise.
1164 (aarch64_class_max_nregs) : Remove CORE_REGS.
1165 * config/aarch64/aarch64.h (enum reg_class) : Remove CORE_REGS.
1166 (REG_CLASS_NAMES) : Likewise.
1167 (REG_CLASS_CONTENTS) : LikeWise.
1168 (INDEX_REG_CLASS) : Change CORE_REGS to GENERAL_REGS.
1170 2014-05-21 Guozhi Wei <carrot@google.com>
1173 * config/aarch64/arm_neon.h (vqdmulh_n_s16): Change the last operand's
1175 (vqdmulhq_n_s16): Likewise.
1177 2014-05-21 Segher Boessenkool <segher@kernel.crashing.org>
1179 * config/rs6000/predicates.md (update_indexed_address_mem): Delete.
1181 2014-05-21 Marek Polacek <polacek@redhat.com>
1184 * ubsan.c (is_ubsan_builtin_p): Turn assert into a condition.
1186 2014-05-21 Martin Jambor <mjambor@suse.cz>
1188 * doc/invoke.texi (Optimize Options): Document parameters
1189 ipa-cp-eval-threshold, ipa-max-agg-items, ipa-cp-loop-hint-bonus and
1190 ipa-cp-array-index-hint-bonus.
1192 2014-05-21 Mark Wielaard <mjw@redhat.com>
1195 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_type if DWARF
1196 version >= 3 or not strict DWARF.
1197 * langhooks.h (struct lang_hooks_for_types): Add
1198 enum_underlying_base_type.
1199 * langhooks.c (lhd_enum_underlying_base_type): New function.
1200 * gcc/langhooks.h (struct lang_hooks_for_types): Add
1201 enum_underlying_base_type.
1202 * langhooks-def.h (lhd_enum_underlying_base_type): New declaration.
1203 (LANG_HOOKS_ENUM_UNDERLYING_BASE_TYPE): New define.
1204 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add new lang hook.
1206 2014-05-21 Richard Biener <rguenther@suse.de>
1208 * doc/invoke.texi (-flto-partition=): Document one and none algorithms.
1210 2014-05-21 John Marino <gnugcc@marino.st>
1212 * config.gcc (*-*-dragonfly*): New target.
1213 * configure.ac: Detect dl_iterate_phdr (*freebsd*, *dragonfly*).
1214 * configure: Regenerate.
1215 * config/dragonfly-stdint.h: New.
1216 * config/dragonfly.h: New.
1217 * config/dragonfly.opt: New.
1218 * config/i386/dragonfly.h: New.
1219 * ginclude/stddef.h: Detect _PTRDIFF_T_DECLARED for DragonFly.
1221 2014-05-21 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1223 * tree.def (VOID_CST): New.
1224 * tree-core.h (TI_VOID): New.
1225 * tree.h (void_node): New.
1226 * tree.c (tree_node_structure_for_code, tree_code_size)
1227 (iterative_hash_expr): Handle VOID_CST.
1228 (build_common_tree_nodes): Initialize void_node.
1230 2014-05-21 Bernd Schmidt <bernds@codesourcery.com>
1232 * reload1.c (remove_init_insns, will_delete_init_insn_p): New static
1234 (reload, calculate_needs_all_insns, reload_as_needed): Use them.
1236 * config/bfin/bfin.c (split_load_immediate): Use gen_int_mode in a few
1239 * cfgrtl.c (cfg_layout_initialize): Weaken assert to only trigger if
1240 flag_reorder_blocks_and_partition.
1241 * hw-doloop.c (reorg_loops): Avoid reordering if that flag is set.
1243 2014-05-21 Oleg Endo <olegendo@gcc.gnu.org>
1246 * config/sh/sh.md (*addc_r_1): Rename to addc_t_r. Remove empty
1248 (*addc_r_t): Add new insn_and_split.
1250 2014-05-21 Jakub Jelinek <jakub@redhat.com>
1253 * omp-low.c (handle_simd_reference): New function.
1254 (lower_rec_input_clauses): Use it. Defer adding reference
1255 initialization even for reduction without placeholder if in simd,
1256 handle it properly later on.
1258 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1260 PR tree-optimization/60899
1261 * gimple-fold.c (can_refer_decl_in_current_unit_p): Cleanup;
1262 assume all static symbols will have definition wile parsing and
1263 check the do have definition later in compilation; check that
1264 variable referring symbol will be output before concluding that
1265 reference is safe; be conservative for referring local statics;
1266 be more precise about when comdat is output in other partition.
1268 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1271 * ipa-inline-transform.c (inline_call): Use add CALLEE_REMOVED
1273 * ipa-inline.c (inline_to_all_callers): If callee was removed; return.
1274 (ipa_inline): Loop inline_to_all_callers until no more aliases
1277 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1279 * ipa.c (ipa_discover_readonly_nonaddressable_var): Fix dumping;
1280 set writeonly flag only for vars actually written to.
1282 2014-05-20 Dehao Chen <dehao@google.com>
1284 * ipa-inline-transform.c (clone_inlined_nodes): Use min of edge count
1285 and callee count to get clone count.
1286 * tree-inline.c (expand_call_inline): Use callee count instead of bb
1289 2014-05-20 Richard Sandiford <rdsandiford@googlemail.com>
1291 PR rtl-optimization/61243
1292 * emit-rtl.c (emit_copy_of_insn_after): Copy CROSSING_JUMP_P.
1294 2014-05-20 Xinliang David Li <davidxl@google.com>
1296 * cgraphunit.c (walk_polymorphic_call_targets): Add
1297 dbgcnt and fopt-info support.
1298 * ipa-prop.c (ipa_make_edge_direct_to_target): Ditto.
1299 * ipa-devirt.c (ipa_devirt): Ditto.
1300 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Ditto.
1301 * ipa.c (walk_polymorphic_call_targets): Ditto.
1302 * gimple-fold.c (fold_gimple_assign): Ditto.
1303 (gimple_fold_call): Ditto.
1304 * dbgcnt.def: New counter.
1306 2014-05-20 DJ Delorie <dj@redhat.com>
1308 * config/msp430/msp430.md (split): Don't allow subregs when
1309 splitting SImode adds.
1310 (andneghi): Fix subtraction logic.
1311 * config/msp430/predicates.md (msp430_nonsubreg_or_imm_operand): New.
1313 2014-05-20 Jan Hubicka <hubicka@ucw.cz>
1315 * tree.h (DECL_ONE_ONLY): Return true only for externally visible
1317 * except.c (switch_to_exception_section, resolve_unique_section,
1318 get_named_text_section, default_function_rodata_section,
1319 align_variable, get_block_for_decl, default_section_type_flags):
1320 Use DECL_COMDAT_GROUP instead of DECL_ONE_ONLY.
1321 * symtab.c (symtab_add_to_same_comdat_group,
1322 symtab_make_decl_local, fixup_same_cpp_alias_visibility,
1323 symtab_nonoverwritable_alias, symtab_get_symbol_partitioning_class):
1325 * cgraphclones.c (cgraph_create_virtual_clone): Likewise.
1326 * bb-reorder.c (pass_partition_blocks::gate): Likewise.
1327 * config/c6x/c6x.c (c6x_elf_unique_section): Likewise.
1328 (c6x_function_in_section_p): Likewise.
1329 * config/darwin.c (machopic_select_section): Likewise.
1330 * config/arm/arm.c (arm_function_in_section_p): Likewise.
1331 * config/mips/mips.c (mips_function_rodata_section): Likewise.
1332 * config/mep/mep.c (mep_select_section): LIkewise.
1333 * config/i386/i386.c (x86_64_elf_unique_section): Likewise.
1335 2014-05-20 Eric Botcazou <ebotcazou@adacore.com>
1337 * tree-ssa-dom.c (hashable_expr_equal_p) <EXPR_CALL>: Also compare the
1338 EH region of calls to pure functions that can throw an exception.
1339 * tree-ssa-sccvn.c (vn_reference_eq): Remove duplicated test.
1340 (copy_reference_ops_from_call): Also copy the EH region of the call if
1341 it can throw an exception.
1343 2014-05-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1345 * simplify-rtx.c (simplify_binary_operation_1): Optimize case of
1346 nested VEC_SELECTs that are inverses of each other.
1348 2014-05-20 Richard Biener <rguenther@suse.de>
1350 * tree-ssa-sccvn.c (process_scc): Dump SCC here, when iterating,
1351 (extract_and_process_scc_for_name): not here.
1352 (cond_dom_walker::before_dom_children): Only process
1353 stmts that end the BB in interesting ways.
1354 (run_scc_vn): Mark param uses as visited.
1356 2014-05-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1358 * config/arm/arm.md (arith_shiftsi): Do not predicate for
1361 2014-05-20 Nick Clifton <nickc@redhat.com>
1363 * config/msp430/msp430.c (TARGET_GIMPLIFY_VA_ARG_EXPR): Define.
1364 (msp430_gimplify_va_arg_expr): New function.
1365 (msp430_print_operand): Handle (CONST (ZERO_EXTRACT)).
1367 * config/msp430/msp430.md (zero_extendpsisi2): Use + constraint on
1368 operand 0 in order to prevent confusion about the number of
1371 2014-05-20 Richard Biener <rguenther@suse.de>
1373 PR tree-optimization/61221
1374 * tree-ssa-pre.c (el_to_update): Remove.
1375 (eliminate_dom_walker::before_dom_children): Handle released
1376 VDEFs by value-numbering them to the associated VUSE. Update
1377 stmt immediately for substituted call address.
1378 (eliminate): Remove delayed stmt updating code.
1379 * tree-ssa-sccvn.c (vuse_ssa_val): New function valueizing
1380 possibly late re-numbered vuses.
1381 (vn_reference_lookup_2): Adjust.
1382 (vn_reference_lookup_pieces): Likewise.
1383 (vn_reference_lookup): Likewise.
1385 2014-05-20 Richard Biener <rguenther@suse.de>
1387 * config.gcc: Remove need_64bit_hwint.
1388 * configure.ac: Do not define NEED_64BIT_HOST_WIDE_INT.
1389 * hwint.h: Do not check NEED_64BIT_HOST_WIDE_INT but assume
1391 * config.in: Regenerate.
1392 * configure: Likewise.
1394 2014-05-19 David Wohlferd <dw@LimeGreenSocks.com>
1396 * doc/extend.texi: Create Label Attributes section,
1397 move all label attributes into it and reference it.
1399 2014-05-19 Richard Earnshaw <rearnsha@arm.com>
1401 * arm.c (thumb1_reorg): When scanning backwards skip anything
1402 that's not a proper insn.
1404 2014-05-19 Richard Biener <rguenther@suse.de>
1406 PR tree-optimization/61221
1407 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
1408 Do nothing for unreachable blocks.
1409 * tree-ssa-sccvn.c (cond_dom_walker::before_dom_children):
1410 Improve unreachability detection.
1412 2014-05-19 Richard Biener <rguenther@suse.de>
1414 PR tree-optimization/61209
1415 * tree-ssa-sccvn.c (visit_phi): Avoid setting expr to VN_TOP.
1417 2014-05-19 Nick Clifton <nickc@redhat.com>
1419 * except.c (init_eh): Fix computation of builtin setjmp buffer
1420 size to allow for targets where POINTER_SIZE > BITS_PER_WORD.
1422 2014-05-19 Richard Biener <rguenther@suse.de>
1424 PR tree-optimization/61184
1425 * tree-vrp.c (is_negative_overflow_infinity): Use
1426 TREE_OVERFLOW_P and do that check first.
1427 (is_positive_overflow_infinity): Likewise.
1428 (is_overflow_infinity): Likewise.
1429 (vrp_operand_equal_p): Properly treat operands with
1430 differing overflow as not equal.
1432 2014-05-19 Bernd Schmidt <bernds@codesourcery.com>
1434 * simplify-rtx.c (simplify_unary_operation_1): Use CONST_INT_P in
1435 shift simplification where it was intended.
1437 2014-05-19 Christian Bruel <christian.bruel@st.com>
1440 * config/sh/sh.md (movsf_ie): Unset fp_mode for fmov.
1442 2014-05-19 Richard Sandiford <r.sandiford@uk.ibm.com>
1445 * config/sparc/sparc.c (sparc_fold_builtin): Use widest_int rather
1448 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1450 * reg-notes.def (CROSSING_JUMP): Likewise.
1451 * rtl.h (rtx_def): Update comment for jump flag.
1452 (CROSSING_JUMP_P): Define.
1453 * cfgcleanup.c (try_forward_edges, try_optimize_cfg): Use it instead
1454 of a REG_CROSSING_JUMP note.
1455 * cfghooks.c (tidy_fallthru_edges): Likewise.
1456 * cfgrtl.c (fixup_partition_crossing, rtl_verify_edges): Likewise.
1457 * emit-rtl.c (try_split): Likewise.
1458 * haifa-sched.c (sched_create_recovery_edges): Likewise.
1459 * ifcvt.c (find_if_case_1, find_if_case_2): Likewise.
1460 * jump.c (redirect_jump_2): Likewise.
1461 * reorg.c (follow_jumps, fill_slots_from_thread): Likewise.
1462 (relax_delay_slots): Likewise.
1463 * config/arc/arc.md (jump_i, cbranchsi4_scratch, *bbit): Likewise.
1464 (bbit_di): Likewise.
1465 * config/arc/arc.c (arc_reorg, arc_can_follow_jump): Likewise.
1466 * config/sh/sh.md (jump_compact): Likewise.
1467 * bb-reorder.c (rotate_loop): Likewise.
1468 (pass_duplicate_computed_gotos::execute): Likewise.
1469 (add_reg_crossing_jump_notes): Rename to...
1470 (update_crossing_jump_flags): ...this.
1471 (pass_partition_blocks::execute): Update accordingly.
1473 2014-05-19 Richard Sandiford <rdsandiford@googlemail.com>
1475 * tree.h: Remove extraneous template <>.
1477 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1479 * ipa.c (symtab_remove_unreachable_nodes): Remove
1480 symbol from comdat group if its body was eliminated.
1481 (comdat_can_be_unshared_p_1): Static symbols can always be privatized.
1482 * symtab.c (symtab_remove_from_same_comdat_group): Break out from ...
1483 (symtab_unregister_node): ... this one.
1484 (verify_symtab_base): More strict checking of comdats.
1485 * cgraph.h (symtab_remove_from_same_comdat_group): Declare.
1487 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1489 * tree-pass.h (make_pass_ipa_comdats): New pass.
1490 * timevar.def (TV_IPA_COMDATS): New timevar.
1491 * passes.def (pass_ipa_comdats): Add.
1492 * Makefile.in (OBJS): Add ipa-comdats.o
1493 * ipa-comdats.c: New file.
1495 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1497 * ipa.c (update_visibility_by_resolution_info): New function.
1498 (function_and_variable_visibility): Use it.
1500 2014-05-17 Jan Hubicka <hubicka@ucw.cz>
1502 * cgraph.h (symtab_first_defined_symbol, symtab_next_defined_symbol):
1504 (FOR_EACH_DEFINED_SYMBOL): New macro.
1505 (varpool_first_static_initializer, varpool_next_static_initializer,
1506 varpool_first_defined_variable, varpool_next_defined_variable):
1508 (symtab_in_same_comdat_p): Correctly deal with inline functions.
1510 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1512 * ggc-page.c (ggc_handle_finalizers): Add comment.
1514 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1516 * ggc-common.c (ggc_internal_cleared_alloc): Adjust.
1517 * ggc-none.c (ggc_internal_alloc): Assert if a finalizer is passed.
1518 (ggc_internal_cleared_alloc): Likewise.
1519 * ggc-page.c (finalizer): New class.
1520 (vec_finalizer): Likewise.
1521 (globals::finalizers): New member.
1522 (globals::vec_finalizers): Likewise.
1523 (ggc_internal_alloc): Record the finalizer if any for the block being
1525 (ggc_handle_finalizers): New function.
1526 (ggc_collect): Call ggc_handle_finalizers.
1527 * ggc.h (ggc_internal_alloc): Add arguments to allow installing a
1529 (ggc_internal_cleared_alloc): Likewise.
1530 (finalize): New function.
1531 (need_finalization_p): Likewise.
1532 (ggc_alloc): Install the type's destructor as the finalizer if it
1534 (ggc_cleared_alloc): Likewise.
1535 (ggc_vec_alloc): Likewise.
1536 (ggc_cleared_vec_alloc): Likewise.
1538 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1540 * ggc.h (ggc_alloc_cleared_simd_clone_stat): Remove function.
1542 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1544 * alias.c (record_alias_subset): Adjust.
1545 * bitmap.c (bitmap_element_allocate): Likewise.
1546 (bitmap_gc_alloc_stat): Likewise.
1547 * cfg.c (init_flow): Likewise.
1548 (alloc_block): Likewise.
1549 (unchecked_make_edge): Likewise.
1550 * cfgloop.c (alloc_loop): Likewise.
1551 (flow_loops_find): Likewise.
1552 (rescan_loop_exit): Likewise.
1553 * cfgrtl.c (init_rtl_bb_info): Likewise.
1554 * cgraph.c (insert_new_cgraph_node_version): Likewise.
1555 (cgraph_allocate_node): Likewise.
1556 (cgraph_create_edge_1): Likewise.
1557 (cgraph_allocate_init_indirect_info): Likewise.
1558 * cgraphclones.c (cgraph_clone_edge): Likewise.
1559 * cgraphunit.c (add_asm_node): Likewise.
1560 (init_lowered_empty_function): Likewise.
1561 * config/aarch64/aarch64.c (aarch64_init_machine_status): Likewise.
1562 * config/alpha/alpha.c (alpha_init_machine_status): Likewise.
1563 (alpha_use_linkage): Likewise.
1564 * config/arc/arc.c (arc_init_machine_status): Likewise.
1565 * config/arm/arm.c (arm_init_machine_status): Likewise.
1566 * config/avr/avr.c (avr_init_machine_status): Likewise.
1567 * config/bfin/bfin.c (bfin_init_machine_status): Likewise.
1568 * config/c6x/c6x.c (c6x_init_machine_status): Likewise.
1569 * config/cris/cris.c (cris_init_machine_status): Likewise.
1570 * config/darwin.c (machopic_indirection_name): Likewise.
1571 (darwin_build_constant_cfstring): Likewise.
1572 (darwin_enter_string_into_cfstring_table): Likewise.
1573 * config/epiphany/epiphany.c (epiphany_init_machine_status): Likewise.
1574 * config/frv/frv.c (frv_init_machine_status): Likewise.
1575 * config/i386/i386.c (get_dllimport_decl): Likewise.
1576 (ix86_init_machine_status): Likewise.
1577 (assign_386_stack_local): Likewise.
1578 * config/i386/winnt.c (i386_pe_record_external_function): Likewise.
1579 (i386_pe_maybe_record_exported_symbol): Likewise.
1580 (i386_pe_record_stub): Likewise.
1581 * config/ia64/ia64.c (ia64_init_machine_status): Likewise.
1582 * config/iq2000/iq2000.c (iq2000_init_machine_status): Likewise.
1583 * config/m32c/m32c.c (m32c_init_machine_status): Likewise.
1584 (m32c_note_pragma_address): Likewise.
1585 * config/mep/mep.c (mep_init_machine_status): Likewise.
1586 (mep_note_pragma_flag): Likewise.
1587 * config/mips/mips.c (mflip_mips16_use_mips16_p): Likewise.
1588 (mips16_local_alias): Likewise.
1589 (mips_init_machine_status): Likewise.
1590 * config/mmix/mmix.c (mmix_init_machine_status): Likewise.
1591 * config/moxie/moxie.c (moxie_init_machine_status): Likewise.
1592 * config/msp430/msp430.c (msp430_init_machine_status): Likewise.
1593 * config/nds32/nds32.c (nds32_init_machine_status): Likewise.
1594 * config/nios2/nios2.c (nios2_init_machine_status): Likewise.
1595 * config/pa/pa.c (pa_init_machine_status): Likewise.
1596 (pa_get_deferred_plabel): Likewise.
1597 * config/rl78/rl78.c (rl78_init_machine_status): Likewise.
1598 * config/rs6000/rs6000.c (builtin_function_type): Likewise.
1599 (rs6000_init_machine_status): Likewise.
1600 (output_toc): Likewise.
1601 * config/s390/s390.c (s390_init_machine_status): Likewise.
1602 * config/score/score.c (score_output_external): Likewise.
1603 * config/sparc/sparc.c (sparc_init_machine_status): Likewise.
1604 * config/spu/spu.c (spu_init_machine_status): Likewise.
1605 * config/tilegx/tilegx.c (tilegx_init_machine_status): Likewise.
1606 * config/tilepro/tilepro.c (tilepro_init_machine_status): Likewise.
1607 * config/xtensa/xtensa.c (xtensa_init_machine_status): Likewise.
1608 * coverage.c (coverage_end_function): Likewise.
1609 * dbxout.c (dbxout_init): Likewise.
1610 * doc/gty.texi: Don't mention variable_size attribute.
1611 * dwarf2cfi.c (new_cfi): Adjust.
1612 (new_cfi_row): Likewise.
1613 (copy_cfi_row): Likewise.
1614 (create_cie_data): Likewise.
1615 * dwarf2out.c (dwarf2out_alloc_current_fde): Likewise.
1616 (new_loc_descr): Likewise.
1617 (find_AT_string_in_table): Likewise.
1618 (add_addr_table_entry): Likewise.
1619 (new_die): Likewise.
1620 (add_var_loc_to_decl): Likewise.
1621 (clone_die): Likewise.
1622 (clone_as_declaration): Likewise.
1623 (break_out_comdat_types): Likewise.
1624 (new_loc_list): Likewise.
1625 (add_loc_descr_to_each): Likewise.
1626 (add_location_or_const_value_attribute): Likewise.
1627 (add_linkage_name): Likewise.
1628 (lookup_filename): Likewise.
1629 (dwarf2out_var_location): Likewise.
1630 (new_line_info_table): Likewise.
1631 (dwarf2out_init): Likewise.
1632 (mem_loc_descriptor): Likewise.
1633 (loc_descriptor): Likewise.
1634 (add_const_value_attribute): Likewise.
1635 (tree_add_const_value_attribute): Likewise.
1636 (comp_dir_string): Likewise.
1637 (dwarf2out_vms_debug_main_pointer): Likewise.
1638 (string_cst_pool_decl): Likewise.
1639 * emit-rtl.c (set_mem_attrs): Likewise.
1640 (get_reg_attrs): Likewise.
1641 (start_sequence): Likewise.
1642 (init_emit): Likewise.
1643 (init_emit_regs): Likewise.
1644 * except.c (init_eh_for_function): Likewise.
1645 (gen_eh_region): Likewise.
1646 (gen_eh_region_catch): Likewise.
1647 (gen_eh_landing_pad): Likewise.
1648 (add_call_site): Likewise.
1649 * function.c (add_frame_space): Likewise.
1650 (insert_temp_slot_address): Likewise.
1651 (assign_stack_temp_for_type): Likewise.
1652 (get_hard_reg_initial_val): Likewise.
1653 (allocate_struct_function): Likewise.
1654 (prepare_function_start): Likewise.
1655 (types_used_by_var_decl_insert): Likewise.
1656 * gengtype.c (variable_size_p): Remove function.
1657 (enum alloc_quantity): Remove enum.
1658 (write_typed_alloc_def): Remove function.
1659 (write_typed_struct_alloc_def): Likewise.
1660 (write_typed_typedef_alloc_def): Likewise.
1661 (write_typed_alloc_defns): Likewise.
1663 * ggc-common.c (ggc_cleared_alloc_htab_ignore_args): Adjust.
1664 (ggc_cleared_alloc_ptr_array_two_args): Likewise.
1665 * ggc.h (ggc_alloc): new function.
1666 (ggc_cleared_alloc): Likewise.
1667 (ggc_vec_alloc): Template on type of vector element, and remove
1668 element size argument.
1669 (ggc_cleared_vec_alloc): Likewise.
1670 * gimple.c (gimple_build_omp_for): Adjust.
1671 (gimple_copy): Likewise.
1672 * ipa-cp.c (get_replacement_map): Likewise.
1673 (find_aggregate_values_for_callers_subset): Likewise.
1674 (known_aggs_to_agg_replacement_list): Likewise.
1675 * ipa-devirt.c (get_odr_type): Likewise.
1676 * ipa-prop.c (ipa_node_duplication_hook): Likewise.
1677 (read_agg_replacement_chain): Likewise.
1678 * loop-iv.c (get_simple_loop_desc): Likewise.
1679 * lto-cgraph.c (input_node_opt_summary): Likewise.
1680 * lto-section-in.c (lto_new_in_decl_state): Likewise.
1681 * lto-streamer-in.c (lto_input_eh_catch_list): Likewise.
1682 (input_eh_region): Likewise.
1683 (input_eh_lp): Likewise.
1684 (input_cfg): Likewise.
1685 * optabs.c (set_optab_libfunc): Likewise.
1686 (init_tree_optimization_optabs): Likewise.
1687 (set_conv_libfunc): Likewise.
1688 * passes.c (do_per_function_toporder): Likewise.
1689 * rtl.h: Don't use variable_size gty attribute.
1690 * sese.c (if_region_set_false_region): Adjust.
1691 * stringpool.c (gt_pch_save_stringpool): Likewise.
1692 * target-globals.c (save_target_globals): Likewise.
1693 * toplev.c (general_init): Likewise.
1694 * trans-mem.c (record_tm_replacement): Likewise.
1695 (split_bb_make_tm_edge): Likewise.
1696 * tree-cfg.c (move_sese_region_to_fn): Likewise.
1697 * tree-data-ref.h (lambda_vector_new): Likewise.
1698 * tree-eh.c (add_stmt_to_eh_lp_fn): Likewise.
1699 * tree-iterator.c (tsi_link_before): Likewise.
1700 (tsi_link_after): Likewise.
1701 * tree-scalar-evolution.c (new_scev_info_str): Likewise.
1702 * tree-ssa-loop-niter.c (record_estimate): Likewise.
1703 * tree-ssa-operands.c (ssa_operand_alloc): Likewise.
1704 * tree-ssa-operands.h: Don't use variable_size gty attribute.
1705 * tree-ssa.c (init_tree_ssa): Adjust.
1706 * tree-ssanames.c (set_range_info): Likewise.
1707 (get_ptr_info): Likewise.
1708 (duplicate_ssa_name_ptr_info): Likewise.
1709 (duplicate_ssa_name_range_info): Likewise.
1710 * tree-streamer-in.c (unpack_ts_real_cst_value_fields): Likewise.
1711 (unpack_ts_fixed_cst_value_fields): Likewise.
1712 * tree.c (build_fixed): Likewise.
1713 (build_real): Likewise.
1714 (build_string): Likewise.
1715 (decl_priority_info): Likewise.
1716 (decl_debug_expr_insert): Likewise.
1717 (decl_value_expr_insert): Likewise.
1718 (decl_debug_args_insert): Likewise.
1719 (type_hash_add): Likewise.
1720 (build_omp_clause): Likewise.
1721 * ubsan.c (decl_for_type_insert): Likewise.
1722 * varasm.c (get_unnamed_section): Likewise.
1723 (get_noswitch_section): Likewise.
1724 (get_section): Likewise.
1725 (get_block_for_section): Likewise.
1726 (create_block_symbol): Likewise.
1727 (build_constant_desc): Likewise.
1728 (create_constant_pool): Likewise.
1729 (force_const_mem): Likewise.
1730 (record_tm_clone_pair): Likewise.
1731 * varpool.c (varpool_create_empty_node): Likewise.
1733 2014-05-17 Trevor Saunders <tsaunders@mozilla.com>
1735 * dwarf2out.c (tree_add_const_value_attribute): Call
1736 ggc_internal_cleared_alloc instead of ggc_alloc_cleared_atomic.
1737 * gengtype.c (write_typed_alloc_def): Call ggc_internal_<x>alloc
1738 instead of ggc_internal_<x>alloc_stat.
1739 * ggc-common.c (ggc_internal_cleared_alloc): Drop _stat suffix.
1740 (ggc_realloc): Likewise.
1741 * ggc-none.c (ggc_internal_alloc): Likewise.
1742 (ggc_internal_cleared_alloc): Likewise.
1743 * ggc-page.c: Likewise.
1744 * ggc.h (ggc_internal_alloc_stat): Likewise.
1745 (ggc_internal_alloc): Remove macro.
1746 (ggc_internal_cleared_alloc_stat): Drop _stat suffix.
1747 (ggc_internal_cleared_alloc): Remove macro.
1748 (GGC_RESIZEVEC): Adjust.
1749 (ggc_resizevar): Remove macro.
1750 (ggc_internal_vec_alloc_stat): Drop _stat suffix.
1751 (ggc_internal_cleared_vec_alloc_stat): Likewise.
1752 (ggc_internal_vec_cleared_alloc): Remove macro.
1753 (ggc_alloc_atomic_stat): Drop _stat suffix.
1754 (ggc_alloc_atomic): Remove macro.
1755 (ggc_alloc_cleared_atomic): Remove macro.
1756 (ggc_alloc_string_stat): Drop _stat suffix.
1757 (ggc_alloc_string): Remove macro.
1758 (ggc_alloc_rtx_def_stat): Adjust.
1759 (ggc_alloc_tree_node_stat): Likewise.
1760 (ggc_alloc_cleared_tree_node_stat): Likewise.
1761 (ggc_alloc_cleared_gimple_statement_stat): Likewise.
1762 (ggc_alloc_cleared_simd_clone_stat): Likewise.
1763 * gimple.c (gimple_build_omp_for): Likewise.
1764 (gimple_copy): Likewise.
1765 * stringpool.c (ggc_alloc_string_stat): Drop _stat suffix.
1766 * toplev.c (realloc_for_line_map): Adjust.
1767 * tree-data-ref.h (lambda_vector_new): Likewise.
1768 * tree-phinodes.c (allocate_phi_node): Likewise.
1769 * tree.c (grow_tree_vec_stat): Likewise.
1770 * vec.h (va_gc::reserve): Adjust.
1772 2014-05-17 Ajit Agarwal <ajitkum@xilinx.com>
1774 * config/microblaze/microblaze.c (break_handler): New Declaration.
1775 (microblaze_break_function_p,microblaze_is_break_handler): New.
1776 (compute_frame_size): Use microblaze_break_function_p.
1777 Add the test of break_handler.
1778 (microblaze_function_prologue) : Add the test of variable
1779 break_handler. Check the fnname by BREAK_HANDLER_NAME.
1780 (microblaze_function_epilogue) : Add the test of break_handler.
1781 (microblaze_globalize_label) : Add the test of break_handler.
1782 Check the name by BREAK_HANDLER_NAME.
1784 * config/microblaze/microblaze.h (BREAK_HANDLER_NAME): New macro
1786 * config/microblaze/microblaze.md (*<optab>,<optab>_internal): Add
1787 microblaze_is_break_handler test.
1788 (call_internal1,call_value_intern): Use microblaze_break_function_p.
1789 Use SYMBOL_REF_DECL.
1791 * config/microblaze/microblaze-protos.h
1792 (microblaze_break_function_p,microblaze_is_break_handler):
1795 * doc/extend.texi (MicroBlaze break_handler Functions): Document
1796 new MicroBlaze break_handler functions.
1798 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1800 * doc/extend.texi (Size of an asm): Move node text according
1801 to its @menu entry position.
1803 2014-05-17 Marc Glisse <marc.glisse@inria.fr>
1805 PR tree-optimization/61140
1806 PR tree-optimization/61150
1807 PR tree-optimization/61197
1808 * tree-ssa-phiopt.c (value_replacement): Punt on multiple phis.
1810 2014-05-17 Uros Bizjak <ubizjak@gmail.com>
1812 * doc/invoke.texi (free): Mention Alpha. Also enabled at -Os.
1814 2014-05-17 Richard Sandiford <r.sandiford@uk.ibm.com>
1816 * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
1817 __SIZEOF_INT128__ is defined.
1819 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1821 * config/rs6000/rs6000.c (rs6000_real_tls_symbol_ref_p): New function.
1822 (rs6000_delegitimize_address): Use it.
1824 2014-05-17 Richard Sandiford <rdsandiford@googlemail.com>
1826 * emit-rtl.h (replace_equiv_address, replace_equiv_address_nv): Add an
1827 inplace argument. Store the new address in the original MEM when true.
1828 * emit-rtl.c (change_address_1): Likewise.
1829 (adjust_address_1, adjust_automodify_address_1, offset_address):
1831 * rtl.h (plus_constant): Add an inplace argument.
1832 * explow.c (plus_constant): Likewise. Try to reuse the original PLUS
1833 when true. Avoid generating (plus X (const_int 0)).
1834 * function.c (instantiate_virtual_regs_in_rtx): Adjust the PLUS
1835 in-place. Pass true to plus_constant.
1836 (instantiate_virtual_regs_in_insn): Pass true to replace_equiv_address.
1838 2014-05-16 Dehao Chen <dehao@google.com>
1840 * tree-cfg.c (gimple_merge_blocks): Updates bb count with max count.
1842 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1845 * config/sh/predicates.md (negt_reg_shl31_operand): Match additional
1847 * config/sh/sh.md (*negt_msb): Merge SH2A and non-SH2A variants.
1849 2014-05-16 Dehao Chen <dehao@google.com>
1851 * ira-int.h (REG_FREQ_FROM_EDGE_FREQ): Use
1852 optimize_function_for_size_p.
1853 * regs.h (REG_FREQ_FROM_BB): Likewise.
1855 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1858 * config/sh/sh.c (sh_eval_treg_value): Handle t_reg_operand and
1859 negt_reg_operand cases.
1860 * config/sh/sh.md (*cset_zero): Likewise by using cbranch_treg_value
1862 * config/sh/predicates.md (cbranch_treg_value): Simplify.
1864 2014-05-16 Oleg Endo <olegendo@gcc.gnu.org>
1866 * config/sh/sh.c (sh_option_override): Set branch cost to 2 for all
1869 2014-05-16 David Malcolm <dmalcolm@redhat.com>
1872 2014-04-29 David Malcolm <dmalcolm@redhat.com>
1874 * tree-cfg.c (dump_function_to_file): Dump the return type of
1875 functions, in a line to itself before the function body, mimicking
1876 the layout of a C function.
1878 2014-05-16 Dehao Chen <dehao@google.com>
1880 * cfghooks.c (make_forwarder_block): Use direct computation to
1881 get fall-through edge's count and frequency.
1883 2014-05-16 Benno Schulenberg <bensberg@justemail.net>
1885 * config/arc/arc.c (arc_init): Fix typo in error message.
1886 * config/i386/i386.c (ix86_expand_builtin): Likewise.
1887 (split_stack_prologue_scratch_regno): Likewise.
1888 * fortran/check.c (gfc_check_fn_rc2008): Remove duplicate
1889 word from error message.
1891 2014-05-16 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
1893 * ira-costs.c: Fix typo in comment.
1895 2014-05-16 David Wohlferd <dw@LimeGreenSocks.com>
1897 * doc/extend.texi: (Visibility Pragmas) Fix misplaced @xref
1899 2014-05-16 Jan Hubicka <hubicka@ucw.cz>
1901 * varpool.c (dump_varpool_node): Dump write-only flag.
1902 * lto-cgraph.c (lto_output_varpool_node, input_varpool_node): Stream
1904 * tree-cfg.c (execute_fixup_cfg): Remove statements setting
1905 write-only variables.
1906 * ipa.c (process_references): New function.
1907 (set_readonly_bit): New function.
1908 (set_writeonly_bit): New function.
1909 (clear_addressable_bit): New function.
1910 (ipa_discover_readonly_nonaddressable_var): Mark write only variables;
1911 fix handling of aliases.
1912 * cgraph.h (struct varpool_node): Add writeonly flag.
1914 2014-05-16 Vladimir Makarov <vmakarov@redhat.com>
1916 PR rtl-optimization/60969
1917 * ira-costs.c (record_reg_classes): Allow only memory for pseudo.
1918 Calculate costs for this case.
1920 2014-05-16 Eric Botcazou <ebotcazou@adacore.com>
1922 * fold-const (fold_unary_loc) <NON_LVALUE_EXPR>: New case.
1923 <CASE_CONVERT>: Pass arg0 instead of op0 to fold_convert_const.
1925 2014-05-16 Richard Biener <rguenther@suse.de>
1927 PR tree-optimization/61194
1928 * tree-vect-patterns.c (adjust_bool_pattern): Also handle
1929 bool patterns ending in a COND_EXPR.
1931 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1933 * config/aarch64/aarch64.c (aarch64_rtx_mult_cost): Fix FNMUL case.
1935 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1937 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle the case
1938 where we were unable to cost an RTX.
1940 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1942 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost SYMBOL_REF,
1945 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1946 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1948 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost TRUNCATE.
1950 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1951 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1953 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost FMA,
1954 FLOAT_EXTEND, FLOAT_TRUNCATE, ABS, SMAX, and SMIN.
1956 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1957 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1959 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost comparison
1962 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1963 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1965 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1968 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1969 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1971 * config/aarch64/aarch64.c (aarch64_rtx_arith_op_extract_p): New.
1972 (aarch64_rtx_costs): Improve costs for SIGN/ZERO_EXTRACT.
1974 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1975 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1977 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costs for
1980 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1981 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1983 * config/aarch64/aarch64.c (aarch64_rtx_costs): Cost
1984 ZERO_EXTEND and SIGN_EXTEND better.
1986 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1987 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1989 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve cost for
1992 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1993 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
1995 * config/aarch64/aarch64.c (aarch64_rtx_costs): Use address
1996 costs when costing loads and stores to memory.
1998 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
1999 Philip Tomsich <philipp.tomsich@theobroma-systems.com>
2001 * config/aarch64/aarch64.c (aarch64_rtx_costs): Improve costing
2004 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2006 * config/aarch64/aarch64.c (aarch64_rtx_costs): Set default costs.
2008 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2009 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2011 * config/aarch64/aarch64.c (aarch64_strip_shift_or_extend): Rename
2013 (aarch64_strip_extend): ...this, don't strip shifts, check RTX is
2015 (aarch64_rtx_mult_cost): New.
2016 (aarch64_rtx_costs): Use it, refactor as appropriate.
2018 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2019 Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
2021 * config/aarch64/aarch64.c (aarch64_build_constant): Conditionally
2022 emit instructions, return number of instructions which would
2024 (aarch64_add_constant): Update call to aarch64_build_constant.
2025 (aarch64_output_mi_thunk): Likewise.
2026 (aarch64_rtx_costs): Estimate cost of a CONST_INT, cost of
2029 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2031 * config/aarch64/aarch64.c (aarch64_rtx_costs_wrapper): New.
2032 (TARGET_RTX_COSTS): Call it.
2034 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2036 * config/aarch64/aarch64.c (cortexa57_addrcost_table): New.
2037 (cortexa57_vector_cost): Likewise.
2038 (cortexa57_tunings): Use them.
2040 2014-05-16 James Greenhalgh <james.greenhalgh@arm.com>
2042 * config/aarch64/aarch64-protos.h (scale_addr_mode_cost): New.
2043 (cpu_addrcost_table): Use it.
2044 * config/aarch64/aarch64.c (generic_addrcost_table): Initialize it.
2045 (aarch64_address_cost): Rewrite using aarch64_classify_address,
2048 2014-05-16 Richard Biener <rguenther@suse.de>
2050 * tree-ssa-sccvn.c: Include tree-cfg.h and domwalk.h.
2051 (set_ssa_val_to): Handle unexpected sets to VN_TOP.
2052 (visit_phi): Ignore edges marked as not executable.
2053 (class cond_dom_walker): New.
2054 (cond_dom_walker::before_dom_children): Value-number
2055 control statements and mark successor edges as not
2056 executable if possible.
2057 (run_scc_vn): First walk all control statements in
2058 dominator order, marking edges as not executable.
2059 * tree-inline.c (copy_edges_for_bb): Be not confused
2060 about random edge flags.
2062 2014-05-16 Richard Biener <rguenther@suse.de>
2064 * tree-ssa-sccvn.c (visit_use): Also constant-fold calls.
2066 2014-05-15 Peter Bergner <bergner@vnet.ibm.com>
2069 * config/rs6000/htmxlintrin.h (_HTM_TBEGIN_STARTED): New define.
2070 (__TM_simple_begin): Use it.
2071 (__TM_begin): Likewise.
2073 2014-05-15 Martin Jambor <mjambor@suse.cz>
2076 * ipa-prop.c (update_indirect_edges_after_inlining): Check
2077 type_preserved flag when the indirect edge is polymorphic.
2079 2014-05-15 Martin Jambor <mjambor@suse.cz>
2081 PR tree-optimization/61090
2082 * tree-sra.c (sra_modify_expr): Pass the current gsi to
2083 build_ref_for_model.
2085 2014-05-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2087 * config/arm/arm.c (arm_option_override): Use the SCHED_PRESSURE_MODEL
2088 enum name for PARAM_SCHED_PRESSURE_ALGORITHM.
2090 2014-05-15 Jakub Jelinek <jakub@redhat.com>
2092 PR tree-optimization/61158
2093 * fold-const.c (fold_binary_loc): If X is zero-extended and
2094 shiftc >= prec, make sure zerobits is all ones instead of
2095 invoking undefined behavior.
2097 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2099 * regcprop.h: New file.
2100 * regcprop.c (skip_debug_insn_p): New decl.
2101 (replace_oldest_value_reg): Check skip_debug_insn_p.
2102 (copyprop_hardreg_forward_bb_without_debug_insn): New function.
2103 * shrink-wrap.c: Include regcprop.h.
2104 (prepare_shrink_wrap): Call
2105 copyprop_hardreg_forward_bb_without_debug_insn.
2107 2014-05-15 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2109 * shrink-wrap.h: Update comment.
2110 * shrink-wrap.c: Update comment.
2111 (next_block_for_reg): Rename to live_edge_for_reg.
2112 (live_edge_for_reg): Allow live_edge->dest has two predecessors.
2113 (move_insn_for_shrink_wrap): Split live_edge.
2114 (prepre_shrink_wrap): One more parameter for move_insn_for_shrink_wrap.
2116 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2118 * config/sparc/sparc-protos.h (sparc_absnegfloat_split_legitimate):
2120 * config/sparc/sparc.c (sparc_absnegfloat_split_legitimate): Likewise.
2121 * config/sparc/sparc.md (fptype_ut699): New attribute.
2122 (in_branch_delay): Return false if -mfix-ut699 is specified and
2123 fptype_ut699 is set to single.
2124 (truncdfsf2): Add fptype_ut699 attribute.
2125 (fix_truncdfsi2): Likewise.
2126 (floatsisf2): Change fptype attribute.
2127 (fix_truncsfsi2): Likewise.
2128 (negtf2_notv9): Delete.
2129 (negtf2_v9): Likewise.
2130 (negtf2_hq): New instruction.
2131 (negtf2): New instruction and splitter.
2132 (negdf2_notv9): Rewrite.
2133 (abstf2_notv9): Delete.
2134 (abstf2_hq_v9): Likewise.
2135 (abstf2_v9): Likewise.
2136 (abstf2_hq): New instruction.
2137 (abstf2): New instruction and splitter.
2138 (absdf2_notv9): Rewrite.
2140 2014-05-14 Cary Coutant <ccoutant@google.com>
2143 * opts.c (common_handle_option): Don't special-case "-g".
2144 (set_debug_level): Default to at least level 2 with "-g".
2146 2014-05-14 DJ Delorie <dj@redhat.com>
2148 * config/msp430/msp430.c (msp430_builtin): Add
2149 MSP430_BUILTIN_DELAY_CYCLES.
2150 (msp430_init_builtins): Register void __delay_cycles(long long).
2151 (msp430_builtin_decl): Add it.
2152 (cg_magic_constant): New.
2153 (msp430_expand_delay_cycles): New.
2154 (msp430_expand_builtin): Call it.
2155 (msp430_print_operand_raw): Change integer printing from "int" to
2157 * config/msp430/msp430.md (define_constants): Add delay_cycles tags.
2158 (delay_cycles_start): New.
2159 (delay_cycles_end): New.
2160 (delay_cycles_32): New.
2161 (delay_cycles_32x): New.
2162 (delay_cycles_16): New.
2163 (delay_cycles_16x): New.
2164 (delay_cycles_2): New.
2165 (delay_cycles_1): New.
2166 * doc/extend.texi: Document __delay_cycles().
2168 2014-05-14 Sandra Loosemore <sandra@codesourcery.com>
2170 * config/nios2/nios2.md (nios2_cbranch): Fix paste-o in
2171 length attribute computation.
2173 2014-05-14 Richard Sandiford <rdsandiford@googlemail.com>
2176 * print-rtl.c (print_rtx): Suppress uids if flag_dump_unnumbered.
2178 2014-05-14 Richard Sandiford <r.sandiford@uk.ibm.com>
2181 * config/sparc/sparc.md: Fix types of low and high in DI constant
2182 splitter. Use gen_int_mode in some other splitters.
2184 2014-05-14 Martin Jambor <mjambor@suse.cz>
2187 * ipa-prop.c (ipa_modify_formal_parameters): Reset DECL_LANG_SPECIFIC.
2189 2014-05-14 James Norris <jnorris@codesourcery.com>
2191 * omp-low.c (expand_parallel_call): Remove shadow variable.
2192 (expand_omp_taskreg): Likewise.
2194 2014-05-14 Ilya Tocar <ilya.tocar@intel.com>
2196 * common/config/i386/i386-common.c
2197 (OPTION_MASK_ISA_CLFLUSHOPT_SET): Define.
2198 (OPTION_MASK_ISA_XSAVES_SET): Ditto.
2199 (OPTION_MASK_ISA_XSAVEC_SET): Ditto.
2200 (OPTION_MASK_ISA_CLFLUSHOPT_UNSET): Ditto.
2201 (OPTION_MASK_ISA_XSAVES_UNSET): Ditto.
2202 (OPTION_MASK_ISA_XSAVEC_UNSET): Ditto.
2203 (ix86_handle_option): Handle OPT_mxsavec, OPT_mxsaves, OPT_mclflushopt.
2204 * config.gcc (i[34567]86-*-*): Add clflushoptintrin.h,
2205 xsavecintrin.h, xsavesintrin.h.
2206 (x86_64-*-*): Ditto.
2207 * config/i386/clflushoptintrin.h: New.
2208 * config/i386/xsavecintrin.h: Ditto.
2209 * config/i386/xsavesintrin.h: Ditto.
2210 * config/i386/cpuid.h (bit_CLFLUSHOPT): Define.
2211 (bit_XSAVES): Ditto.
2212 (bit_XSAVES): Ditto.
2213 * config/i386/driver-i386.c (host_detect_local_cpu): Handle
2214 -mclflushopt, -mxsavec, -mxsaves, -mno-xsaves, -mno-xsavec,
2216 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
2217 OPTION_MASK_ISA_CLFLUSHOPT, OPTION_MASK_ISA_XSAVEC,
2218 OPTION_MASK_ISA_XSAVES.
2219 * config/i386/i386.c (ix86_target_string): Handle -mclflushopt,
2221 (PTA_CLFLUSHOPT) Define.
2222 (PTA_XSAVEC): Ditto.
2223 (PTA_XSAVES): Ditto.
2224 (ix86_option_override_internal): Handle new options.
2225 (ix86_valid_target_attribute_inner_p): Ditto.
2226 (ix86_builtins): Add IX86_BUILTIN_XSAVEC, IX86_BUILTIN_XSAVEC64,
2227 IX86_BUILTIN_XSAVES, IX86_BUILTIN_XRSTORS, IX86_BUILTIN_XSAVES64,
2228 IX86_BUILTIN_XRSTORS64, IX86_BUILTIN_CLFLUSHOPT.
2229 (bdesc_special_args): Add __builtin_ia32_xsaves,
2230 __builtin_ia32_xrstors, __builtin_ia32_xsavec, __builtin_ia32_xsaves64,
2231 __builtin_ia32_xrstors64, __builtin_ia32_xsavec64.
2232 (ix86_init_mmx_sse_builtins): Add __builtin_ia32_clflushopt.
2233 (ix86_expand_builtin): Handle new builtins.
2234 * config/i386/i386.h (TARGET_CLFLUSHOPT) Define.
2235 (TARGET_CLFLUSHOPT_P): Ditto.
2236 (TARGET_XSAVEC): Ditto.
2237 (TARGET_XSAVEC_P): Ditto.
2238 (TARGET_XSAVES): Ditto.
2239 (TARGET_XSAVES_P): Ditto.
2240 * config/i386/i386.md (ANY_XSAVE): Add UNSPECV_XSAVEC, UNSPECV_XSAVES.
2241 (ANY_XSAVE64)" Add UNSPECV_XSAVEC64, UNSPECV_XSAVES64.
2242 (attr xsave): Add xsavec, xsavec64, xsaves, xsaves64.
2244 (ANY_XRSTOR64): Ditto.
2246 (xrstor): Change into <xrstor>.
2247 (xrstor_rex64): Change into <xrstor>_rex64.
2248 (xrstor64): Change into <xrstor>64
2250 * config/i386/i386.opt (mclflushopt): New.
2253 * config/i386/x86intrin.h: Add clflushoptintrin.h, xsavesintrin.h,
2255 * doc/invoke.texi: Document new options.
2257 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2259 PR rtl-optimization/60866
2260 * sel-sched-ir (sel_init_new_insn): New parameter old_seqno.
2261 Default it to -1. Pass it down to init_simplejump_data.
2262 (init_simplejump_data): New parameter old_seqno. Pass it down
2263 to get_seqno_for_a_jump.
2264 (get_seqno_for_a_jump): New parameter old_seqno. Use it for
2265 initializing new jump seqno as a last resort. Add comment.
2266 (sel_redirect_edge_and_branch): Save old seqno of the conditional
2267 jump and pass it down to sel_init_new_insn.
2268 (sel_redirect_edge_and_branch_force): Likewise.
2270 2014-05-14 Georg-Johann Lay <avr@gjlay.de>
2272 * config/avr/avr.h (REG_CLASS_CONTENTS): Use unsigned suffix for
2273 shifted values to avoid build warning.
2275 2014-05-14 Eric Botcazou <ebotcazou@adacore.com>
2277 * cfgcleanup.c (try_forward_edges): Use location_t for locations.
2278 * cfgrtl.c (rtl_merge_blocks): Fix comment.
2279 (cfg_layout_merge_blocks): Likewise.
2280 * except.c (emit_to_new_bb_before): Remove prev_bb local variable.
2282 2014-05-14 Andrey Belevantsev <abel@ispras.ru>
2284 PR rtl-optimization/60901
2285 * config/i386/i386.c (ix86_dependencies_evaluation_hook): Check that
2286 bb predecessor belongs to the same scheduling region. Adjust comment.
2288 2014-05-13 Peter Bergner <bergner@vnet.ibm.com>
2290 * doc/sourcebuild.texi: (dfp_hw): Document.
2291 (p8vector_hw): Likewise.
2292 (powerpc_eabi_ok): Likewise.
2293 (powerpc_elfv2): Likewise.
2294 (powerpc_htm_ok): Likewise.
2295 (ppc_recip_hw): Likewise.
2298 2014-05-13 Cary Coutant <ccoutant@google.com>
2300 * opts.c (finish_options): Use -ggnu-pubnames with -gsplit-dwarf.
2302 2014-05-13 David Malcolm <dmalcolm@redhat.com>
2304 * gengtype-parse.c (require3): Eliminate in favor of...
2306 (require_template_declaration): Update to support optional single *
2309 * gengtype.c (get_ultimate_base_class): Add a non-const overload.
2310 (create_user_defined_type): Handle a single level of explicit
2311 pointerness within template arguments.
2312 (struct write_types_data): Add field "kind".
2313 (filter_type_name): Handle "*" character.
2314 (write_user_func_for_structure_ptr): Require a write_types_data
2315 rather than just a prefix string, so that we can look up the kind
2316 of the wtd and use it as an index into wrote_user_func_for_ptr,
2317 ensuring that such functions are written at most once. Support
2318 subclasses by invoking the marking function of the ultimate base class.
2319 (write_user_func_for_structure_body): Require a write_types_data
2320 rather than just a prefix string, so that we can pass this to
2321 write_user_func_for_structure_ptr.
2322 (write_func_for_structure): Likewise.
2323 (ggc_wtd): Add initializer of new "kind" field.
2324 (pch_wtd): Likewise.
2326 * gengtype.h (enum write_types_kinds): New.
2327 (struct type): Add field wrote_user_func_for_ptr to the "s"
2330 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2332 * fold-const.c (optimize_bit_field_compare): Use wi:: operations
2333 instead of const_binop.
2334 (fold_binary_loc): Likewise.
2336 2014-05-13 Richard Sandiford <r.sandiford@uk.ibm.com>
2338 * tree-dfa.h (get_addr_base_and_unit_offset_1): Update array index
2339 calculation to match get_ref_base_and_extent.
2341 2014-05-13 Catherine Moore <clm@codesourcery.com>
2342 Sandra Loosemore <sandra@codesourcery.com>
2344 * configure.ac: Fix assembly for explicit JALR relocation check.
2345 * configure: Regenerate.
2347 2014-05-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2349 * config/arm/arm.c (neon_itype): Remove NEON_RESULTPAIR.
2350 (arm_init_neon_builtins): Remove handling of NEON_RESULTPAIR.
2351 Remove associated type declarations and initialisations.
2352 (arm_expand_neon_builtin): Likewise.
2353 (neon_emit_pair_result_insn): Delete.
2354 * config/arm/arm_neon_builtins (vtrn, vzip, vuzp): Delete.
2355 * config/arm/neon.md (neon_vtrn<mode>): Delete.
2356 (neon_vzip<mode>): Likewise.
2357 (neon_vuzp<mode>): Likewise.
2359 2014-05-13 Richard Biener <rguenther@suse.de>
2362 * tree-inline.c (remap_gimple_stmt): Clear tail call flag,
2363 it needs revisiting whether the call still may be tail-called.
2365 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2367 * rtl.def (SYMBOL_REF): Remove middle "0" field.
2368 * rtl.h (block_symbol): Reduce number of fields to 2.
2369 (rtx_def): Add u2.symbol_ref_flags.
2370 (SYMBOL_REF_FLAGS): Use it.
2371 (SYMBOL_REF_DATA, SET_SYMBOL_REF_DECL, SYMBOL_REF_DECL)
2372 (SET_SYMBOL_REF_CONSTANT, SYMBOL_REF_CONSTANT): Lower index.
2373 * gengtype.c (adjust_field_rtx_def): Remove SYMBOL_REF_FLAGS handling.
2374 Lower index of SYMBOL_REF_DATA.
2375 * print-rtl.c (print_rtx): Lower index for SYMBOL_REF_DATA.
2376 Print SYMBOL_REF_FLAGS at the same time.
2377 * genattrtab.c (attr_rtx_1): Only initialize 1 "0" SYMBOL_REF field.
2379 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2381 * rtl.def (VAR_LOCATION): Remove "i" field.
2382 * rtl.h (rtx_def): Add u2.var_location_status.
2383 (PAT_VAR_LOCATION_STATUS): Use it.
2384 (gen_rtx_VAR_LOCATION): Declare.
2385 * gengenrtl.c (excluded_rtx): Add VAR_LOCATION.
2386 * emit-rtl.c (gen_rtx_VAR_LOCATION): New function.
2387 * var-tracking.c (emit_note_insn_var_location): Remove casts.
2389 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2391 * rtl.def (scratch): Fix outdated comment and remove "0" field.
2392 * gengtype.c (adjust_field_rtx_def): Update accordingly.
2394 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2396 * rtl.def (DEBUG_INSN, INSN, JUMP_INSN, CALL_INSN, JUMP_TABLE_DATA)
2397 (BARRIER, CODE_LABEL, NOTE): Remove first "i" field.
2398 * rtl.h (rtx_def): Add insn_uid to u2 field.
2399 (RTX_FLAG_CHECK8): Delete in favor of...
2400 (RTL_INSN_CHAIN_FLAG_CHECK): ...this new macro.
2401 (INSN_DELETED_P): Update accordingly.
2402 (INSN_UID): Use u2.insn_uid.
2403 (INSN_CHAIN_CODE_P): Define.
2404 (PREV_INSN, NEXT_INSN, BLOCK_FOR_INSN, PATTERN, INSN_LOCATION)
2405 (INSN_CODE, REG_NOTES, CALL_INSN_FUNCTION_USAGE, CODE_LABEL_NUMBER)
2406 (NOTE_DATA, NOTE_DELETED_LABEL_NAME, NOTE_BLOCK, NOTE_EH_HANDLER)
2407 (NOTE_BASIC_BLOCK, NOTE_VAR_LOCATION, NOTE_CFI, NOTE_LABEL_NUMBER)
2408 (NOTE_KIND, LABEL_NAME, LABEL_NUSES, JUMP_LABEL, LABEL_REFS): Lower
2409 indices accordingly.
2410 * print-rtl.c (print_rtx): Print INSN_UIDs before the main loop.
2411 Update indices for insn-chain rtxes.
2412 * gengtype.c (gen_rtx_next): Adjust test for insn-chain rtxes.
2413 (adjust_field_rtx_def): Lower '0' indices for all insn-chain rtxes.
2414 * emit-rtl.c (gen_label_rtx): Update gen_rtx_LABEL call.
2415 * caller-save.c (init_caller_save): Update gen_rtx_INSN calls.
2416 * combine.c (try_combine): Likewise.
2417 * ira.c (setup_prohibited_mode_move_regs): Likewise.
2419 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2421 * rtl.def (REG): Remove middle field.
2422 * rtl.h (rtx_def): Add orignal_regno to u2.
2423 (ORIGINAL_REGNO): Use it instead of field 1.
2424 (REG_ATTRS): Lower field index accordingly.
2425 * gengtype.c (adjust_field_rtx_def): Remove handling of
2426 ORIGINAL_REGNO. Move REG_ATTRS index down.
2427 * print-rtl.c (print_rtx): Move ORIGINAL_REGNO handling to the
2428 code that prints the REGNO.
2430 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2432 * print-rtl.c (print_rtx): Guard whole '0' block with ifndef
2435 2014-05-13 Richard Sandiford <rdsandiford@googlemail.com>
2437 * rtl.h (rtx_def): Mark u2 as GTY ((skip)).
2439 2014-05-13 Bin Cheng <bin.cheng@arm.com>
2441 * tree-ssa-loop-ivopts.c (contain_complex_addr_expr): New.
2442 (alloc_iv): Lower base expressions containing ADDR_EXPR.
2444 2014-05-13 Ian Bolton <ian.bolton@arm.com>
2446 * config/aarch64/aarch64-protos.h
2447 (aarch64_hard_regno_caller_save_mode): New prototype.
2448 * config/aarch64/aarch64.c (aarch64_hard_regno_caller_save_mode):
2450 * config/aarch64/aarch64.h (HARD_REGNO_CALLER_SAVE_MODE): New macro.
2452 2014-05-13 Christian Bruel <christian.bruel@st.com>
2454 * target.def (mode_switching): New hook vector.
2455 (mode_emit, mode_needed, mode_after, mode_entry): New hooks.
2456 (mode_exit, modepriority_to_mode): Likewise.
2457 * mode-switching.c (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Hookify.
2458 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2459 * target.h: Include tm.h and hard-reg-set.h.
2460 * doc/tm.texi.in (EMIT_MODE_SET, MODE_NEEDED, MODE_AFTER, MODE_ENTRY)
2461 (MODE_EXIT, MODE_PRIORITY_TO_MODE): Delete and hookify.
2462 * doc/tm.texi Regenerate.
2463 * config/sh/sh.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2464 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2465 * config/sh/sh.c (sh_emit_mode_set, sh_mode_priority): Hookify.
2466 (sh_mode_needed, sh_mode_after, sh_mode_entry, sh_mode_exit): Likewise.
2467 * config/i386/i386.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY): Delete
2468 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2469 * config/i386/i386-protos.h (ix86_mode_needed, ix86_mode_after)
2470 (ix86_mode_entrym, ix86_emit_mode_set): Remove external declaration.
2471 * config/i386/i386.c (ix86_mode_needed, ix86_mode_after,
2472 (ix86_mode_exit, ix86_mode_entry, ix86_mode_priority)
2473 (ix86_emit_mode_set): Hookify.
2474 * config/epiphany/epiphany.h (MODE_NEEDED, MODE_AFTER, MODE_ENTRY):
2476 (MODE_EXIT, MODE_PRIORITY_TO_MODE, EMIT_MODE_SET): Likewise.
2477 * config/epiphany/epiphany-protos.h (epiphany_mode_needed)
2478 (emit_set_fp_mode, epiphany_mode_entry_exit, epiphany_mode_after)
2479 (epiphany_mode_priority_to_mode): Remove declaration.
2480 * config/epiphany/epiphany.c (emit_set_fp_mode): Hookify.
2481 (epiphany_mode_needed, epiphany_mode_priority_to_mode): Likewise.
2482 (epiphany_mode_entry, epiphany_mode_exit, epiphany_mode_after):
2484 (epiphany_mode_priority_to_mode): Change priority type. Hookify.
2485 (epiphany_mode_needed, epiphany_mode_entry_exit): Hookify.
2486 (epiphany_mode_after, epiphany_mode_entry, emit_set_fp_mode): Hookify.
2488 2014-05-13 Jakub Jelinek <jakub@redhat.com>
2491 * config/i386/i386.c (ix86_expand_set_or_movmem): If count_exp
2492 is const0_rtx, return immediately. Don't test count == 0 when
2495 2014-05-13 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2497 * Makefile.in: add shrink-wrap.o.
2498 * config/i386/i386.c: include "shrink-wrap.h"
2499 * function.c: Likewise.
2500 (requires_stack_frame_p, next_block_for_reg,
2501 move_insn_for_shrink_wrap, prepare_shrink_wrap,
2502 dup_block_and_redirect): Move to shrink-wrap.c
2503 (thread_prologue_and_epilogue_insns): Extract three code segments
2504 as functions in shrink-wrap.c
2505 * function.h: Move #ifdef HAVE_simple_return ... #endif block to
2507 * shrink-wrap.c: New file.
2508 * shrink-wrap.h: New file.
2510 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2512 * doc/extend.texi: Reflect current numbers of pragmas. Remove
2513 reference to Solaris.
2515 2014-05-12 Mike Stump <mikestump@comcast.net>
2518 * genattrtab.c (filename): Add.
2519 (convert_set_attr_alternative): Improve error message.
2520 (check_defs): Restore read_md_filename for error messages.
2521 (gen_insn): Save filename.
2523 2014-05-12 Dimitris Papavasiliou <dpapavas@gmail.com>
2525 * doc/invoke.texi: Document new switches -Wno-shadow-ivar,
2526 -fno-local-ivars and -fivar-visibility.
2527 * c-family/c.opt: Make -Wshadow also implicitly enable
2530 2014-05-12 David Wohlferd <dw@LimeGreenSocks.com>
2532 * doc/tm.texi: Remove reference to deleted macro.
2533 * doc/tm.texi.in: Likewise.
2535 2014-05-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
2538 * config/avr/avr.c (avr_out_store_psi): Use correct constant
2541 2014-05-12 Georg-Johann Lay <avr@gjlay.de>
2544 * config/arm/arm.h (License): Add GCC Runtime Library Exception.
2545 * config/arm/aout.h (License): Same.
2546 * config/arm/bpabi.h (License): Same.
2547 * config/arm/elf.h (License): Same.
2548 * config/arm/linux-elf.h (License): Same.
2549 * config/arm/linux-gas.h (License): Same.
2550 * config/arm/netbsd-elf.h (License): Same.
2551 * config/arm/uclinux-eabi.h (License): Same.
2552 * config/arm/uclinux-elf.h (License): Same.
2553 * config/arm/vxworks.h (License): Same.
2555 2014-05-11 Jakub Jelinek <jakub@redhat.com>
2557 * tree.h (OMP_CLAUSE_LINEAR_STMT): Define.
2558 * tree.c (omp_clause_num_ops): Increase OMP_CLAUSE_LINEAR
2559 number of operands to 3.
2560 (walk_tree_1): Walk all operands of OMP_CLAUSE_LINEAR.
2561 * tree-nested.c (convert_nonlocal_omp_clauses,
2562 convert_local_omp_clauses): Handle OMP_CLAUSE_DEPEND.
2563 * gimplify.c (gimplify_scan_omp_clauses): Handle
2564 OMP_CLAUSE_LINEAR_STMT.
2565 * omp-low.c (lower_rec_input_clauses): Fix typo.
2566 (maybe_add_implicit_barrier_cancel, lower_omp_1): Add
2567 cast between Fortran boolean_type_node and C _Bool if
2570 2014-05-11 Richard Sandiford <rdsandiford@googlemail.com>
2572 PR tree-optimization/61136
2573 * wide-int.h (multiple_of_p): Define a version that doesn't return
2575 * fold-const.c (extract_muldiv_1): Use wi::multiple_of_p instead of an
2576 integer_zerop/const_binop pair.
2577 (multiple_of_p): Likewise, converting both operands to widest_int
2580 2014-05-09 Teresa Johnson <tejohnson@google.com>
2582 * cgraphunit.c (analyze_functions): Use correct dump file.
2584 2014-05-09 Florian Weimer <fweimer@redhat.com>
2586 * cfgexpand.c (stack_protect_decl_p): New function, extracted from
2588 (stack_protect_return_slot_p): New function.
2589 (expand_used_vars): Call stack_protect_decl_p and
2590 stack_protect_return_slot_p for -fstack-protector-strong.
2592 2014-05-09 Kenneth Zadeck <zadeck@naturalbridge.com>
2595 * fold-const.c (fold_binary_loc): Changed width of mask.
2597 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2599 * config/avr/avr-fixed.md (round<mode>3): Use -1U instead of -1 in
2600 unsigned int initializers for regno_in, regno_out.
2602 2014-05-09 Georg-Johann Lay <avr@gjlay.de>
2605 * config/avr/avr.md (cc): Add new attribute set_vzn.
2606 (addqi3, addqq3, adduqq3, subqi3, subqq3, subuqq3, negqi2) [cc]:
2607 Set cc insn attribute to set_vzn instead of set_zn for alternatives
2608 with INC, DEC or NEG.
2609 * config/avr/avr.c (avr_notice_update_cc): Handle SET_VZN.
2610 (avr_out_plus_1): ADIW sets cc0 to CC_SET_CZN.
2611 INC, DEC and ADD+ADC set cc0 to CC_CLOBBER.
2613 2014-05-09 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2616 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2618 * wide-int.cc (UTItype): Define.
2619 (UDWtype): Define for appropriate W_TYPE_SIZE.
2621 2014-05-09 Richard Biener <rguenther@suse.de>
2623 * Makefile.in (GTFILES): Remove tree-ssa-propagate.c.
2624 * tree-ssa-propagate.c: Do not include gt-tree-ssa-propagate.h.
2625 (interesting_ssa_edges, varying_ssa_edges): Move out of GC space.
2626 (add_ssa_edge, process_ssa_edge_worklist, ssa_prop_init,
2627 ssa_propagate): Adjust.
2629 2014-05-08 Jeff Law <law@redhat.com>
2631 PR tree-optimization/61009
2632 * tree-ssa-threadedge.c (thread_through_normal_block): Return a
2633 tri-state rather than a boolean. When a block is too big to
2634 thread through, inform caller via negative return value.
2635 (thread_across_edge): If a block was too big for normal threading,
2636 then it's too big for a joiner too, so remove temporary equivalences
2637 and return immediately.
2639 2014-05-08 Manuel López-Ibáñez <manu@gcc.gnu.org>
2640 Matthias Klose <doko@ubuntu.com>
2643 * optc-gen.awk: Fix option handling for -Wunused-parameter.
2645 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2648 * config/i386/i386.c (PTA_HASWELL): Remove PTA_RTM.
2650 2014-05-08 Uros Bizjak <ubizjak@gmail.com>
2653 * config/alpha/alpha.c: Include gimple-iterator.h.
2654 (alpha_gimple_fold_builtin): New function. Move
2655 ALPHA_BUILTIN_UMULH folding from ...
2656 (alpha_fold_builtin): ... here.
2657 (TARGET_GIMPLE_FOLD_BUILTIN): New define.
2659 2014-05-08 Wei Mi <wmi@google.com>
2662 * config/i386/i386.c (ix86_compute_frame_layout): Update
2663 preferred_stack_boundary for call, expanded from tls descriptor.
2664 * config/i386/i386.md (*tls_global_dynamic_32_gnu): Update RTX
2665 to depend on SP register.
2666 (*tls_local_dynamic_base_32_gnu): Ditto.
2667 (*tls_local_dynamic_32_once): Ditto.
2668 (tls_global_dynamic_64_<mode>): Set
2669 ix86_tls_descriptor_calls_expanded_in_cfun.
2670 (tls_local_dynamic_base_64_<mode>): Ditto.
2671 (tls_global_dynamic_32): Set
2672 ix86_tls_descriptor_calls_expanded_in_cfun. Update RTX
2673 to depend on SP register.
2674 (tls_local_dynamic_base_32): Ditto.
2676 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2678 * config/arm/arm_neon.h: Update comment.
2679 * config/arm/neon-docgen.ml: Delete.
2680 * config/arm/neon-gen.ml: Delete.
2681 * doc/arm-neon-intrinsics.texi: Update comment.
2683 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2685 * config/arm/arm_neon_builtins.def (vadd, vsub): Only define the v2sf
2687 (vand, vorr, veor, vorn, vbic): Remove.
2688 * config/arm/neon.md (neon_vadd, neon_vsub, neon_vadd_unspec): Adjust
2690 (neon_vsub_unspec): Likewise.
2691 (neon_vorr, neon_vand, neon_vbic, neon_veor, neon_vorn): Remove.
2693 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2695 * config/arm/arm_neon.h (vadd_s8): GNU C implementation
2696 (vadd_s16): Likewise.
2697 (vadd_s32): Likewise.
2698 (vadd_f32): Likewise.
2699 (vadd_u8): Likewise.
2700 (vadd_u16): Likewise.
2701 (vadd_u32): Likewise.
2702 (vadd_s64): Likewise.
2703 (vadd_u64): Likewise.
2704 (vaddq_s8): Likewise.
2705 (vaddq_s16): Likewise.
2706 (vaddq_s32): Likewise.
2707 (vaddq_s64): Likewise.
2708 (vaddq_f32): Likewise.
2709 (vaddq_u8): Likewise.
2710 (vaddq_u16): Likewise.
2711 (vaddq_u32): Likewise.
2712 (vaddq_u64): Likewise.
2713 (vmul_s8): Likewise.
2714 (vmul_s16): Likewise.
2715 (vmul_s32): Likewise.
2716 (vmul_f32): Likewise.
2717 (vmul_u8): Likewise.
2718 (vmul_u16): Likewise.
2719 (vmul_u32): Likewise.
2720 (vmul_p8): Likewise.
2721 (vmulq_s8): Likewise.
2722 (vmulq_s16): Likewise.
2723 (vmulq_s32): Likewise.
2724 (vmulq_f32): Likewise.
2725 (vmulq_u8): Likewise.
2726 (vmulq_u16): Likewise.
2727 (vmulq_u32): Likewise.
2728 (vsub_s8): Likewise.
2729 (vsub_s16): Likewise.
2730 (vsub_s32): Likewise.
2731 (vsub_f32): Likewise.
2732 (vsub_u8): Likewise.
2733 (vsub_u16): Likewise.
2734 (vsub_u32): Likewise.
2735 (vsub_s64): Likewise.
2736 (vsub_u64): Likewise.
2737 (vsubq_s8): Likewise.
2738 (vsubq_s16): Likewise.
2739 (vsubq_s32): Likewise.
2740 (vsubq_s64): Likewise.
2741 (vsubq_f32): Likewise.
2742 (vsubq_u8): Likewise.
2743 (vsubq_u16): Likewise.
2744 (vsubq_u32): Likewise.
2745 (vsubq_u64): Likewise.
2746 (vand_s8): Likewise.
2747 (vand_s16): Likewise.
2748 (vand_s32): Likewise.
2749 (vand_u8): Likewise.
2750 (vand_u16): Likewise.
2751 (vand_u32): Likewise.
2752 (vand_s64): Likewise.
2753 (vand_u64): Likewise.
2754 (vandq_s8): Likewise.
2755 (vandq_s16): Likewise.
2756 (vandq_s32): Likewise.
2757 (vandq_s64): Likewise.
2758 (vandq_u8): Likewise.
2759 (vandq_u16): Likewise.
2760 (vandq_u32): Likewise.
2761 (vandq_u64): Likewise.
2762 (vorr_s8): Likewise.
2763 (vorr_s16): Likewise.
2764 (vorr_s32): Likewise.
2765 (vorr_u8): Likewise.
2766 (vorr_u16): Likewise.
2767 (vorr_u32): Likewise.
2768 (vorr_s64): Likewise.
2769 (vorr_u64): Likewise.
2770 (vorrq_s8): Likewise.
2771 (vorrq_s16): Likewise.
2772 (vorrq_s32): Likewise.
2773 (vorrq_s64): Likewise.
2774 (vorrq_u8): Likewise.
2775 (vorrq_u16): Likewise.
2776 (vorrq_u32): Likewise.
2777 (vorrq_u64): Likewise.
2778 (veor_s8): Likewise.
2779 (veor_s16): Likewise.
2780 (veor_s32): Likewise.
2781 (veor_u8): Likewise.
2782 (veor_u16): Likewise.
2783 (veor_u32): Likewise.
2784 (veor_s64): Likewise.
2785 (veor_u64): Likewise.
2786 (veorq_s8): Likewise.
2787 (veorq_s16): Likewise.
2788 (veorq_s32): Likewise.
2789 (veorq_s64): Likewise.
2790 (veorq_u8): Likewise.
2791 (veorq_u16): Likewise.
2792 (veorq_u32): Likewise.
2793 (veorq_u64): Likewise.
2794 (vbic_s8): Likewise.
2795 (vbic_s16): Likewise.
2796 (vbic_s32): Likewise.
2797 (vbic_u8): Likewise.
2798 (vbic_u16): Likewise.
2799 (vbic_u32): Likewise.
2800 (vbic_s64): Likewise.
2801 (vbic_u64): Likewise.
2802 (vbicq_s8): Likewise.
2803 (vbicq_s16): Likewise.
2804 (vbicq_s32): Likewise.
2805 (vbicq_s64): Likewise.
2806 (vbicq_u8): Likewise.
2807 (vbicq_u16): Likewise.
2808 (vbicq_u32): Likewise.
2809 (vbicq_u64): Likewise.
2810 (vorn_s8): Likewise.
2811 (vorn_s16): Likewise.
2812 (vorn_s32): Likewise.
2813 (vorn_u8): Likewise.
2814 (vorn_u16): Likewise.
2815 (vorn_u32): Likewise.
2816 (vorn_s64): Likewise.
2817 (vorn_u64): Likewise.
2818 (vornq_s8): Likewise.
2819 (vornq_s16): Likewise.
2820 (vornq_s32): Likewise.
2821 (vornq_s64): Likewise.
2822 (vornq_u8): Likewise.
2823 (vornq_u16): Likewise.
2824 (vornq_u32): Likewise.
2825 (vornq_u64): Likewise.
2827 2014-05-08 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2829 * wide-int.cc (UTItype): Define.
2830 (UDWtype): Define for appropriate W_TYPE_SIZE.
2832 2014-05-08 Marc Glisse <marc.glisse@inria.fr>
2834 PR tree-optimization/59100
2835 * tree-ssa-phiopt.c: Include tree-inline.h.
2836 (neutral_element_p, absorbing_element_p): New functions.
2837 (value_replacement): Handle conditional binary operations with a
2838 neutral or absorbing element.
2840 2014-05-08 Richard Biener <rguenther@suse.de>
2842 * tree-ssa-sccvn.c (vn_get_expr_for): Valueize operands before
2843 folding the expression.
2844 (valueize_expr): Remove.
2845 (visit_reference_op_load): Do not valueize the result of
2847 (simplify_binary_expression): Likewise.
2848 (simplify_unary_expression): Likewise.
2850 2014-05-08 Richard Biener <rguenther@suse.de>
2852 * gimplify.c (gimplify_call_expr): Use saved fnptrtype for
2853 looking at TYPE_ARG_TYPES.
2855 2014-05-08 Richard Biener <rguenther@suse.de>
2857 * gimple-fold.c (gimple_fold_stmt_to_constant_1): Remove
2858 pointer propagation special-case.
2860 2014-05-08 Bin Cheng <bin.cheng@arm.com>
2862 * tree-affine.c (tree_to_aff_combination): Handle MEM_REF for
2863 core part of address expressions.
2865 2014-05-08 Alan Modra <amodra@gmail.com>
2868 * config/rs6000/rs6000.c (expand_block_move): Allow 64-bit
2869 loads and stores when -mno-strict-align at any alignment.
2870 (expand_block_clear): Similarly. Also correct calculation of
2873 2014-05-07 Thomas Preud'homme <thomas.preudhomme@arm.com>
2876 * tree-complex.c (expand_complex_move): Keep line info when expanding
2878 * tree-ssa-uninit.c (warn_uninit): New argument. Ignore assignment
2879 of complex expression. Use new argument to display correct location
2880 for values coming from phi statement.
2881 (warn_uninitialized_vars): Adapt to new signature of warn_uninit.
2882 (warn_uninitialized_phi): Pass location of phi argument to
2884 * tree-ssa.c (ssa_undefined_value_p): For SSA_NAME initialized by a
2885 COMPLEX_EXPR, recurse on each part of the COMPLEX_EXPR.
2887 2014-05-07 Segher Boessenkool <segher@kernel.crashing.org>
2889 * config/rs6000/predicates.md (indexed_address_mem): New.
2890 * config/rs6000/rs6000.md (type): Remove load_ext, load_ext_u,
2891 load_ext_ux, load_ux, load_u, store_ux, store_u, fpload_ux, fpload_u,
2892 fpstore_ux, fpstore_u.
2893 (sign_extend, indexed, update): New.
2894 (cell_micro): Adjust.
2895 (*zero_extend<mode>di2_internal1, *zero_extendsidi2_lfiwzx,
2896 *extendsidi2_lfiwax, *extendsidi2_nocell, *extendsfdf2_fpr,
2897 *movsi_internal1, *movsi_internal1_single, *movhi_internal,
2898 *movqi_internal, *movcc_internal1, mov<mode>_hardfloat,
2899 *mov<mode>_softfloat, *mov<mode>_hardfloat32, *mov<mode>_hardfloat64,
2900 *mov<mode>_softfloat64, *movdi_internal32, *movdi_internal64,
2901 *mov<mode>_string, *ldmsi8, *ldmsi7, *ldmsi6, *ldmsi5, *ldmsi4,
2902 *ldmsi3, *stmsi8, *stmsi7, *stmsi6, *stmsi5, *stmsi4, *stmsi3,
2903 *movdi_update1, movdi_<mode>_update, movdi_<mode>_update_stack,
2904 *movsi_update1, *movsi_update2, movsi_update, movsi_update_stack,
2905 *movhi_update1, *movhi_update2, *movhi_update3, *movhi_update4,
2906 *movqi_update1, *movqi_update2, *movqi_update3, *movsf_update1,
2907 *movsf_update2, *movsf_update3, *movsf_update4, *movdf_update1,
2908 *movdf_update2, load_toc_aix_si, load_toc_aix_di, probe_stack_<mode>,
2909 *stmw, *lmw, as well as 10 anonymous patterns): Adjust.
2911 * config/rs6000/dfp.md (movsd_store, movsd_load): Adjust.
2912 * config/rs6000/vsx.md (*vsx_movti_32bit, *vsx_extract_<mode>_load,
2913 *vsx_extract_<mode>_store): Adjust.
2914 * config/rs6000/rs6000.c (rs6000_adjust_cost, is_microcoded_insn,
2915 is_cracked_insn, insn_must_be_first_in_group,
2916 insn_must_be_last_in_group): Adjust.
2918 * config/rs6000/40x.md (ppc403-load, ppc403-store, ppc405-float):
2920 * config/rs6000/440.md (ppc440-load, ppc440-store, ppc440-fpload,
2921 ppc440-fpstore): Adjust.
2922 * config/rs6000/476.md (ppc476-load, ppc476-store, ppc476-fpload,
2923 ppc476-fpstore): Adjust.
2924 * config/rs6000/601.md (ppc601-load, ppc601-store, ppc601-fpload,
2925 ppc601-fpstore): Adjust.
2926 * config/rs6000/603.md (ppc603-load, ppc603-store, ppc603-fpload):
2928 * config/rs6000/6xx.md (ppc604-load, ppc604-store, ppc604-fpload):
2930 * config/rs6000/7450.md (ppc7450-load, ppc7450-store, ppc7450-fpload,
2931 ppc7450-fpstore): Adjust.
2932 * config/rs6000/7xx.md (ppc750-load, ppc750-store): Adjust.
2933 * config/rs6000/8540.md (ppc8540_load, ppc8540_store): Adjust.
2934 * config/rs6000/a2.md (ppca2-load, ppca2-fp-load, ppca2-fp-store):
2936 * config/rs6000/cell.md (cell-load, cell-load-ux, cell-load-ext,
2937 cell-fpload, cell-fpload-update, cell-store, cell-store-update,
2938 cell-fpstore, cell-fpstore-update): Adjust.
2939 * config/rs6000/e300c2c3.md (ppce300c3_load, ppce300c3_fpload,
2940 ppce300c3_store, ppce300c3_fpstore): Adjust.
2941 * config/rs6000/e500mc.md (e500mc_load, e500mc_fpload, e500mc_store,
2942 e500mc_fpstore): Adjust.
2943 * config/rs6000/e500mc64.md (e500mc64_load, e500mc64_fpload,
2944 e500mc64_store, e500mc64_fpstore): Adjust.
2945 * config/rs6000/e5500.md (e5500_load, e5500_fpload, e5500_store,
2946 e5500_fpstore): Adjust.
2947 * config/rs6000/e6500.md (e6500_load, e6500_fpload, e6500_store,
2948 e6500_fpstore): Adjust.
2949 * config/rs6000/mpc.md (mpccore-load, mpccore-store, mpccore-fpload):
2951 * config/rs6000/power4.md (power4-load, power4-load-ext,
2952 power4-load-ext-update, power4-load-ext-update-indexed,
2953 power4-load-update-indexed, power4-load-update, power4-fpload,
2954 power4-fpload-update, power4-store, power4-store-update,
2955 power4-store-update-indexed, power4-fpstore, power4-fpstore-update):
2957 * config/rs6000/power5.md (power5-load, power5-load-ext,
2958 power5-load-ext-update, power5-load-ext-update-indexed,
2959 power5-load-update-indexed, power5-load-update, power5-fpload,
2960 power5-fpload-update, power5-store, power5-store-update,
2961 power5-store-update-indexed, power5-fpstore, power5-fpstore-update):
2963 * config/rs6000/power6.md (power6-load, power6-load-ext,
2964 power6-load-update, power6-load-update-indexed,
2965 power6-load-ext-update, power6-load-ext-update-indexed, power6-fpload,
2966 power6-fpload-update, power6-store, power6-store-update,
2967 power6-store-update-indexed, power6-fpstore, power6-fpstore-update):
2969 * config/rs6000/power7.md (power7-load, power7-load-ext,
2970 power7-load-update, power7-load-update-indexed,
2971 power7-load-ext-update, power7-load-ext-update-indexed, power7-fpload,
2972 power7-fpload-update, power7-store, power7-store-update,
2973 power7-store-update-indexed, power7-fpstore, power7-fpstore-update):
2975 * config/rs6000/power8.md (power8-load, power8-load-update,
2976 power8-load-ext, power8-load-ext-update, power8-fpload,
2977 power8-fpload-update, power8-store, power8-store-update-indexed,
2978 power8-fpstore, power8-fpstore-update): Adjust.
2979 * config/rs6000/rs64.md (rs64a-load, rs64a-store, rs64a-fpload):
2981 * config/rs6000/titan.md (titan_lsu_load, titan_lsu_fpload,
2982 titan_lsu_store, titan_lsu_fpstore): Adjust.
2983 * config/rs6000/xfpu.md (fp-load, fp-store): Adjust.
2985 2014-05-07 Oleg Endo <olegendo@gcc.gnu.org>
2988 * config/sh/sh-mem.cc (sh_expand_strlen): Use loop when emitting
2989 unrolled byte insns. Emit address increments after move insns.
2991 2014-05-07 David Malcolm <dmalcolm@redhat.com>
2993 * gimple.h (gimple_builtin_call_types_compatible_p): Accept a
2994 const_gimple, rather than a gimple.
2995 (gimple_call_builtin_p): Likewise, for the three variants.
2997 * gimple.c (gimple_builtin_call_types_compatible_p): Likewise.
2998 (gimple_call_builtin_p): Likewise, for the three variants.
3000 2014-05-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
3002 PR tree-optimization/61095
3003 * tree-ssanames.c (get_nonzero_bits): Fix type extension in wi::shwi.
3005 2014-05-07 Richard Biener <rguenther@suse.de>
3007 PR tree-optimization/61034
3008 * tree-ssa-alias.c (call_may_clobber_ref_p_1): Export.
3009 (maybe_skip_until): Use translate to take into account
3010 lattices when trying to do disambiguations.
3011 (get_continuation_for_phi_1): Likewise.
3012 (get_continuation_for_phi): Adjust for added translate arguments.
3013 (walk_non_aliased_vuses): Likewise.
3014 * tree-ssa-alias.h (get_continuation_for_phi): Adjust prototype.
3015 (walk_non_aliased_vuses): Likewise.
3016 (call_may_clobber_ref_p_1): Declare.
3017 * tree-ssa-sccvn.c (vn_reference_lookup_3): Also disambiguate against
3018 calls. Stop early if we are only supposed to disambiguate.
3019 * tree-ssa-pre.c (translate_vuse_through_block): Adjust.
3021 2014-05-07 Joern Rennecke <joern.rennecke@embecosm.com>
3023 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
3024 Emit an error when the function has arguments.
3026 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3028 * cfgloop.h (unswitch_loops): Remove.
3029 * doc/passes.texi: Remove references to loop-unswitch.c
3030 * timevar.def (TV_LOOP_UNSWITCH): Remove.
3032 2014-05-07 Evgeny Stupachenko <evstupac@gmail.com>
3034 * tree-vect-data-refs.c (vect_grouped_load_supported): New
3035 check for loads group of length 3.
3036 (vect_permute_load_chain): New permutations for loads group of
3038 * tree-vect-stmts.c (vect_model_load_cost): Change cost
3039 of vec_perm_shuffle for the new permutations.
3041 2014-05-07 Alan Lawrence <alan.lawrence@arm.com>
3043 * config/aarch64/arm_neon.h (vtrn1_f32, vtrn1_p8, vtrn1_p16, vtrn1_s8,
3044 vtrn1_s16, vtrn1_s32, vtrn1_u8, vtrn1_u16, vtrn1_u32, vtrn1q_f32,
3045 vtrn1q_f64, vtrn1q_p8, vtrn1q_p16, vtrn1q_s8, vtrn1q_s16, vtrn1q_s32,
3046 vtrn1q_s64, vtrn1q_u8, vtrn1q_u16, vtrn1q_u32, vtrn1q_u64, vtrn2_f32,
3047 vtrn2_p8, vtrn2_p16, vtrn2_s8, vtrn2_s16, vtrn2_s32, vtrn2_u8,
3048 vtrn2_u16, vtrn2_u32, vtrn2q_f32, vtrn2q_f64, vtrn2q_p8, vtrn2q_p16,
3049 vtrn2q_s8, vtrn2q_s16, vtrn2q_s32, vtrn2q_s64, vtrn2q_u8, vtrn2q_u16,
3050 vtrn2q_u32, vtrn2q_u64): Replace temporary asm with __builtin_shuffle.
3052 2014-05-07 Thomas Schwinge <thomas@codesourcery.com>
3054 * loop-unswitch.c: Delete.
3056 2014-05-07 Richard Biener <rguenther@suse.de>
3058 * config.gcc: Always set need_64bit_hwint to yes.
3060 2014-05-07 Chung-Ju Wu <jasonwucj@gmail.com>
3062 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Have it in favor
3063 of using optimize_size.
3065 2014-05-06 Mike Stump <mikestump@comcast.net>
3067 * wide-int.h (wi::int_traits <HOST_WIDE_INT>): Always define.
3069 2014-05-06 Joseph Myers <joseph@codesourcery.com>
3071 * config/i386/sse.md (*mov<mode>_internal)
3072 (*<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>)
3073 (*<sse2_avx_avx512f>_loaddqu<mode><mask_name>)
3074 (<sse>_andnot<mode>3, <code><mode>3, *andnot<mode>3)
3075 (*<code><mode>3, *andnot<mode>3<mask_name>)
3076 (<mask_codefor><code><mode>3<mask_name>): Only consider
3077 TARGET_SSE_PACKED_SINGLE_INSN_OPTIMAL for modes of size 16.
3079 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3082 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
3084 * lra-constraints.c (valid_address_p): Move earlier in file.
3085 Add a constraint argument to the address_info version.
3086 (satisfies_memory_constraint_p): New function.
3087 (satisfies_address_constraint_p): Likewise.
3088 (process_alt_operands, curr_insn_transform): Use them.
3089 (process_address): Pass the constraint to valid_address_p when
3090 checking address operands.
3092 2014-05-06 Richard Sandiford <r.sandiford@uk.ibm.com>
3094 * lto-cgraph.c (compute_ltrans_boundary): Make node variables local
3095 to their respective blocks. Fix inadvertent use of "node".
3097 2014-05-06 Richard Sandiford <rdsandiford@googlemail.com>
3099 * emit-rtl.c (init_derived_machine_modes): New functionm, split
3101 (init_emit_once): ...here.
3102 * rtl.h (init_derived_machine_modes): Declare.
3103 * toplev.c (do_compile): Call it even if no_backend.
3105 2014-05-06 Kenneth Zadeck <zadeck@naturalbridge.com>
3106 Mike Stump <mikestump@comcast.net>
3107 Richard Sandiford <rdsandiford@googlemail.com>
3108 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3110 * alias.c (ao_ref_from_mem): Use wide-int interfaces.
3111 (rtx_equal_for_memref_p): Update comment.
3112 (adjust_offset_for_component_ref): Use wide-int interfaces.
3113 * builtins.c (get_object_alignment_2): Likewise.
3114 (c_readstr): Likewise.
3115 (target_char_cast): Add comment.
3116 (determine_block_size): Use wide-int interfaces.
3117 (expand_builtin_signbit): Likewise.
3118 (fold_builtin_int_roundingfn): Likewise.
3119 (fold_builtin_bitop): Likewise.
3120 (fold_builtin_bswap): Likewise.
3121 (fold_builtin_logarithm): Use signop.
3122 (fold_builtin_pow): Likewise.
3123 (fold_builtin_memory_op): Use wide-int interfaces.
3124 (fold_builtin_object_size): Likewise.
3125 * cfgloop.c (alloc_loop): Initialize nb_iterations_upper_bound and
3126 nb_iterations_estimate.
3127 (record_niter_bound): Use wide-int interfaces.
3128 (get_estimated_loop_iterations_int): Likewise.
3129 (get_estimated_loop_iterations): Likewise.
3130 (get_max_loop_iterations): Likewise.
3131 * cfgloop.h: Include wide-int.h.
3132 (struct nb_iter_bound): Change bound to widest_int.
3133 (struct loop): Change nb_iterations_upper_bound and
3134 nb_iterations_estimate to widest_int.
3135 (record_niter_bound): Switch to use widest_int.
3136 (get_estimated_loop_iterations): Likewise.
3137 (get_max_loop_iterations): Likewise.
3138 (gcov_type_to_double_int): Rename to gcov_type_to_wide_int and
3139 update for wide-int.
3140 * cgraph.c (cgraph_add_thunk): Use wide-int interfaces.
3141 * combine.c (try_combine): Likewise.
3142 (subst): Use CONST_SCALAR_INT_P rather than CONST_INT_P.
3143 * config/aarch64/aarch64.c (aapcs_vfp_sub_candidate): Use wide-int
3145 (aarch64_float_const_representable_p): Likewise.
3146 * config/arc/arc.c: Include wide-int.h.
3147 (arc_can_use_doloop_p): Use wide-int interfaces.
3148 * config/arm/arm.c (aapcs_vfp_sub_candidate): Likewise.
3149 (vfp3_const_double_index): Likewise.
3150 * config/avr/avr.c (avr_out_round): Likewise.
3151 (avr_fold_builtin): Likewise.
3152 * config/bfin/bfin.c (bfin_local_alignment): Likewise.
3153 (bfin_can_use_doloop_p): Likewise.
3154 * config/darwin.c (darwin_mergeable_constant_section): Likewise.
3155 (machopic_select_rtx_section): Update to handle CONST_WIDE_INT.
3156 * config/i386/i386.c: Include wide-int.h.
3157 (ix86_data_alignment): Use wide-int interfaces.
3158 (ix86_local_alignment): Likewise.
3159 (ix86_emit_swsqrtsf): Update real_from_integer.
3160 * config/msp430/msp430.c (msp430_attr): Use wide-int interfaces.
3161 * config/nds32/nds32.c (nds32_insert_attributes): Likewise.
3162 * config/rs6000/predicates.md (any_operand): Add const_wide_int.
3163 (zero_constant): Likewise.
3164 (input_operand): Likewise.
3165 (splat_input_operand): Likewise.
3166 (non_logical_cint_operand): Change const_double to const_wide_int.
3167 * config/rs6000/rs6000.c (num_insns_constant): Handle CONST_WIDE_INT.
3168 (easy_altivec_constant): Remove comment.
3169 (paired_expand_vector_init): Use CONSTANT_P.
3170 (rs6000_legitimize_address): Handle CONST_WIDE_INT.
3171 (rs6000_emit_move): Update checks.
3172 (rs6000_aggregate_candidate): Use wide-int interfaces.
3173 (rs6000_expand_ternop_builtin): Likewise.
3174 (rs6000_output_move_128bit): Handle CONST_WIDE_INT.
3175 (rs6000_assemble_integer): Likewise.
3176 (rs6000_hash_constant): Likewise.
3177 (output_toc): Likewise.
3178 (rs6000_rtx_costs): Likewise.
3179 (rs6000_emit_swrsqrt); Update call to real_from_integer.
3180 * config/rs6000/rs6000-c.c: Include wide-int.h.
3181 (altivec_resolve_overloaded_builtin): Use wide-int interfaces.
3182 * config/rs6000/rs6000.h (TARGET_SUPPORTS_WIDE_INT): New.
3183 * config/rs6000/rs6000.md: Use const_scalar_int_operand.
3184 Handle CONST_WIDE_INT.
3185 * config/sol2-c.c (solaris_pragma_align): Change low to unsigned HWI.
3186 Use tree_fits_uhwi_p.
3187 * config/sparc/sparc.c: Include wide-int.h.
3188 (sparc_fold_builtin): Use wide-int interfaces.
3189 * config/vax/vax.c: Include wide-int.h.
3190 (vax_float_literal): Use real_from_integer.
3191 * coretypes.h (struct hwivec_def): New.
3193 (const_hwivec): New.
3194 * cse.c (hash_rtx_cb): Handle CONST_WIDE_INT.
3195 (equiv_constant): Handle CONST_WIDE_INT.
3196 * cselib.c (rtx_equal_for_cselib_1): Use CASE_CONST_UNIQUE.
3197 (cselib_hash_rtx): Handle CONST_WIDE_INT.
3198 * dbxout.c (stabstr_U): Use wide-int interfaces.
3199 (dbxout_type): Update to use cst_fits_shwi_p.
3200 * defaults.h (LOG2_BITS_PER_UNIT): Define.
3201 (TARGET_SUPPORTS_WIDE_INT): Add default.
3202 * dfp.c: Include wide-int.h.
3203 (decimal_real_to_integer2): Use wide-int interfaces and rename to
3204 decimal_real_to_integer.
3205 * dfp.h (decimal_real_to_integer2): Return a wide_int and rename to
3206 decimal_real_to_integer.
3207 * doc/generic.texi (Constant expressions): Update for wide_int.
3208 * doc/rtl.texi (const_double): Likewise.
3209 (const_wide_int, CONST_WIDE_INT, CONST_WIDE_INT_VEC): New.
3210 (CONST_WIDE_INT_NUNITS, CONST_WIDE_INT_ELT): New.
3211 * doc/tm.texi.in (REAL_VALUE_TO_INT): Remove.
3212 (REAL_VALUE_FROM_INT): Remove.
3213 (TARGET_SUPPORTS_WIDE_INT): New.
3214 * doc/tm.texi: Regenerate.
3215 * dojump.c (prefer_and_bit_test): Use wide-int interfaces.
3216 * double-int.h: Include wide-int.h.
3217 (struct wi::int_traits): New.
3218 * dwarf2out.c (get_full_len): New.
3219 (dw_val_equal_p): Add case dw_val_class_wide_int.
3220 (size_of_loc_descr): Likewise.
3221 (output_loc_operands): Likewise.
3222 (insert_double): Remove.
3223 (insert_wide_int): New.
3225 (print_die): Add case dw_val_class_wide_int.
3226 (attr_checksum): Likewise.
3227 (attr_checksum_ordered): Likewise.
3228 (same_dw_val_p): Likewise.
3229 (size_of_die): Likewise.
3230 (value_format): Likewise.
3231 (output_die): Likewise.
3232 (double_int_type_size_in_bits): Rename to offset_int_type_size_in_bits.
3234 (clz_loc_descriptor): Use wide-int interfaces.
3235 (mem_loc_descriptor): Likewise. Handle CONST_WIDE_INT.
3236 (loc_descriptor): Use wide-int interfaces. Handle CONST_WIDE_INT.
3237 (round_up_to_align): Use wide-int interfaces.
3238 (field_byte_offset): Likewise.
3239 (insert_double): Rename to insert_wide_int. Use wide-int interfaces.
3240 (add_const_value_attribute): Handle CONST_WIDE_INT. Update
3241 CONST_DOUBLE handling. Use wide-int interfaces.
3242 (add_bound_info): Use tree_fits_uhwi_p. Use wide-int interfaces.
3243 (gen_enumeration_type_die): Use add_AT_wide.
3244 (hash_loc_operands): Add case dw_val_class_wide_int.
3245 (compare_loc_operands): Likewise.
3246 * dwarf2out.h: Include wide-int.h.
3247 (wide_int_ptr): New.
3248 (enum dw_val_class): Add dw_val_class_wide_int.
3249 (struct dw_val_struct): Add val_wide.
3250 * emit-rtl.c (const_wide_int_htab): New.
3251 (const_wide_int_htab_hash): New.
3252 (const_wide_int_htab_eq): New.
3253 (lookup_const_wide_int): New.
3254 (const_double_htab_hash): Use wide-int interfaces.
3255 (const_double_htab_eq): Likewise.
3256 (rtx_to_double_int): Conditionally compile for wide-int.
3257 (immed_double_int_const): Rename to immed_wide_int_const and
3258 update for wide-int.
3259 (immed_double_const): Conditionally compile for wide-int.
3260 (init_emit_once): Use wide-int interfaces.
3261 * explow.c (plus_constant): Likewise.
3262 * expmed.c (mask_rtx): Move further up file. Use wide-int interfaces.
3263 (lshift_value): Use wide-int interfaces.
3264 (expand_mult): Likewise.
3265 (choose_multiplier): Likewise.
3266 (expand_smod_pow2): Likewise.
3267 (make_tree): Likewise.
3268 * expr.c (convert_modes): Consolidate handling of constants.
3269 Use wide-int interfaces.
3270 (emit_group_load_1): Add note.
3271 (store_expr): Update comment.
3272 (get_inner_reference): Use wide-int interfaces.
3273 (expand_constructor): Update comment.
3274 (expand_expr_real_2): Use wide-int interfaces.
3275 (expand_expr_real_1): Likewise.
3276 (reduce_to_bit_field_precision): Likewise.
3277 (const_vector_from_tree): Likewise.
3278 * final.c: Include wide-int-print.h.
3279 (output_addr_const): Handle CONST_WIDE_INT. Use CONST_DOUBLE_AS_INT_P.
3280 * fixed-value.c: Include wide-int.h.
3281 (fixed_from_string): Use wide-int interfaces.
3282 (fixed_to_decimal): Likewise.
3283 (fixed_convert_from_real): Likewise.
3284 (real_convert_from_fixed): Likewise.
3285 * fold-const.h (mem_ref_offset): Return an offset_int.
3286 (div_if_zero_remainder): Remove code parameter.
3287 * fold-const.c (div_if_zero_remainder): Remove code parameter.
3288 Use wide-int interfaces.
3289 (may_negate_without_overflow_p): Use wide-int interfaces.
3290 (negate_expr_p): Likewise.
3291 (fold_negate_expr): Likewise.
3292 (int_const_binop_1): Likewise.
3293 (const_binop): Likewise.
3294 (fold_convert_const_int_from_int): Likewise.
3295 (fold_convert_const_int_from_real): Likewise.
3296 (fold_convert_const_int_from_fixed): Likewise.
3297 (fold_convert_const_fixed_from_int): Likewise.
3298 (all_ones_mask_p): Take an unsigned size. Use wide-int interfaces.
3299 (sign_bit_p): Use wide-int interfaces.
3300 (make_range_step): Likewise.
3301 (build_range_check): Likewise. Pass an integer of the correct type
3302 instead of using integer_one_node.
3303 (range_predecessor): Pass an integer of the correct type instead
3304 of using integer_one_node.
3305 (range_successor): Likewise.
3306 (merge_ranges): Likewise.
3307 (unextend): Use wide-int interfaces.
3308 (extract_muldiv_1): Likewise.
3309 (fold_div_compare): Likewise.
3310 (fold_single_bit_test): Likewise.
3311 (fold_sign_changed_comparison): Likewise.
3312 (try_move_mult_to_index): Update calls to div_if_zero_remainder.
3313 (fold_plusminus_mult_expr): Use wide-int interfaces.
3314 (native_encode_int): Likewise.
3315 (native_interpret_int): Likewise.
3316 (fold_unary_loc): Likewise.
3317 (pointer_may_wrap_p): Likewise.
3318 (size_low_cst): Likewise.
3319 (mask_with_tz): Likewise.
3320 (fold_binary_loc): Likewise.
3321 (fold_ternary_loc): Likewise.
3322 (multiple_of_p): Likewise.
3323 (tree_call_nonnegative_warnv_p): Update calls to
3324 tree_int_cst_min_precision and real_from_integer.
3325 (fold_negate_const): Use wide-int interfaces.
3326 (fold_abs_const): Likewise.
3327 (fold_relational_const): Use tree_int_cst_lt.
3328 (round_up_loc): Use wide-int interfaces.
3329 * genemit.c (gen_exp): Add CONST_WIDE_INT case.
3330 * gengenrtl.c (excluded_rtx): Add CONST_WIDE_INT case.
3331 * gengtype.c: Remove include of double-int.h.
3332 (do_typedef): Use wide-int interfaces.
3333 (open_base_files): Add wide-int.h.
3334 (main): Add offset_int and widest_int typedefs.
3335 * gengtype-lex.l: Handle "^".
3336 (CXX_KEYWORD): Add "static".
3337 * gengtype-parse.c (require3): New.
3338 (require_template_declaration): Handle constant template arguments
3339 and nested templates.
3340 * gengtype-state.c: Don't include "double-int.h".
3341 * genpreds.c (write_one_predicate_function): Update comment.
3342 (write_tm_constrs_h): Add check for hval and lval use in
3344 * genrecog.c (validate_pattern): Add CONST_WIDE_INT case.
3345 (add_to_sequence): Likewise.
3346 * gensupport.c (struct std_pred_table): Add const_scalar_int_operand
3347 and const_double_operand.
3348 * gimple.c (preprocess_case_label_vec_for_gimple): Use wide-int
3350 * gimple-fold.c (get_base_constructor): Likewise.
3351 (fold_array_ctor_reference): Likewise.
3352 (fold_nonarray_ctor_reference): Likewise.
3353 (fold_const_aggregate_ref_1): Likewise.
3354 (gimple_val_nonnegative_real_p): Likewise.
3355 (gimple_fold_indirect_ref): Likewise.
3356 * gimple-pretty-print.c (dump_ssaname_info): Likewise.
3357 * gimple-ssa-strength-reduction.c: Include wide-int-print.h.
3358 (struct slsr_cand_d): Change index to be widest_int.
3359 (struct incr_info_d): Change incr to be widest_int.
3360 (alloc_cand_and_find_basis): Use wide-int interfaces.
3361 (slsr_process_phi): Likewise.
3362 (backtrace_base_for_ref): Likewise. Return a widest_int.
3363 (restructure_reference): Take a widest_int instead of a double_int.
3364 (slsr_process_ref): Use wide-int interfaces.
3365 (create_mul_ssa_cand): Likewise.
3366 (create_mul_imm_cand): Likewise.
3367 (create_add_ssa_cand): Likewise.
3368 (create_add_imm_cand): Take a widest_int instead of a double_int.
3369 (slsr_process_add): Use wide-int interfaces.
3370 (slsr_process_cast): Likewise.
3371 (slsr_process_copy): Likewise.
3372 (dump_candidate): Likewise.
3373 (dump_incr_vec): Likewise.
3374 (replace_ref): Likewise.
3375 (cand_increment): Likewise. Return a widest_int.
3376 (cand_abs_increment): Likewise.
3377 (replace_mult_candidate): Take a widest_int instead of a double_int.
3378 (replace_unconditional_candidate): Use wide-int interfaces.
3379 (incr_vec_index): Take a widest_int instead of a double_int.
3380 (create_add_on_incoming_edge): Likewise.
3381 (create_phi_basis): Use wide-int interfaces.
3382 (replace_conditional_candidate): Likewise.
3383 (record_increment): Take a widest_int instead of a double_int.
3384 (record_phi_increments): Use wide-int interfaces.
3385 (phi_incr_cost): Take a widest_int instead of a double_int.
3386 (lowest_cost_path): Likewise.
3387 (total_savings): Likewise.
3388 (analyze_increments): Use wide-int interfaces.
3389 (ncd_with_phi): Take a widest_int instead of a double_int.
3390 (ncd_of_cand_and_phis): Likewise.
3391 (nearest_common_dominator_for_cands): Likewise.
3392 (insert_initializers): Use wide-int interfaces.
3393 (all_phi_incrs_profitable): Likewise.
3394 (replace_one_candidate): Likewise.
3395 (replace_profitable_candidates): Likewise.
3396 * godump.c: Include wide-int-print.h.
3397 (go_output_typedef): Use wide-int interfaces.
3398 * graphite-clast-to-gimple.c (gmp_cst_to_tree): Likewise.
3399 * graphite-sese-to-poly.c (tree_int_to_gmp): Likewise.
3400 (build_loop_iteration_domains): Likewise.
3401 * hooks.h: Include wide-int.h rather than double-int.h.
3402 (hook_bool_dint_dint_uint_bool_true): Delete.
3403 (hook_bool_wint_wint_uint_bool_true): Declare.
3404 * hooks.c (hook_bool_dint_dint_uint_bool_true): Removed.
3405 (hook_bool_wint_wint_uint_bool_true): New.
3406 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): Use wide-int
3408 (ubsan_expand_si_overflow_mul_check): Likewise.
3409 * ipa-devirt.c (get_polymorphic_call_info): Likewise.
3410 * ipa-prop.c (compute_complex_assign_jump_func): Likewise.
3411 (get_ancestor_addr_info): Likewise.
3412 (ipa_modify_call_arguments): Likewise.
3413 * loop-doloop.c (doloop_modify): Likewise.
3414 (doloop_optimize): Likewise.
3415 * loop-iv.c (iv_number_of_iterations): Likewise.
3416 * loop-unroll.c (decide_unroll_constant_iterations): Likewise.
3417 (unroll_loop_constant_iterations): Likewise.
3418 (decide_unroll_runtime_iterations): Likewise.
3419 (unroll_loop_runtime_iterations): Likewise.
3420 (decide_peel_simple): Likewise.
3421 (decide_unroll_stupid): Likewise.
3422 * lto-streamer-in.c (streamer_read_wi): Add.
3423 (input_cfg): Use wide-int interfaces.
3424 (lto_input_tree_1): Likewise.
3425 * lto-streamer-out.c (streamer_write_wi): Add.
3426 (hash_tree): Use wide-int interfaces.
3427 (output_cfg): Likewise.
3428 * Makefile.in (OBJS): Add wide-int.o and wide-int-print.o.
3429 (GTFILES): Add wide-int.h and signop.h.
3430 (TAGS): Look for .cc files too.
3431 * omp-low.c (scan_omp_1_op): Use wide-int interfaces.
3432 * optabs.c (expand_subword_shift): Likewise.
3433 (expand_doubleword_shift): Likewise.
3434 (expand_absneg_bit): Likewise.
3435 (expand_copysign_absneg): Likewise.
3436 (expand_copysign_bit): Likewise.
3437 * postreload.c (reload_cse_simplify_set): Likewise.
3438 * predict.c (predict_iv_comparison): Likewise.
3439 * pretty-print.h: Include wide-int-print.h.
3441 * print-rtl.c (print_rtx): Add CONST_WIDE_INT case.
3442 * print-tree.c: Include wide-int-print.h.
3443 (print_node_brief): Use wide-int interfaces.
3444 (print_node): Likewise.
3445 * read-rtl.c (validate_const_wide_int): New.
3446 (read_rtx_code): Add CONST_WIDE_INT case.
3447 * real.c: Include wide-int.h.
3448 (real_to_integer2): Delete.
3449 (real_to_integer): New function, returning a wide_int.
3450 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3451 (ten_to_ptwo): Update call to real_from_integer.
3452 (real_digit): Likewise.
3453 * real.h: Include signop.h, wide-int.h and insn-modes.h.
3454 (real_to_integer2, REAL_VALUE_FROM_INT, REAL_VALUE_FROM_UNSIGNED_INT)
3455 (REAL_VALUE_TO_INT): Delete.
3456 (real_to_integer): Declare a wide-int form.
3457 (real_from_integer): Take a wide_int rather than two HOST_WIDE_INTs.
3458 * recog.c (const_int_operand): Improve comment.
3459 (const_scalar_int_operand): New.
3460 (const_double_operand): Add a separate definition for CONST_WIDE_INT.
3461 * rtlanal.c (commutative_operand_precedence): Handle CONST_WIDE_INT.
3462 (split_double): Likewise.
3463 * rtl.c (DEF_RTL_EXPR): Handle CONST_WIDE_INT.
3464 (rtx_size): Likewise.
3465 (rtx_alloc_stat_v): New.
3466 (rtx_alloc_stat): Now calls rtx_alloc_stat_v.
3467 (cwi_output_hex): New.
3468 (iterative_hash_rtx): Handle CONST_WIDE_INT.
3469 (cwi_check_failed_bounds): New.
3470 * rtl.def (CONST_WIDE_INT): New.
3471 * rtl.h: Include <utility> and wide-int.h.
3472 (struct hwivec_def): New.
3473 (CWI_GET_NUM_ELEM): New.
3474 (CWI_PUT_NUM_ELEM): New.
3475 (struct rtx_def): Add num_elem and hwiv.
3476 (CASE_CONST_SCALAR_INT): Modify for TARGET_SUPPORTS_WIDE_INT.
3477 (CASE_CONST_UNIQUE): Likewise.
3478 (CASE_CONST_ANY): Likewise.
3479 (CONST_SCALAR_INT_P): Likewise.
3480 (CONST_WIDE_INT_P): New.
3482 (HWIVEC_CHECK): New.
3483 (cwi_check_failed_bounds): New.
3485 (HWIVEC_CHECK): New.
3486 (CONST_WIDE_INT_VEC) New.
3487 (CONST_WIDE_INT_NUNITS) New.
3488 (CONST_WIDE_INT_ELT) New.
3489 (rtx_mode_t): New type.
3490 (wi::int_traits <rtx_mode_t>): New.
3492 (wi::min_value): New.
3493 (wi::max_value): New.
3495 (const_wide_int_alloc): New.
3496 (immed_wide_int_const): New.
3497 * sched-vis.c (print_value): Handle CONST_WIDE_INT.
3498 * sel-sched-ir.c (lhs_and_rhs_separable_p): Update comment.
3499 * signop.h: New file.
3500 * simplify-rtx.c (mode_signbit_p): Handle CONST_WIDE_INT.
3501 (simplify_const_unary_operation): Use wide-int interfaces.
3502 (simplify_binary_operation_1): Likewise.
3503 (simplify_const_binary_operation): Likewise.
3504 (simplify_const_relational_operation): Likewise.
3505 (simplify_immed_subreg): Likewise.
3506 * stmt.c (expand_case): Likewise.
3507 * stor-layout.h (set_min_and_max_values_for_integral_type): Take a
3508 signop rather than a bool.
3509 * stor-layout.c (layout_type): Use wide-int interfaces.
3510 (initialize_sizetypes): Update calls to
3511 set_min_and_max_values_for_integral_type.
3512 (set_min_and_max_values_for_integral_type): Take a signop rather
3513 than a bool. Use wide-int interfaces.
3514 (fixup_signed_type): Update accordingly. Remove
3515 HOST_BITS_PER_DOUBLE_INT limit.
3516 (fixup_unsigned_type): Likewise.
3517 * system.h (STATIC_CONSTANT_P): New.
3518 (STATIC_ASSERT): New.
3519 * target.def (can_use_doloop_p): Take widest_ints rather than
3521 * target.h: Include wide-int.h rather than double-int.h.
3522 * targhooks.h (can_use_doloop_if_innermost): Take widest_ints rather
3524 * targhooks.c (default_cxx_get_cookie_size): Use tree_int_cst_lt
3525 rather than INT_CST_LT_UNSIGNED.
3526 (can_use_doloop_if_innermost): Take widest_ints rather than
3528 * tree-affine.c: Include wide-int-print.h.
3529 (double_int_ext_for_comb): Delete.
3530 (wide_int_ext_for_comb): New.
3531 (aff_combination_zero): Use wide-int interfaces.
3532 (aff_combination_const): Take a widest_int instead of a double_int.
3533 (aff_combination_elt): Use wide-int interfaces.
3534 (aff_combination_scale): Take a widest_int instead of a double_int.
3535 (aff_combination_add_elt): Likewise.
3536 (aff_combination_add_cst): Likewise.
3537 (aff_combination_add): Use wide-int interfaces.
3538 (aff_combination_convert): Likewise.
3539 (tree_to_aff_combination): Likewise.
3540 (add_elt_to_tree): Take a widest_int instead of a double_int.
3541 (aff_combination_to_tree): Use wide-int interfaces.
3542 (aff_combination_remove_elt): Likewise.
3543 (aff_combination_add_product): Take a widest_int instead of
3545 (aff_combination_mult): Use wide-int interfaces.
3546 (aff_combination_expand): Likewise.
3547 (double_int_constant_multiple_p): Delete.
3548 (wide_int_constant_multiple_p): New.
3549 (aff_combination_constant_multiple_p): Take a widest_int pointer
3550 instead of a double_int pointer.
3551 (print_aff): Use wide-int interfaces.
3552 (get_inner_reference_aff): Take a widest_int pointer
3553 instead of a double_int pointer.
3554 (aff_comb_cannot_overlap_p): Take widest_ints instead of double_ints.
3555 * tree-affine.h: Include wide-int.h.
3556 (struct aff_comb_elt): Change type of coef to widest_int.
3557 (struct affine_tree_combination): Change type of offset to widest_int.
3558 (double_int_ext_for_comb): Delete.
3559 (wide_int_ext_for_comb): New.
3560 (aff_combination_const): Use widest_int instead of double_int.
3561 (aff_combination_scale): Likewise.
3562 (aff_combination_add_elt): Likewise.
3563 (aff_combination_constant_multiple_p): Likewise.
3564 (get_inner_reference_aff): Likewise.
3565 (aff_comb_cannot_overlap_p): Likewise.
3566 (aff_combination_zero_p): Use wide-int interfaces.
3567 * tree.c: Include tree.h.
3568 (init_ttree): Use make_int_cst.
3569 (tree_code_size): Removed code for INTEGER_CST case.
3570 (tree_size): Add INTEGER_CST case.
3571 (make_node_stat): Update comment.
3572 (get_int_cst_ext_nunits, build_new_int_cst, build_int_cstu): New.
3573 (build_int_cst_type): Use wide-int interfaces.
3574 (double_int_to_tree): Likewise.
3575 (double_int_fits_to_tree_p): Delete.
3576 (force_fit_type_double): Delete.
3577 (force_fit_type): New.
3578 (int_cst_hash_hash): Use wide-int interfaces.
3579 (int_cst_hash_eq): Likewise.
3580 (build_int_cst_wide): Delete.
3581 (wide_int_to_tree): New.
3582 (cache_integer_cst): Use wide-int interfaces.
3583 (build_low_bits_mask): Likewise.
3584 (cst_and_fits_in_hwi): Likewise.
3585 (real_value_from_int_cst): Likewise.
3586 (make_int_cst_stat): New.
3587 (integer_zerop): Use wide_int interfaces.
3588 (integer_onep): Likewise.
3589 (integer_all_onesp): Likewise.
3590 (integer_pow2p): Likewise.
3591 (integer_nonzerop): Likewise.
3592 (tree_log2): Likewise.
3593 (tree_floor_log2): Likewise.
3594 (tree_ctz): Likewise.
3595 (int_size_in_bytes): Likewise.
3596 (mem_ref_offset): Return an offset_int rather than a double_int.
3597 (build_type_attribute_qual_variant): Use wide_int interfaces.
3598 (type_hash_eq): Likewise
3599 (tree_int_cst_equal): Likewise.
3600 (tree_int_cst_lt): Delete.
3601 (tree_int_cst_compare): Likewise.
3602 (tree_fits_shwi_p): Use wide_int interfaces.
3603 (tree_fits_uhwi_p): Likewise.
3604 (tree_int_cst_sign_bit): Likewise.
3605 (tree_int_cst_sgn): Likewise.
3606 (tree_int_cst_min_precision): Take a signop rather than a bool.
3607 (simple_cst_equal): Use wide_int interfaces.
3608 (compare_tree_int): Likewise.
3609 (iterative_hash_expr): Likewise.
3610 (int_fits_type_p): Likewise. Use tree_int_cst_lt rather than
3612 (get_type_static_bounds): Use wide_int interfaces.
3613 (tree_int_cst_elt_check_failed): New.
3614 (build_common_tree_nodes): Reordered to set prec before filling in
3616 (int_cst_value): Check cst_and_fits_in_hwi.
3617 (widest_int_cst_value): Use wide_int interfaces.
3618 (upper_bound_in_type): Likewise.
3619 (lower_bound_in_type): Likewise.
3620 (num_ending_zeros): Likewise.
3621 (drop_tree_overflow): Likewise.
3622 * tree-call-cdce.c (check_pow): Update call to real_from_integer.
3623 (gen_conditions_for_pow_cst_base): Likewise.
3624 * tree-cfg.c: Include wide-int.h and wide-int-print.h.
3625 (group_case_labels_stmt): Use wide-int interfaces.
3626 (verify_gimple_assign_binary): Likewise.
3627 (print_loop): Likewise.
3628 * tree-chrec.c (tree_fold_binomial): Likewise.
3629 * tree-core.h (struct tree_base): Add int_length.
3630 (struct tree_int_cst): Change rep of value.
3631 * tree-data-ref.c (dr_analyze_innermost): Use wide-int interfaces.
3632 (dr_may_alias_p): Likewise.
3633 (max_stmt_executions_tree): Likewise.
3634 * tree.def (INTEGER_CST): Update comment.
3635 * tree-dfa.c (get_ref_base_and_extent): Use wide-int interfaces.
3636 * tree-dfa.h (get_addr_base_and_unit_offset_1): Likewise.
3637 * tree-dump.c: Include wide-int.h and wide-int-print.h.
3638 (dequeue_and_dump): Use wide-int interfaces.
3639 * tree.h: Include wide-int.h.
3640 (NULL_TREE): Moved to earlier loc in file.
3641 (TREE_INT_CST_ELT_CHECK): New.
3642 (tree_int_cst_elt_check_failed): New.
3644 (TREE_INT_CST): Delete.
3645 (TREE_INT_CST_LOW): Use wide-int interfaces.
3646 (TREE_INT_CST_HIGH): Delete.
3647 (TREE_INT_CST_NUNITS): New.
3648 (TREE_INT_CST_EXT_NUNITS): Likewise.
3649 (TREE_INT_CST_OFFSET_NUNITS): Likewise.
3650 (TREE_INT_CST_ELT): Likewise.
3651 (INT_CST_LT): Delete.
3652 (tree_int_cst_elt_check): New (two forms).
3653 (type_code_size): Update comment.
3654 (make_int_cst_stat, make_int_cst): New.
3655 (tree_to_double_int): Delete.
3656 (double_int_fits_to_tree_p): Delete.
3657 (force_fit_type_double): Delete.
3658 (build_int_cstu): Replace with out-of-line function.
3659 (build_int_cst_wide): Delete.
3660 (tree_int_cst_lt): Define inline.
3661 (tree_int_cst_le): New.
3662 (tree_int_cst_compare): Define inline.
3663 (tree_int_cst_min_precision): Take a signop rather than a bool.
3664 (wi::int_traits <const_tree>): New.
3665 (wi::int_traits <tree>): New.
3666 (wi::extended_tree): New.
3667 (wi::int_traits <wi::extended_tree>): New.
3668 (wi::to_widest): New.
3669 (wi::to_offset): New.
3670 (wi::fits_to_tree_p): New.
3671 (wi::min_value): New.
3672 (wi::max_value): New.
3673 * tree-inline.c (remap_gimple_op_r): Use wide-int interfaces.
3674 (copy_tree_body_r): Likewise.
3675 * tree-object-size.c (compute_object_offset): Likewise.
3676 (addr_object_size): Likewise.
3677 * tree-predcom.c: Include wide-int-print.h.
3678 (struct dref_d): Change type of offset to widest_int.
3679 (dump_dref): Call wide-int printer.
3680 (aff_combination_dr_offset): Use wide-int interfaces.
3681 (determine_offset): Take a widest_int pointer rather than a
3683 (split_data_refs_to_components): Use wide-int interfaces.
3684 (suitable_component_p): Likewise.
3685 (order_drefs): Likewise.
3686 (add_ref_to_chain): Likewise.
3687 (valid_initializer_p): Likewise.
3688 (determine_roots_comp): Likewise.
3689 * tree-pretty-print.c: Include wide-int-print.h.
3690 (dump_generic_node): Use wide-int interfaces.
3691 * tree-sra.c (sra_ipa_modify_expr): Likewise.
3692 * tree-ssa-address.c (addr_for_mem_ref): Likewise.
3693 (move_fixed_address_to_symbol): Likewise.
3694 (move_hint_to_base): Likewise.
3695 (move_pointer_to_base): Likewise.
3696 (move_variant_to_index): Likewise.
3697 (most_expensive_mult_to_index): Likewise.
3698 (addr_to_parts): Likewise.
3699 (copy_ref_info): Likewise.
3700 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
3701 (indirect_refs_may_alias_p): Likewise.
3702 (stmt_kills_ref_p_1): Likewise.
3703 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
3704 * tree-ssa-ccp.c: Update comment at top of file. Include
3706 (struct prop_value_d): Change type of mask to widest_int.
3707 (extend_mask): New function.
3708 (dump_lattice_value): Use wide-int interfaces.
3709 (get_default_value): Likewise.
3710 (set_constant_value): Likewise.
3711 (set_value_varying): Likewise.
3712 (valid_lattice_transition): Likewise.
3713 (set_lattice_value): Likewise.
3714 (value_to_double_int): Delete.
3715 (value_to_wide_int): New.
3716 (get_value_from_alignment): Use wide-int interfaces.
3717 (get_value_for_expr): Likewise.
3718 (do_dbg_cnt): Likewise.
3719 (ccp_finalize): Likewise.
3720 (ccp_lattice_meet): Likewise.
3721 (bit_value_unop_1): Use widest_ints rather than double_ints.
3722 (bit_value_binop_1): Likewise.
3723 (bit_value_unop): Use wide-int interfaces.
3724 (bit_value_binop): Likewise.
3725 (bit_value_assume_aligned): Likewise.
3726 (evaluate_stmt): Likewise.
3727 (ccp_fold_stmt): Likewise.
3728 (visit_cond_stmt): Likewise.
3729 (ccp_visit_stmt): Likewise.
3730 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Likewise.
3731 (constant_pointer_difference): Likewise.
3732 (associate_pointerplus): Likewise.
3733 (combine_conversions): Likewise.
3734 * tree-ssa-loop.h: Include wide-int.h.
3735 (struct tree_niter_desc): Change type of max to widest_int.
3736 * tree-ssa-loop-im.c (mem_refs_may_alias_p): Use wide-int interfaces.
3737 * tree-ssa-loop-ivcanon.c (remove_exits_and_undefined_stmts): Likewise.
3738 (remove_redundant_iv_tests): Likewise.
3739 (canonicalize_loop_induction_variables): Likewise.
3740 * tree-ssa-loop-ivopts.c (alloc_iv): Likewise.
3741 (constant_multiple_of): Take a widest_int pointer instead of
3742 a double_int pointer.
3743 (get_computation_aff): Use wide-int interfaces.
3744 (ptr_difference_cost): Likewise.
3745 (difference_cost): Likewise.
3746 (get_loop_invariant_expr_id): Likewise.
3747 (get_computation_cost_at): Likewise.
3748 (iv_elimination_compare_lt): Likewise.
3749 (may_eliminate_iv): Likewise.
3750 * tree-ssa-loop-niter.h (estimated_loop_iterations): Use widest_int
3751 instead of double_int.
3752 (max_loop_iterations): Likewise.
3753 (max_stmt_executions): Likewise.
3754 (estimated_stmt_executions): Likewise.
3755 * tree-ssa-loop-niter.c: Include wide-int-print.h.
3756 (split_to_var_and_offset): Use wide-int interfaces.
3757 (determine_value_range): Likewise.
3758 (bound_difference_of_offsetted_base): Likewise.
3759 (bounds_add): Take a widest_int instead of a double_int.
3760 (number_of_iterations_ne_max): Use wide-int interfaces.
3761 (number_of_iterations_ne): Likewise.
3762 (number_of_iterations_lt_to_ne): Likewise.
3763 (assert_loop_rolls_lt): Likewise.
3764 (number_of_iterations_lt): Likewise.
3765 (number_of_iterations_le): Likewise.
3766 (number_of_iterations_cond): Likewise.
3767 (number_of_iterations_exit): Likewise.
3768 (finite_loop_p): Likewise.
3769 (derive_constant_upper_bound_assign): Likewise.
3770 (derive_constant_upper_bound): Return a widest_int.
3771 (derive_constant_upper_bound_ops): Likewise.
3772 (do_warn_aggressive_loop_optimizations): Use wide-int interfaces.
3773 (record_estimate): Take a widest_int rather than a double_int.
3774 (record_nonwrapping_iv): Use wide-int interfaces.
3775 (double_int_cmp): Delete.
3776 (wide_int_cmp): New.
3777 (bound_index): Take a widest_int rather than a double_int.
3778 (discover_iteration_bound_by_body_walk): Use wide-int interfaces.
3779 (maybe_lower_iteration_bound): Likewise.
3780 (estimate_numbers_of_iterations_loop): Likewise.
3781 (estimated_loop_iterations): Take a widest_int pointer than than
3782 a double_int pointer.
3783 (estimated_loop_iterations_int): Use wide-int interfaces.
3784 (max_loop_iterations): Take a widest_int pointer than than
3785 a double_int pointer.
3786 (max_loop_iterations_int): Use wide-int interfaces.
3787 (max_stmt_executions): Take a widest_int pointer than than
3788 a double_int pointer.
3789 (estimated_stmt_executions): Likewise.
3790 (n_of_executions_at_most): Use wide-int interfaces.
3791 (scev_probably_wraps_p): Likewise.
3792 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Update calls
3794 * tree-scalar-evolution.c (simplify_peeled_chrec): Use wide-int
3796 * tree-ssanames.c (set_range_info): Use wide_int_refs rather than
3797 double_ints. Adjust for trailing_wide_ints <3> representation.
3798 (set_nonzero_bits): Likewise.
3799 (get_range_info): Return wide_ints rather than double_ints.
3800 Adjust for trailing_wide_ints <3> representation.
3801 (get_nonzero_bits): Likewise.
3802 (duplicate_ssa_name_range_info): Adjust for trailing_wide_ints <3>
3804 * tree-ssanames.h (struct range_info_def): Replace min, max and
3805 nonzero_bits with a trailing_wide_ints <3>.
3806 (set_range_info): Use wide_int_refs rather than double_ints.
3807 (set_nonzero_bits): Likewise.
3808 (get_range_info): Return wide_ints rather than double_ints.
3809 (get_nonzero_bits): Likewise.
3810 * tree-ssa-phiopt.c (jump_function_from_stmt): Use wide-int interfaces.
3811 * tree-ssa-pre.c (phi_translate_1): Likewise.
3812 * tree-ssa-reassoc.c (decrement_power): Use calls to real_from_integer.
3813 (acceptable_pow_call): Likewise.
3814 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Use wide-int
3816 (vn_reference_fold_indirect): Likewise.
3817 (vn_reference_maybe_forwprop_address): Likewise.
3818 (valueize_refs_1): Likewise.
3819 * tree-ssa-structalias.c (get_constraint_for_ptr_offset): Likewise.
3820 * tree-ssa-uninit.c (is_value_included_in): Use wide-int interfaces,
3821 tree_int_cst_lt and tree_int_cst_le.
3822 * tree-streamer-in.c (unpack_ts_base_value_fields): Use wide-int
3824 (streamer_alloc_tree): Likewise.
3825 * tree-streamer-out.c (pack_ts_int_cst_value_fields): Likewise.
3826 (streamer_write_tree_header): Likewise.
3827 (streamer_write_integer_cst): Likewise.
3828 * tree-switch-conversion.c (emit_case_bit_tests): Likewise.
3829 (build_constructors): Likewise.
3830 (array_value_type): Likewise.
3831 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list): Likewise.
3832 (vect_check_gather): Likewise.
3833 * tree-vect-generic.c (build_replicated_const): Likewise.
3834 (expand_vector_divmod): Likewise.
3835 * tree-vect-loop.c (vect_transform_loop): Likewise.
3836 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
3837 (vect_do_peeling_for_alignment): Likewise.
3838 * tree-vect-patterns.c (vect_recog_divmod_pattern): Likewise.
3839 * tree-vrp.c: Include wide-int.h.
3840 (operand_less_p): Use wide-int interfaces and tree_int_cst_lt.
3841 (extract_range_from_assert): Use wide-int interfaces.
3842 (vrp_int_const_binop): Likewise.
3843 (zero_nonzero_bits_from_vr): Take wide_int pointers rather than
3844 double_int pointers.
3845 (ranges_from_anti_range): Use wide-int interfaces.
3846 (quad_int_cmp): Delete.
3847 (quad_int_pair_sort): Likewise.
3848 (extract_range_from_binary_expr_1): Use wide-int interfaces.
3849 (extract_range_from_unary_expr_1): Likewise.
3850 (adjust_range_with_scev): Likewise.
3851 (masked_increment): Take and return wide_ints rather than double_ints.
3852 (register_edge_assert_for_2): Use wide-int interfaces.
3853 (check_array_ref): Likewise.
3854 (search_for_addr_array): Likewise.
3855 (maybe_set_nonzero_bits): Likewise.
3856 (union_ranges): Pass an integer of the correct type instead of
3857 using integer_one_node.
3858 (intersect_ranges): Likewise.
3859 (simplify_truth_ops_using_ranges): Likewise.
3860 (simplify_bit_ops_using_ranges): Use wide-int interfaces.
3861 (range_fits_type_p): Likewise.
3862 (simplify_cond_using_ranges): Likewise. Take a signop rather than
3864 (simplify_conversion_using_ranges): Use wide-int interfaces.
3865 (simplify_float_conversion_using_ranges): Likewise.
3866 (vrp_finalize): Likewise.
3867 * value-prof.c (gimple_divmod_fixed_value_transform): Likewise.
3868 (gimple_stringops_transform): Likewise.
3869 * varasm.c (decode_addr_const): Likewise.
3870 (const_hash_1): Likewise.
3871 (const_rtx_hash_1): Likewise
3872 (output_constant): Likewise.
3873 (array_size_for_constructor): Likewise.
3874 (output_constructor_regular_field): Likewise.
3875 (output_constructor_bitfield): Likewise.
3876 * var-tracking.c (loc_cmp): Handle CONST_WIDE_INT.
3877 * mkconfig.sh: Include machmode.h to pick up BITS_PER_UNIT for
3879 * gencheck.c: Define BITS_PER_UNIT.
3882 * wide-int-print.cc: New.
3883 * wide-int-print.h: New.
3885 2014-05-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3887 * config/avr/avr.c (avr_can_eliminate): Mark unused argument.
3889 2014-05-06 Richard Biener <rguenther@suse.de>
3891 * tree-pass.h (TODO_verify_ssa, TODO_verify_flow,
3892 TODO_verify_stmts, TODO_verify_rtl_sharing): Remove.
3893 (TODO_verify_all): Adjust.
3894 * asan.c: Remove references to TODO_verify_ssa, TODO_verify_flow,
3895 TODO_verify_stmts and TODO_verify_rtl_sharing.
3896 * bb-reorder.c: Likewise.
3897 * cfgexpand.c: Likewise.
3898 * cprop.c: Likewise.
3900 * function.c: Likewise.
3901 * fwprop.c: Likewise.
3903 * gimple-ssa-isolate-paths.c: Likewise.
3904 * gimple-ssa-strength-reduction.c: Likewise.
3905 * ipa-split.c: Likewise.
3906 * loop-init.c: Likewise.
3907 * loop-unroll.c: Likewise.
3908 * lower-subreg.c: Likewise.
3909 * modulo-sched.c: Likewise.
3910 * postreload-gcse.c: Likewise.
3911 * predict.c: Likewise.
3912 * recog.c: Likewise.
3913 * sched-rgn.c: Likewise.
3914 * store-motion.c: Likewise.
3915 * tracer.c: Likewise.
3916 * trans-mem.c: Likewise.
3917 * tree-call-cdce.c: Likewise.
3918 * tree-cfg.c: Likewise.
3919 * tree-cfgcleanup.c: Likewise.
3920 * tree-complex.c: Likewise.
3921 * tree-eh.c: Likewise.
3922 * tree-emutls.c: Likewise.
3923 * tree-if-conv.c: Likewise.
3924 * tree-into-ssa.c: Likewise.
3925 * tree-loop-distribution.c: Likewise.
3926 * tree-object-size.c: Likewise.
3927 * tree-parloops.c: Likewise.
3928 * tree-pass.h: Likewise.
3929 * tree-sra.c: Likewise.
3930 * tree-ssa-ccp.c: Likewise.
3931 * tree-ssa-copy.c: Likewise.
3932 * tree-ssa-copyrename.c: Likewise.
3933 * tree-ssa-dce.c: Likewise.
3934 * tree-ssa-dom.c: Likewise.
3935 * tree-ssa-dse.c: Likewise.
3936 * tree-ssa-forwprop.c: Likewise.
3937 * tree-ssa-ifcombine.c: Likewise.
3938 * tree-ssa-loop-ch.c: Likewise.
3939 * tree-ssa-loop-ivcanon.c: Likewise.
3940 * tree-ssa-loop.c: Likewise.
3941 * tree-ssa-math-opts.c: Likewise.
3942 * tree-ssa-phiopt.c: Likewise.
3943 * tree-ssa-phiprop.c: Likewise.
3944 * tree-ssa-pre.c: Likewise.
3945 * tree-ssa-reassoc.c: Likewise.
3946 * tree-ssa-sink.c: Likewise.
3947 * tree-ssa-strlen.c: Likewise.
3948 * tree-ssa-tail-merge.c: Likewise.
3949 * tree-ssa-uncprop.c: Likewise.
3950 * tree-switch-conversion.c: Likewise.
3951 * tree-tailcall.c: Likewise.
3952 * tree-vect-generic.c: Likewise.
3953 * tree-vectorizer.c: Likewise.
3954 * tree-vrp.c: Likewise.
3956 * var-tracking.c: Likewise.
3957 * bt-load.c: Likewise.
3958 * cfgcleanup.c: Likewise.
3959 * combine-stack-adj.c: Likewise.
3960 * combine.c: Likewise.
3961 * compare-elim.c: Likewise.
3962 * config/epiphany/resolve-sw-modes.c: Likewise.
3963 * config/i386/i386.c: Likewise.
3964 * config/mips/mips.c: Likewise.
3965 * config/s390/s390.c: Likewise.
3966 * config/sh/sh_treg_combine.cc: Likewise.
3967 * config/sparc/sparc.c: Likewise.
3970 * final.c: Likewise.
3971 * ifcvt.c: Likewise.
3972 * mode-switching.c: Likewise.
3973 * passes.c: Likewise.
3974 * postreload.c: Likewise.
3976 * reg-stack.c: Likewise.
3977 * regcprop.c: Likewise.
3978 * regrename.c: Likewise.
3981 2014-05-06 Richard Biener <rguenther@suse.de>
3984 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
3985 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
3987 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
3990 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
3992 2014-05-05 Radovan Obradovic <robradovic@mips.com>
3993 Tom de Vries <tom@codesourcery.com>
3995 * target.def (call_fusage_contains_non_callee_clobbers): New
3997 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
3999 (@node Miscellaneous Register Hooks): New node.
4000 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
4001 * doc/tm.texi: Regenerate.
4003 2014-05-05 Marek Polacek <polacek@redhat.com>
4006 * opts.c (common_handle_option): Call error_at instead of warning_at.
4008 2014-05-05 Richard Biener <rguenther@suse.de>
4010 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4011 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
4012 under the TODO_verify_il umbrella.
4014 2014-05-05 Richard Biener <rguenther@suse.de>
4016 * passes.c (execute_function_todo): Move TODO_verify_flow under
4017 the TODO_verify_ul umbrella.
4019 2014-05-05 Richard Biener <rguenther@suse.de>
4022 * fold-const.c (fold_binary_loc): Consistently avoid canonicalizing
4023 X & CST away from a CST that is the mask of a mode.
4025 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4027 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
4028 int argument to enum machine_mode.
4029 (picochip_class_max_nregs): Ditto.
4030 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
4031 (picochip_class_max_nregs): Ditto.
4033 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4035 * target.def: Add new target hook.
4036 * doc/tm.texi: Regenerate.
4037 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
4038 * targhooks.c (default_keep_leaf_when_profiled): New function.
4040 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
4041 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
4043 2014-05-05 Bin Cheng <bin.cheng@arm.com>
4045 PR tree-optimization/60363
4046 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
4047 (copy_phi_args): New parameters. Call get_value_locus_in_path.
4048 (update_destination_phis): New parameter.
4049 (create_edge_and_update_destination_phis): Ditto.
4050 (ssa_fix_duplicate_block_edges): Pass new arguments.
4051 (thread_single_edge): Ditto.
4053 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
4055 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
4056 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
4057 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
4058 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
4059 Use RS6000_BTM_HARD_FLOAT.
4060 (BU_MISC_2): Likewise.
4061 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
4062 RS6000_BTM_HARD_FLOAT.
4063 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
4065 (rs6000_invalid_builtin): Add hard floating builtin support.
4066 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
4067 hard float builtins.
4068 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
4070 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4072 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
4073 Add missing function* argument.
4075 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4077 * lra-constraints.c (valid_address_p): Move earlier in file.
4078 Add a constraint argument to the address_info version.
4079 (satisfies_memory_constraint_p): New function.
4080 (satisfies_address_constraint_p): Likewise.
4081 (process_alt_operands, curr_insn_transform): Use them.
4082 (process_address): Pass the constraint to valid_address_p when
4083 checking address operands.
4085 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
4087 * config/mips/mips.c (mips_isa_rev): New variable.
4088 (mips_set_architecture): Set it.
4089 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
4091 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
4092 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
4093 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
4094 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
4095 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
4096 conditions in terms of mips_isa_rev.
4097 (mips_isa_rev): Declare.
4099 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4101 * config/sh/sh-mem.cc: Use tabs instead of spaces.
4102 (prob_unlikely, prob_likely): Make variables const.
4104 2014-05-03 Denis Chertykov <chertykov@gmail.com>
4106 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
4108 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4110 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4112 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4114 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
4115 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
4116 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
4118 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
4119 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
4121 Replace usage of ROUND_REG with sh_round_reg.
4122 Use CEIL instead of ROUND_ADVANCE.
4124 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
4127 * config/sh/sh.c: Include stdlib headers before everything else.
4129 2014-05-02 Jakub Jelinek <jakub@redhat.com>
4131 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
4132 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
4133 (gimplify_adjust_omp_clauses): Simd region is never
4134 directly nested in combined parallel. Instead, for linear
4135 with copyin/copyout, if in combined for simd loop, make decl
4136 firstprivate/lastprivate on OMP_FOR.
4137 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
4138 expand_omp_for_static_chunk): When setting endvar, also set
4139 fd->loop.v to the same value.
4141 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
4143 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
4145 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
4147 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
4150 2014-05-02 Marek Polacek <polacek@redhat.com>
4152 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4154 2014-05-02 Kito Cheng <kito@0xlab.org>
4156 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
4157 to a C expression marco.
4158 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
4159 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
4160 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
4161 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
4162 HONOR_REG_ALLOC_ORDER.
4163 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
4165 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4167 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4169 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4171 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
4173 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
4175 * tree-if-conv.c (is_cond_scalar_reduction): New function.
4176 (convert_scalar_cond_reduction): Likewise.
4177 (predicate_scalar_phi): Add recognition and transformation
4178 of simple conditioanl reduction to be vectorizable.
4180 2014-05-01 Marek Polacek <polacek@redhat.com>
4183 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
4185 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
4187 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
4188 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
4189 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
4190 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
4191 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
4192 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
4193 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
4194 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
4196 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
4198 * config/arc/arc.opt (mlra): Move comment above option name
4199 to avoid mis-parsing as language options.
4201 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4203 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
4204 * config/sol2.h: ... here.
4205 * config/sol2-10.h: Remove.
4207 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
4208 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
4209 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
4210 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
4211 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
4212 * config/sol2.h: ... here.
4213 (SECTION_NAME_FORMAT): Don't redefine.
4214 (STARTFILE_ARCH32_SPEC): Rename to ...
4215 (STARTFILE_ARCH_SPEC): ... this.
4216 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
4217 * config/sparc/sol2.h: ... here.
4218 (SECTION_NAME_FORMAT): Don't undef.
4219 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
4220 (SUBTARGET_EXTRA_SPECS): Remove.
4221 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
4223 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
4224 (MD_STARTFILE_PREFIX): Remove.
4225 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
4226 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
4227 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
4228 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
4229 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
4230 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
4231 * config/i386/sol2.h: ... here.
4232 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
4233 * config/i386/sol2-bi.h: Remove.
4234 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
4235 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
4237 * config/i386/t-sol2-64: Rename to ...
4238 * config/i386/t-sol2: ... this.
4239 * config/sparc/t-sol2-64: Rename to ...
4240 * config/sparc/t-sol2: ... this.
4242 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
4243 sol2_tm_file_head, sol2_tm_file_tail.
4244 Include ${cpu_type}/sol2.h before sol2.h.
4246 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
4247 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
4248 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
4249 Reflect i386/t-sol2-64 renaming.
4250 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
4251 Reflect sparc/t-sol2-64 renaming.
4253 2014-04-30 Richard Biener <rguenther@suse.de>
4255 * passes.c (execute_function_todo): Move TODO_verify_stmts
4256 and TODO_verify_ssa under the TODO_verify_il umbrella.
4257 * tree-ssa.h (verify_ssa): Adjust prototype.
4258 * tree-ssa.c (verify_ssa): Add parameter to tell whether
4259 we should verify SSA operands.
4260 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
4261 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
4262 whether we should verify whether not throwing stmts have EH info.
4263 * graphite-scop-detection.c (create_sese_edges): Adjust.
4264 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
4265 * tree-eh.c (lower_try_finally_switch): Do not add the
4266 default case label twice.
4268 2014-04-30 Marek Polacek <polacek@redhat.com>
4270 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4271 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
4272 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
4273 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
4275 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
4277 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4278 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
4279 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
4280 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
4281 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
4282 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
4283 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
4284 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
4286 2014-04-29 David Malcolm <dmalcolm@redhat.com>
4288 * tree-cfg.c (dump_function_to_file): Dump the return type of
4289 functions, in a line to itself before the function body, mimicking
4290 the layout of a C function.
4292 2014-04-29 Jakub Jelinek <jakub@redhat.com>
4294 PR tree-optimization/60971
4295 * tree-tailcall.c (process_assignment): Reject conversions which
4298 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
4300 * calls.c (initialize_argument_information): Always treat
4301 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
4302 (expand_call): Likewise.
4303 (emit_library_call_calue_1): Likewise.
4304 * expr.c (PUSH_ARGS_REVERSED): Do not define.
4305 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
4308 2014-04-29 Nick Clifton <nickc@redhat.com>
4310 * config/msp430/msp430.md (umulsidi): Fix typo.
4311 (mulhisi3): Enable even inside interrupt handlers.
4312 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
4313 bigger return address pushed in large mode.
4315 2014-04-29 Nick Clifton <nickc@redhat.com>
4317 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4318 (arc_init_reg_tables): Use a machine_mode enum to iterate over
4320 * config/m32r/m32r.c (init_reg_tables): Likewise.
4321 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
4322 enum to hold the modes.
4324 2014-04-29 Richard Biener <rguenther@suse.de>
4326 * dominance.c (free_dominance_info): Add overload with
4328 (dom_info_state): Likewise.
4329 (dom_info_available_p): Likewise.
4330 * basic-block.h (free_dominance_info, dom_info_state,
4331 dom_info_available_p): Declare overloads.
4332 * passes.c (execute_function_todo): Verify that verifiers
4333 don't change dominator info state. Drop dominator info
4334 for IPA pass invocations.
4335 * cgraph.c (release_function_body): Restore asserts that
4336 dominator information is released.
4338 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
4340 * doc/invoke.texi: Fix typo.
4341 * tree-vrp.c: Fix typos.
4342 * gimple.c (infer_nonnull_range): Reorder operands of an && condition.
4344 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4346 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
4348 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4350 * config/aarch64/aarch64-builtins.c
4351 (aarch64_types_storestruct_lane_qualifiers): New.
4352 (TYPES_STORESTRUCT_LANE): Likewise.
4353 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
4354 (st3_lane): Likewise.
4355 (st4_lane): Likewise.
4356 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
4357 (vec_store_lanesci_lane<mode>): Likewise.
4358 (vec_store_lanesxi_lane<mode>): Likewise.
4359 (aarch64_st2_lane<VQ:mode>): Likewise.
4360 (aarch64_st3_lane<VQ:mode>): Likewise.
4361 (aarch64_st4_lane<VQ:mode>): Likewise.
4362 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
4363 * config/aarch64/arm_neon.h
4364 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
4365 use new macro arguments.
4366 (__ST3_LANE_FUNC): Likewise.
4367 (__ST4_LANE_FUNC): Likewise.
4368 * config/aarch64/iterators.md (V_TWO_ELEM): New.
4369 (V_THREE_ELEM): Likewise.
4370 (V_FOUR_ELEM): Likewise.
4372 2014-04-28 David Malcolm <dmalcolm@redhat.com>
4374 * doc/gimple.texi: Replace the description of the now-defunct
4375 union gimple_statement_d with a diagram showing the
4376 gimple_statement_base class hierarchy and its relationships to
4377 the GSS_ and GIMPLE_ enums.
4379 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
4381 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4382 * config/aarch64/aarch64.c
4383 (aarch64_cannot_change_mode_class): Weaken conditions.
4384 (aarch64_modes_tieable_p): New.
4385 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
4387 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
4389 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
4390 (loadsync_<mode>): Change mode.
4391 (load_quadpti, store_quadpti): New.
4392 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
4393 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
4395 2014-04-28 Martin Jambor <mjambor@suse.cz>
4397 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
4398 same alias type as the original statement.
4399 (subreplacement_assignment_data): New type.
4400 (handle_unscalarized_data_in_subtree): New type of parameter,
4401 generate new memory accesses with same alias type as the original
4403 (load_assign_lhs_subreplacements): Likewise.
4404 (sra_modify_constructor_assign): Generate new memory accesses with
4405 same alias type as the original statement.
4407 2014-04-28 Richard Biener <rguenther@suse.de>
4409 * tree-pass.h (TODO_verify_il): Define.
4410 (TODO_verify_all): Complete properly.
4411 * passes.c (execute_function_todo): Move existing loop-closed
4412 SSA verification under TODO_verify_il.
4413 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
4414 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
4415 Fix tree sharing issue.
4417 2014-04-28 Richard Biener <rguenther@suse.de>
4420 * builtins.def (DEF_C11_BUILTIN): Add.
4421 (BUILT_IN_ALIGNED_ALLOC): Likewise.
4422 * coretypes.h (enum function_class): Add function_c11_misc.
4423 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
4424 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
4425 (call_may_clobber_ref_p_1): Likewise.
4426 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
4427 (mark_all_reaching_defs_necessary_1): Likewise.
4428 (propagate_necessity): Likewise.
4429 (eliminate_unnecessary_stmts): Likewise.
4430 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
4432 2014-04-28 Richard Biener <rguenther@suse.de>
4434 * tree-vrp.c (vrp_var_may_overflow): Remove.
4435 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
4436 with overflow immediately bump to one before that value and
4437 let iteration figure out overflow status.
4439 2014-04-28 Richard Biener <rguenther@suse.de>
4441 * configure.ac: Do valgrind header checks unconditionally.
4442 Add --enable-valgrind-annotations.
4443 * system.h: Guard valgrind header inclusion with
4444 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
4445 * alloc-pool.c (pool_alloc, pool_free): Use
4446 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
4447 to guard possibly dead code.
4448 * config.in: Regenerated.
4449 * configure: Likewise.
4451 2014-04-28 Jeff Law <law@redhat.com>
4453 PR tree-optimization/60902
4454 * tree-ssa-threadedge.c
4455 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
4456 over real defs when invalidating outputs from statements that do not
4457 produce useful outputs for threading.
4459 2014-04-28 Richard Biener <rguenther@suse.de>
4461 PR tree-optimization/60979
4462 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
4463 SCOPs that end in a block with a successor with abnormal
4466 2014-04-28 Richard Biener <rguenther@suse.de>
4468 * tree-pass.h (execute_pass_list): Adjust prototype.
4469 * passes.c (pass_manager::execute_early_local_passes): Adjust.
4470 (do_per_function): Change callback signature, push all actual
4471 work to the callbals.
4472 (do_per_function_toporder): Likewise.
4473 (execute_function_dump): Adjust.
4474 (execute_function_todo): Likewise.
4475 (clear_last_verified): Likewise.
4476 (verify_curr_properties): Likewise.
4477 (update_properties_after_pass): Likewise.
4478 (execute_pass_list_1): Split out from ...
4479 (execute_pass_list): ... here. Adjust.
4480 (execute_ipa_pass_list): Likewise.
4481 * cgraphunit.c (cgraph_add_new_function): Adjust.
4482 (analyze_function): Likewise.
4483 (expand_function): Likewise.
4484 * cgraph.c (release_function_body): Free dominance info
4485 here instead of asserting it was magically freed elsewhere.
4487 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
4489 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
4490 * configure: Regenerate.
4491 * config/sparc/sparc.opt (muser-mode): New option.
4492 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
4494 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
4495 * doc/invoke.texi (SPARC options): Document -muser-mode.
4497 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
4499 * cselib.c (find_slot_memmode): Delete.
4500 (cselib_hasher): Change compare_type to a struct.
4501 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
4503 (preserve_constants_and_equivs): Adjust for new compare_type.
4504 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
4505 (wrap_constant): Delete.
4506 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
4508 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
4510 * doc/install.texi (Building with profile feedback): Remove
4513 2014-04-26 Tom de Vries <tom@codesourcery.com>
4515 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
4518 2014-04-25 Cary Coutant <ccoutant@google.com>
4521 * dwarf2out.c (should_move_die_to_comdat): A type definition
4522 can contain a subprogram definition, but don't move it to a
4524 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
4525 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
4527 (clone_tree_hash): Rename to...
4528 (clone_tree_partial): ...this; change callers. Copy
4529 DW_TAG_subprogram DIEs as declarations.
4530 (copy_decls_walk): Don't copy children of a declaration into a
4533 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
4536 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
4539 2014-04-25 Jiong Wang <jiong.wang@arm.com>
4541 * config/arm/predicates.md (call_insn_operand): Add long_call check.
4542 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
4544 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
4547 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4549 * config/arm/arm.c (arm_cortex_a8_tune): Initialise T16-related fields.
4551 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4553 PR tree-optimization/60930
4554 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
4555 creating a multiply candidate by folding two constant
4556 multiplicands when the result overflows.
4558 2014-04-25 Jakub Jelinek <jakub@redhat.com>
4560 PR tree-optimization/60960
4561 * tree-vect-generic.c (expand_vector_operation): Only call
4562 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
4564 2014-04-25 Tom de Vries <tom@codesourcery.com>
4566 * expr.c (clobber_reg_mode): New function.
4567 * expr.h (clobber_reg): New function.
4569 2014-04-25 Tom de Vries <tom@codesourcery.com>
4571 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
4574 2014-04-25 Radovan Obradovic <robradovic@mips.com>
4575 Tom de Vries <tom@codesourcery.com>
4577 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
4579 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
4580 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
4581 new argument to find_all_hard_reg_sets call.
4583 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4585 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
4586 Use HOST_WIDE_INT_C for mask literal.
4587 (aarch_rev16_shleft_mask_imm_p): Likewise.
4589 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
4592 * config/sparc/sparc.md (ashlsi3_extend): Delete.
4594 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
4596 PR preprocessor/56540
4597 * config/i386/i386-c.c (ix86_target_macros): Define
4598 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
4600 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4602 * configure.ac (tga_func): Remove.
4603 (LIB_TLS_SPEC): Remove.
4604 * configure: Regenerate.
4605 * config.in: Regenerate.
4606 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
4608 2014-04-25 Richard Biener <rguenther@suse.de>
4611 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
4612 call stmt use/clobber sets during stmt walk instead of
4613 walking the possibly incomplete set of caller edges.
4615 2014-04-25 Richard Biener <rguenther@suse.de>
4618 * passes.c (apply_ipa_transforms): Inline into only caller ...
4619 (execute_one_pass): ... here. Properly bring in function
4620 bodies for nodes we want to apply IPA transforms to.
4622 2014-04-24 Cong Hou <congh@google.com>
4624 PR tree-optimization/60896
4625 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
4626 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
4627 (vect_mark_pattern_stmts): Set the def type of all statements in
4628 PATTERN_DEF_SEQ as vect_internal_def.
4630 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
4632 * doc/extend.texi (PowerPC Built-in Functions): Document new
4633 powerpc extended divide, bcd, pack/unpack 128-bit, builtin functions.
4634 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
4636 * config/rs6000/predicates.md (const_0_to_3_operand): New
4637 predicate to match 0..3 integer constants.
4639 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
4640 to support adding miscellaneous builtin functions.
4641 (BU_DFP_MISC_2): Likewise.
4642 (BU_P7_MISC_1): Likewise.
4643 (BU_P7_MISC_2): Likewise.
4644 (BU_P8V_MISC_3): Likewise.
4645 (BU_MISC_1): Likewise.
4646 (BU_MISC_2): Likewise.
4647 (DIVWE): Add extended divide builtin functions.
4650 (DIVWEUO): Likewise.
4654 (DIVDEUO): Likewise.
4655 (DXEX): Add decimal floating-point builtin functions.
4658 (DDEDPDQ): Likewise.
4660 (DENBCDQ): Likewise.
4667 (CDTBCD): Add new BCD builtin functions.
4671 (BCDADD_LT): Likewise.
4672 (BCDADD_EQ): Likewise.
4673 (BCDADD_GT): Likewise.
4674 (BCDADD_OV): Likewise.
4676 (BCDSUB_LT): Likewise.
4677 (BCDSUB_EQ): Likewise.
4678 (BCDSUB_GT): Likewise.
4679 (BCDSUB_OV): Likewise.
4680 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
4681 (UNPACK_TD): Likewise.
4682 (PACK_TF): Likewise.
4683 (UNPACK_TF): Likewise.
4684 (UNPACK_TF_0): Likewise.
4685 (UNPACK_TF_1): Likewise.
4686 (PACK_V1TI): Likewise.
4687 (UNPACK_V1TI): Likewise.
4689 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
4690 support for decimal floating point builtin functions.
4691 (rs6000_expand_ternop_builtin): Add checks for the new builtin
4692 functions that take constant arguments.
4693 (rs6000_invalid_builtin): Add decimal floating point builtin support.
4694 (rs6000_init_builtins): Setup long double, _Decimal64, and
4695 _Decimal128 types for new builtin functions.
4696 (builtin_function_type): Set the unsigned flags appropriately for
4697 the new builtin functions.
4698 (rs6000_opt_masks): Add support for decimal floating point builtin
4701 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
4702 floating point builtin functions.
4703 (RS6000_BTM_COMMON): Likewise.
4704 (RS6000_BTI_long_double): Likewise.
4705 (RS6000_BTI_dfloat64): Likewise.
4706 (RS6000_BTI_dfloat128): Likewise.
4707 (long_double_type_internal_node): Likewise.
4708 (dfloat64_type_internal_node): Likewise.
4709 (dfloat128_type_internal_node): Likewise.
4711 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
4712 2.07 bcd arithmetic instructions.
4713 (UNSPEC_BCDSUB): Likewise.
4714 (UNSPEC_BCD_OVERFLOW): Likewise.
4715 (UNSPEC_BCD_ADD_SUB): Likewise.
4716 (bcd_add_sub): Likewise.
4717 (BCD_TEST): Likewise.
4718 (bcd<bcd_add_sub>): Likewise.
4719 (bcd<bcd_add_sub>_test): Likewise.
4720 (bcd<bcd_add_sub>_test2): Likewise.
4721 (bcd<bcd_add_sub>_<code>): Likewise.
4722 (peephole2 for combined bcd ops): Likewise.
4724 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
4725 decimal floating point builtin functions.
4726 (UNSPEC_DENBCD): Likewise.
4727 (UNSPEC_DXEX): Likewise.
4728 (UNSPEC_DIEX): Likewise.
4729 (UNSPEC_DSCLI): Likewise.
4730 (UNSPEC_DSCRI): Likewise.
4731 (D64_D128): Likewise.
4732 (dfp_suffix): Likewise.
4733 (dfp_ddedpd_<mode>): Likewise.
4734 (dfp_denbcd_<mode>): Likewise.
4735 (dfp_dxex_<mode>): Likewise.
4736 (dfp_diex_<mode>): Likewise.
4737 (dfp_dscli_<mode>): Likewise.
4738 (dfp_dscri_<mode>): Likewise.
4740 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
4742 (UNSPEC_CDTBCD): Likewise.
4743 (UNSPEC_CBCDTD): Likewise.
4744 (UNSPEC_DIVE): Add support for new extended divide builtin functions.
4745 (UNSPEC_DIVEO): Likewise.
4746 (UNSPEC_DIVEU): Likewise.
4747 (UNSPEC_DIVEUO): Likewise.
4748 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
4749 pack/unpack 128-bit types.
4750 (UNSPEC_PACK_128BIT): Likewise.
4751 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
4752 (udiv<mode>3): Use idiv_ldiv mode attribute.
4753 (div<mode>3): Likewise.
4754 (addg6s): Add new BCD builtin functions.
4757 (UNSPEC_DIV_EXTEND): Add support for new extended divide instructions.
4758 (div_extend): Likewise.
4759 (div<div_extend>_<mode>"): Likewise.
4760 (FP128_64): Add support for new builtin functions to pack/unpack
4762 (unpack<mode>): Likewise.
4763 (unpacktf_0): Likewise.
4764 (unpacktf_1): Likewise.
4765 (unpack<mode>_dm): Likewise.
4766 (unpack<mode>_nodm): Likewise.
4767 (pack<mode>): Likewise.
4768 (unpackv1ti): Likewise.
4769 (packv1ti): Likewise.
4771 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
4773 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
4776 2014-04-24 Jakub Jelinek <jakub@redhat.com>
4778 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
4779 * gimplify.c (omp_is_private): Change last argument's type to int.
4780 Only diagnose lastprivate if the simd argument is 1, only diagnose
4781 linear if the simd argument is 2.
4782 (gimplify_omp_for): Adjust omp_is_private callers. When adding
4783 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
4784 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
4785 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
4786 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
4787 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4788 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
4789 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
4790 * tree-nested.c (convert_nonlocal_omp_clauses,
4791 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
4793 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
4796 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
4799 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
4801 * flag-types.h (enum ivar_visibility): Add.
4803 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
4805 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
4806 function * argument.
4808 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
4810 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
4812 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4813 Tom de Vries <tom@codesourcery.com>
4815 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
4816 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
4818 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
4819 * emit-rtl.c (try_split): Same.
4821 2014-04-24 Radovan Obradovic <robradovic@mips.com>
4822 Tom de Vries <tom@codesourcery.com>
4824 * common.opt (fuse-caller-save): New option.
4826 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
4828 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of
4829 elements for big-endian.
4831 2014-04-24 Richard Biener <rguenther@suse.de>
4833 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
4834 during TER and instead use the sepops interface for expanding
4835 non-GIMPLE_SINGLE_RHS.
4837 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4839 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
4840 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
4842 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4844 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
4845 assembler 64-bit option.
4846 * configure: Regenerate.
4848 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4850 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
4851 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
4852 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
4853 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
4854 (TARGET_CRYPTO): Take TARGET_SIMD into account.
4856 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4858 * config/aarch64/aarch64-builtins.c
4859 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
4860 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
4861 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
4862 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
4864 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
4865 (Vrevsuff): New mode attribute.
4867 2014-04-24 Terry Guo <terry.guo@arm.com>
4869 * config/arm/arm.h (machine_function): Define variable
4870 after_arm_reorg here.
4871 * config/arm/arm.c (after_arm_reorg): Remove the definition.
4872 (arm_split_constant): Update the way to access variable
4875 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
4877 2014-04-23 Tom de Vries <tom@codesourcery.com>
4879 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
4881 2014-04-23 David Malcolm <dmalcolm@redhat.com>
4883 * is-a.h: Update comments to reflect the following changes to the
4884 "pointerness" of the API, making the template parameter match the
4885 return type, allowing use of is-a.h with typedefs of pointers.
4886 (is_a_helper::cast): Return a T rather then a pointer to a T, so
4887 that the return type matches the parameter to the is_a_helper.
4889 (dyn_cast): Likewise.
4891 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
4892 pointer from the is-a.h API.
4894 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
4895 (is_a_helper <cgraph_node *>::test): ...this, matching change to
4897 (is_a_helper <varpool_node>::test): Likewise, convert to...
4898 (is_a_helper <varpool_node *>::test): ...this.
4900 (varpool_first_variable): Update for removal of implicit pointer
4901 from the is-a.h API.
4902 (varpool_next_variable): Likewise.
4903 (varpool_first_static_initializer): Likewise.
4904 (varpool_next_static_initializer): Likewise.
4905 (varpool_first_defined_variable): Likewise.
4906 (varpool_next_defined_variable): Likewise.
4907 (cgraph_first_defined_function): Likewise.
4908 (cgraph_next_defined_function): Likewise.
4909 (cgraph_first_function): Likewise.
4910 (cgraph_next_function): Likewise.
4911 (cgraph_first_function_with_gimple_body): Likewise.
4912 (cgraph_next_function_with_gimple_body): Likewise.
4913 (cgraph_alias_target): Likewise.
4914 (varpool_alias_target): Likewise.
4915 (cgraph_function_or_thunk_node): Likewise.
4916 (varpool_variable_node): Likewise.
4917 (symtab_real_symbol_p): Likewise.
4918 * cgraphunit.c (referred_to_p): Likewise.
4919 (analyze_functions): Likewise.
4920 (handle_alias_pairs): Likewise.
4921 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
4922 * gimple-ssa.h (gimple_vuse_op): Likewise.
4923 (gimple_vdef_op): Likewise.
4924 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
4925 * gimple.c (gimple_build_asm_1): Likewise.
4926 (gimple_build_try): Likewise.
4927 (gimple_build_resx): Likewise.
4928 (gimple_build_eh_dispatch): Likewise.
4929 (gimple_build_omp_for): Likewise.
4930 (gimple_omp_for_set_clauses): Likewise.
4932 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
4933 (is_a_helper <gimple_statement_asm *>::test): ...this.
4934 (is_a_helper <gimple_statement_bind>::test): Convert to...
4935 (is_a_helper <gimple_statement_bind *>::test): ...this.
4936 (is_a_helper <gimple_statement_call>::test): Convert to...
4937 (is_a_helper <gimple_statement_call *>::test): ...this.
4938 (is_a_helper <gimple_statement_catch>::test): Convert to...
4939 (is_a_helper <gimple_statement_catch *>::test): ...this.
4940 (is_a_helper <gimple_statement_resx>::test): Convert to...
4941 (is_a_helper <gimple_statement_resx *>::test): ...this.
4942 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
4943 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
4944 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
4945 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
4946 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
4947 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
4948 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
4949 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
4950 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
4951 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
4952 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
4953 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
4954 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
4955 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
4956 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
4957 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
4958 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
4959 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
4960 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
4961 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
4962 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
4963 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
4964 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
4965 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
4966 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
4967 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
4968 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
4969 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
4970 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
4971 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
4972 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
4973 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
4974 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
4975 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
4976 (is_a_helper <gimple_statement_phi>::test): Convert to...
4977 (is_a_helper <gimple_statement_phi *>::test): ...this.
4978 (is_a_helper <gimple_statement_transaction>::test): Convert to...
4979 (is_a_helper <gimple_statement_transaction *>::test): ...this.
4980 (is_a_helper <gimple_statement_try>::test): Convert to...
4981 (is_a_helper <gimple_statement_try *>::test): ...this.
4982 (is_a_helper <gimple_statement_wce>::test): Convert to...
4983 (is_a_helper <gimple_statement_wce *>::test): ...this.
4984 (is_a_helper <const gimple_statement_asm>::test): Convert to...
4985 (is_a_helper <const gimple_statement_asm *>::test): ...this.
4986 (is_a_helper <const gimple_statement_bind>::test): Convert to...
4987 (is_a_helper <const gimple_statement_bind *>::test): ...this.
4988 (is_a_helper <const gimple_statement_call>::test): Convert to...
4989 (is_a_helper <const gimple_statement_call *>::test): ...this.
4990 (is_a_helper <const gimple_statement_catch>::test): Convert to...
4991 (is_a_helper <const gimple_statement_catch *>::test): ...this.
4992 (is_a_helper <const gimple_statement_resx>::test): Convert to...
4993 (is_a_helper <const gimple_statement_resx *>::test): ...this.
4994 (is_a_helper <const gimple_statement_eh_dispatch>::test): Convert to...
4995 (is_a_helper <const gimple_statement_eh_dispatch *>::test): ...this.
4996 (is_a_helper <const gimple_statement_eh_filter>::test): Convert to...
4997 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
4998 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
5000 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
5002 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
5004 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
5006 (is_a_helper <const gimple_statement_omp_return>::test): Convert to...
5007 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
5008 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
5010 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
5011 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
5013 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
5014 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
5015 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
5016 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert to...
5017 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
5018 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
5020 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
5021 (is_a_helper <const gimple_statement_omp_target>::test): Convert to...
5022 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
5023 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
5025 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
5026 (is_a_helper <const gimple_statement_omp_single>::test): Convert to...
5027 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
5028 (is_a_helper <const gimple_statement_omp_teams>::test): Convert to...
5029 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
5030 (is_a_helper <const gimple_statement_omp_task>::test): Convert to...
5031 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
5032 (is_a_helper <const gimple_statement_phi>::test): Convert to...
5033 (is_a_helper <const gimple_statement_phi *>::test): ...this.
5034 (is_a_helper <const gimple_statement_transaction>::test): Convert to...
5035 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
5036 (is_a_helper <const gimple_statement_with_ops>::test): Convert to...
5037 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
5038 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
5039 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
5040 (is_a_helper <const gimple_statement_with_memory_ops>::test): Convert
5042 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
5044 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert to...
5045 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
5047 (gimple_use_ops): Update for removal of implicit pointer from the
5049 (gimple_set_use_ops): Likewise.
5050 (gimple_vuse): Likewise.
5051 (gimple_vdef): Likewise.
5052 (gimple_vuse_ptr): Likewise.
5053 (gimple_vdef_ptr): Likewise.
5054 (gimple_set_vuse): Likewise.
5055 (gimple_set_vdef): Likewise.
5056 (gimple_omp_return_set_lhs): Likewise.
5057 (gimple_omp_return_lhs): Likewise.
5058 (gimple_omp_return_lhs_ptr): Likewise.
5059 (gimple_call_fntype): Likewise.
5060 (gimple_call_set_fntype): Likewise.
5061 (gimple_call_set_internal_fn): Likewise.
5062 (gimple_call_use_set): Likewise.
5063 (gimple_call_clobber_set): Likewise.
5064 (gimple_bind_vars): Likewise.
5065 (gimple_bind_set_vars): Likewise.
5066 (gimple_bind_body_ptr): Likewise.
5067 (gimple_bind_set_body): Likewise.
5068 (gimple_bind_add_stmt): Likewise.
5069 (gimple_bind_block): Likewise.
5070 (gimple_bind_set_block): Likewise.
5071 (gimple_asm_ninputs): Likewise.
5072 (gimple_asm_noutputs): Likewise.
5073 (gimple_asm_nclobbers): Likewise.
5074 (gimple_asm_nlabels): Likewise.
5075 (gimple_asm_input_op): Likewise.
5076 (gimple_asm_input_op_ptr): Likewise.
5077 (gimple_asm_output_op): Likewise.
5078 (gimple_asm_output_op_ptr): Likewise.
5079 (gimple_asm_set_output_op): Likewise.
5080 (gimple_asm_clobber_op): Likewise.
5081 (gimple_asm_set_clobber_op): Likewise.
5082 (gimple_asm_label_op): Likewise.
5083 (gimple_asm_set_label_op): Likewise.
5084 (gimple_asm_string): Likewise.
5085 (gimple_catch_types): Likewise.
5086 (gimple_catch_types_ptr): Likewise.
5087 (gimple_catch_handler_ptr): Likewise.
5088 (gimple_catch_set_types): Likewise.
5089 (gimple_catch_set_handler): Likewise.
5090 (gimple_eh_filter_types): Likewise.
5091 (gimple_eh_filter_types_ptr): Likewise.
5092 (gimple_eh_filter_failure_ptr): Likewise.
5093 (gimple_eh_filter_set_types): Likewise.
5094 (gimple_eh_filter_set_failure): Likewise.
5095 (gimple_eh_must_not_throw_fndecl): Likewise.
5096 (gimple_eh_must_not_throw_set_fndecl): Likewise.
5097 (gimple_eh_else_n_body_ptr): Likewise.
5098 (gimple_eh_else_e_body_ptr): Likewise.
5099 (gimple_eh_else_set_n_body): Likewise.
5100 (gimple_eh_else_set_e_body): Likewise.
5101 (gimple_try_eval_ptr): Likewise.
5102 (gimple_try_cleanup_ptr): Likewise.
5103 (gimple_try_set_eval): Likewise.
5104 (gimple_try_set_cleanup): Likewise.
5105 (gimple_wce_cleanup_ptr): Likewise.
5106 (gimple_wce_set_cleanup): Likewise.
5107 (gimple_phi_capacity): Likewise.
5108 (gimple_phi_num_args): Likewise.
5109 (gimple_phi_result): Likewise.
5110 (gimple_phi_result_ptr): Likewise.
5111 (gimple_phi_set_result): Likewise.
5112 (gimple_phi_arg): Likewise.
5113 (gimple_phi_set_arg): Likewise.
5114 (gimple_resx_region): Likewise.
5115 (gimple_resx_set_region): Likewise.
5116 (gimple_eh_dispatch_region): Likewise.
5117 (gimple_eh_dispatch_set_region): Likewise.
5118 (gimple_omp_critical_name): Likewise.
5119 (gimple_omp_critical_name_ptr): Likewise.
5120 (gimple_omp_critical_set_name): Likewise.
5121 (gimple_omp_for_clauses): Likewise.
5122 (gimple_omp_for_clauses_ptr): Likewise.
5123 (gimple_omp_for_set_clauses): Likewise.
5124 (gimple_omp_for_collapse): Likewise.
5125 (gimple_omp_for_index): Likewise.
5126 (gimple_omp_for_index_ptr): Likewise.
5127 (gimple_omp_for_set_index): Likewise.
5128 (gimple_omp_for_initial): Likewise.
5129 (gimple_omp_for_initial_ptr): Likewise.
5130 (gimple_omp_for_set_initial): Likewise.
5131 (gimple_omp_for_final): Likewise.
5132 (gimple_omp_for_final_ptr): Likewise.
5133 (gimple_omp_for_set_final): Likewise.
5134 (gimple_omp_for_incr): Likewise.
5135 (gimple_omp_for_incr_ptr): Likewise.
5136 (gimple_omp_for_set_incr): Likewise.
5137 (gimple_omp_for_pre_body_ptr): Likewise.
5138 (gimple_omp_for_set_pre_body): Likewise.
5139 (gimple_omp_parallel_clauses): Likewise.
5140 (gimple_omp_parallel_clauses_ptr): Likewise.
5141 (gimple_omp_parallel_set_clauses): Likewise.
5142 (gimple_omp_parallel_child_fn): Likewise.
5143 (gimple_omp_parallel_child_fn_ptr): Likewise.
5144 (gimple_omp_parallel_set_child_fn): Likewise.
5145 (gimple_omp_parallel_data_arg): Likewise.
5146 (gimple_omp_parallel_data_arg_ptr): Likewise.
5147 (gimple_omp_parallel_set_data_arg): Likewise.
5148 (gimple_omp_task_clauses): Likewise.
5149 (gimple_omp_task_clauses_ptr): Likewise.
5150 (gimple_omp_task_set_clauses): Likewise.
5151 (gimple_omp_task_child_fn): Likewise.
5152 (gimple_omp_task_child_fn_ptr): Likewise.
5153 (gimple_omp_task_set_child_fn): Likewise.
5154 (gimple_omp_task_data_arg): Likewise.
5155 (gimple_omp_task_data_arg_ptr): Likewise.
5156 (gimple_omp_task_set_data_arg): Likewise.
5157 (gimple_omp_taskreg_clauses): Likewise.
5158 (gimple_omp_taskreg_clauses_ptr): Likewise.
5159 (gimple_omp_taskreg_set_clauses): Likewise.
5160 (gimple_omp_taskreg_child_fn): Likewise.
5161 (gimple_omp_taskreg_child_fn_ptr): Likewise.
5162 (gimple_omp_taskreg_set_child_fn): Likewise.
5163 (gimple_omp_taskreg_data_arg): Likewise.
5164 (gimple_omp_taskreg_data_arg_ptr): Likewise.
5165 (gimple_omp_taskreg_set_data_arg): Likewise.
5166 (gimple_omp_task_copy_fn): Likewise.
5167 (gimple_omp_task_copy_fn_ptr): Likewise.
5168 (gimple_omp_task_set_copy_fn): Likewise.
5169 (gimple_omp_task_arg_size): Likewise.
5170 (gimple_omp_task_arg_size_ptr): Likewise.
5171 (gimple_omp_task_set_arg_size): Likewise.
5172 (gimple_omp_task_arg_align): Likewise.
5173 (gimple_omp_task_arg_align_ptr): Likewise.
5174 (gimple_omp_task_set_arg_align): Likewise.
5175 (gimple_omp_single_clauses): Likewise.
5176 (gimple_omp_single_clauses_ptr): Likewise.
5177 (gimple_omp_single_set_clauses): Likewise.
5178 (gimple_omp_target_clauses): Likewise.
5179 (gimple_omp_target_clauses_ptr): Likewise.
5180 (gimple_omp_target_set_clauses): Likewise.
5181 (gimple_omp_target_child_fn): Likewise.
5182 (gimple_omp_target_child_fn_ptr): Likewise.
5183 (gimple_omp_target_set_child_fn): Likewise.
5184 (gimple_omp_target_data_arg): Likewise.
5185 (gimple_omp_target_data_arg_ptr): Likewise.
5186 (gimple_omp_target_set_data_arg): Likewise.
5187 (gimple_omp_teams_clauses): Likewise.
5188 (gimple_omp_teams_clauses_ptr): Likewise.
5189 (gimple_omp_teams_set_clauses): Likewise.
5190 (gimple_omp_sections_clauses): Likewise.
5191 (gimple_omp_sections_clauses_ptr): Likewise.
5192 (gimple_omp_sections_set_clauses): Likewise.
5193 (gimple_omp_sections_control): Likewise.
5194 (gimple_omp_sections_control_ptr): Likewise.
5195 (gimple_omp_sections_set_control): Likewise.
5196 (gimple_omp_for_set_cond): Likewise.
5197 (gimple_omp_for_cond): Likewise.
5198 (gimple_omp_atomic_store_set_val): Likewise.
5199 (gimple_omp_atomic_store_val): Likewise.
5200 (gimple_omp_atomic_store_val_ptr): Likewise.
5201 (gimple_omp_atomic_load_set_lhs): Likewise.
5202 (gimple_omp_atomic_load_lhs): Likewise.
5203 (gimple_omp_atomic_load_lhs_ptr): Likewise.
5204 (gimple_omp_atomic_load_set_rhs): Likewise.
5205 (gimple_omp_atomic_load_rhs): Likewise.
5206 (gimple_omp_atomic_load_rhs_ptr): Likewise.
5207 (gimple_omp_continue_control_def): Likewise.
5208 (gimple_omp_continue_control_def_ptr): Likewise.
5209 (gimple_omp_continue_set_control_def): Likewise.
5210 (gimple_omp_continue_control_use): Likewise.
5211 (gimple_omp_continue_control_use_ptr): Likewise.
5212 (gimple_omp_continue_set_control_use): Likewise.
5213 (gimple_transaction_body_ptr): Likewise.
5214 (gimple_transaction_label): Likewise.
5215 (gimple_transaction_label_ptr): Likewise.
5216 (gimple_transaction_set_body): Likewise.
5217 (gimple_transaction_set_label): Likewise.
5219 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
5220 * ipa-inline-analysis.c (inline_write_summary): Likewise.
5221 * ipa-ref.c (ipa_record_reference): Likewise.
5222 * ipa-reference.c (analyze_function): Likewise.
5223 (ipa_reference_write_optimization_summary): Likewise.
5224 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
5225 (address_taken_from_non_vtable_p): Likewise.
5226 (comdat_can_be_unshared_p_1): Likewise.
5227 * lto-cgraph.c (lto_output_ref): Likewise.
5228 (add_references): Likewise.
5229 (compute_ltrans_boundary): Likewise.
5230 (output_symtab): Likewise.
5231 (input_ref): Likewise.
5232 (input_cgraph_1): Likewise.
5233 (output_cgraph_opt_summary): Likewise.
5234 * lto-streamer-out.c (lto_output): Likewise.
5235 (output_symbol_p): Likewise.
5236 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
5237 (lsei_start_function_in_partition): Likewise.
5238 (lsei_next_variable_in_partition): Likewise.
5239 (lsei_start_variable_in_partition): Likewise.
5240 * symtab.c (insert_to_assembler_name_hash): Likewise.
5241 (unlink_from_assembler_name_hash): Likewise.
5242 (symtab_unregister_node): Likewise.
5243 (symtab_remove_node): Likewise.
5244 (dump_symtab_node): Likewise.
5245 (verify_symtab_base): Likewise.
5246 (verify_symtab_node): Likewise.
5247 (symtab_make_decl_local): Likewise.
5248 (symtab_alias_ultimate_target): Likewise.
5249 (symtab_resolve_alias): Likewise.
5250 (symtab_get_symbol_partitioning_class): Likewise.
5251 * tree-phinodes.c (allocate_phi_node): Likewise.
5252 (reserve_phi_args_for_new_edge): Likewise.
5253 (remove_phi_args): Likewise.
5254 * varpool.c (varpool_node_for_asm): Likewise.
5255 (varpool_remove_unreferenced_decls): Likewise.
5257 2014-04-23 Jeff Law <law@redhat.com>
5259 PR tree-optimization/60902
5260 * tree-ssa-threadedge.c
5261 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
5262 invalidate outputs from statements that do not produce useful
5263 outputs for threading.
5265 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
5267 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
5268 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
5269 machine descriptions for Stack Smashing Protector.
5271 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
5273 * aarch64.md (<optab>_rol<mode>3): New pattern.
5274 (<optab>_rolsi3_uxtw): Likewise.
5275 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
5277 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
5279 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
5280 (arm_cortex_a12_tune): Likewise.
5282 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5284 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
5286 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5288 * config/arm/arm.md (arm_rev16si2): New pattern.
5289 (arm_rev16si2_alt): Likewise.
5290 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
5292 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5294 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
5295 (rev16<mode>2_alt): Likewise.
5296 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
5297 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
5298 (aarch_rev16_shleft_mask_imm_p): Likewise.
5299 (aarch_rev16_p_1): Likewise.
5300 (aarch_rev16_p): Likewise.
5301 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
5302 (aarch_rev16_shright_mask_imm_p): Likewise.
5303 (aarch_rev16_shleft_mask_imm_p): Likewise.
5305 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5307 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
5308 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
5310 (cortex_a53_extra_costs): Likewise.
5311 (cortex_a57_extra_costs): Likewise.
5312 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
5313 (cortexa7_extra_costs): Likewise.
5314 (cortexa8_extra_costs): Likewise.
5315 (cortexa12_extra_costs): Likewise.
5316 (cortexa15_extra_costs): Likewise.
5317 (v7m_extra_costs): Likewise.
5318 (arm_new_rtx_costs): Handle BSWAP.
5320 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5322 * config/arm/arm.c (cortexa8_extra_costs): New table.
5323 (arm_cortex_a8_tune): New tuning struct.
5324 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
5326 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5328 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
5330 2014-04-23 Richard Biener <rguenther@suse.de>
5332 * Makefile.in (OBJS): Remove loop-unswitch.o.
5333 * tree-pass.h (make_pass_rtl_unswitch): Remove.
5334 * passes.def (pass_rtl_unswitch): Likewise.
5335 * loop-init.c (gate_rtl_unswitch): Likewise.
5336 (rtl_unswitch): Likewise.
5337 (pass_data_rtl_unswitch): Likewise.
5338 (pass_rtl_unswitch): Likewise.
5339 (make_pass_rtl_unswitch): Likewise.
5340 * rtl.h (reversed_condition): Likewise.
5341 (compare_and_jump_seq): Likewise.
5342 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
5344 * loop-unroll.c (compare_and_jump_seq): Likewise.
5346 2014-04-23 Richard Biener <rguenther@suse.de>
5348 PR tree-optimization/60903
5349 * tree-ssa-loop-im.c (analyze_memory_references): Remove
5350 commented code block.
5351 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
5352 loop flags to newly created BBs and edges.
5354 2014-04-23 Nick Clifton <nickc@redhat.com>
5356 * config/msp430/msp430.c (msp430_handle_option): Move function
5358 (msp430_option_override): Simplify mcu and mcpu option handling.
5359 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
5360 support for -mhwmult command line option.
5361 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
5362 -mhwmult command line option.
5363 (msp430_hwmult_enabled): Delete.
5364 (msp43o_output_labelref): Add support for -mhwmult command line option.
5365 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
5366 (umulsidi3): Likewise.
5367 * config/msp430/msp430.opt (mmcu): Add Report attribute.
5368 (mcpu, mlarge, msmall): Likewise.
5369 (mhwmult): New option.
5370 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
5372 (msp430_is_f5_mcu): Remove prototype.
5373 (msp430_use_f5_series_hwmult): Add prototype.
5374 * config/msp430/msp430-opts.h: New file.
5375 * common/config/msp430: New directory.
5376 * common/config/msp430/msp430-common.c: New file.
5377 * config.gcc (msp430): Remove target_has_targetm_common.
5378 * doc/invoke.texi: Document -mhwmult command line option.
5380 2014-04-23 Nick Clifton <nickc@redhat.com>
5382 * config/i386/cygwin.h (ENDFILE_SPEC): Include
5383 default-manifest.o if it can be found in the search path.
5384 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
5386 2014-04-23 Terry Guo <terry.guo@arm.com>
5388 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
5390 2014-04-23 Richard Biener <rguenther@suse.de>
5393 * tree-inline.c (declare_return_variable): Use mark_addressable.
5395 2014-04-23 Richard Biener <rguenther@suse.de>
5398 * loop-init.c (loop_optimizer_init): Make sure to apply
5399 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
5401 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5404 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
5406 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
5407 if flag_sanitize_undefined_trap_on_error.
5408 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
5409 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
5410 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
5411 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
5412 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
5413 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
5414 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
5415 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
5416 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
5417 * ubsan.c (ubsan_instrument_unreachable): Return
5418 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
5419 (ubsan_expand_null_ifn): Emit __builtin_trap ()
5420 if flag_sanitize_undefined_trap_on_error and
5421 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
5422 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
5423 instrument_bool_enum_load): Emit __builtin_trap () if
5424 flag_sanitize_undefined_trap_on_error and
5425 __builtin_handle_*_abort () if !flag_sanitize_recover.
5426 * doc/invoke.texi (-fsanitize-recover,
5427 -fsanitize-undefined-trap-on-error): Document.
5429 2014-04-22 Christian Bruel <christian.bruel@st.com>
5431 * config/sh/sh.md (mov<mode>): Replace movQIHI.
5432 Force immediates to SImode.
5434 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
5436 * config/nios2/nios2.md (UNSPEC_ROUND): New.
5438 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
5439 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
5440 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
5441 (nios2_fpu_insn): Add entry for round.
5442 (N2FPU_NO_ERRNO_P): Define.
5443 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
5445 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
5447 2014-04-22 Richard Henderson <rth@redhat.com>
5449 * config/aarch64/aarch64 (addti3, subti3): New expanders.
5450 (add<GPI>3_compare0): Remove leading * from name.
5451 (add<GPI>3_carryin): Likewise.
5452 (sub<GPI>3_compare0): Likewise.
5453 (sub<GPI>3_carryin): Likewise.
5454 (<su_optab>mulditi3): New expander.
5455 (multi3): New expander.
5456 (madd<GPI>): Remove leading * from name.
5458 2014-04-22 Martin Jambor <mjambor@suse.cz>
5460 * cgraphclones.c (cgraph_function_versioning): Copy
5461 ipa_transforms_to_apply instead of asserting it is empty.
5463 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
5466 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
5467 on count_exp to get mode.
5469 2014-04-22 Andrew Pinski <apinski@cavium.com>
5471 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
5472 Handle TLS for ILP32.
5473 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
5474 (tlsie_small_<mode>): this and handle PTR.
5475 (tlsie_small_sidi): New pattern.
5476 (tlsle_small): Change to an expand to handle ILP32.
5477 (tlsle_small_<mode>): New pattern.
5478 (tlsdesc_small): Rename to ...
5479 (tlsdesc_small_<mode>): this and handle PTR.
5481 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5483 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
5485 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5487 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5488 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
5489 (aarch64_types_signed_poly_qualifiers): Likewise.
5490 (aarch64_types_unsigned_signed_qualifiers): Likewise.
5491 (aarch64_types_poly_signed_qualifiers): Likewise.
5492 (TYPES_REINTERP_SS): Type macro added.
5493 (TYPES_REINTERP_SU): Likewise.
5494 (TYPES_REINTERP_SP): Likewise.
5495 (TYPES_REINTERP_US): Likewise.
5496 (TYPES_REINTERP_PS): Likewise.
5497 (aarch64_fold_builtin): New expression folding added.
5498 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
5499 Declarations removed.
5500 (REINTERP_SS): Declarations added.
5501 (REINTERP_US): Likewise.
5502 (REINTERP_PS): Likewise.
5503 (REINTERP_SU): Likewise.
5504 (REINTERP_SP): Likewise.
5505 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
5506 (vreinterpretq_p8_f64): Likewise.
5507 (vreinterpret_p16_f64): Likewise.
5508 (vreinterpretq_p16_f64): Likewise.
5509 (vreinterpret_f32_f64): Likewise.
5510 (vreinterpretq_f32_f64): Likewise.
5511 (vreinterpret_f64_f32): Likewise.
5512 (vreinterpret_f64_p8): Likewise.
5513 (vreinterpret_f64_p16): Likewise.
5514 (vreinterpret_f64_s8): Likewise.
5515 (vreinterpret_f64_s16): Likewise.
5516 (vreinterpret_f64_s32): Likewise.
5517 (vreinterpret_f64_s64): Likewise.
5518 (vreinterpret_f64_u8): Likewise.
5519 (vreinterpret_f64_u16): Likewise.
5520 (vreinterpret_f64_u32): Likewise.
5521 (vreinterpret_f64_u64): Likewise.
5522 (vreinterpretq_f64_f32): Likewise.
5523 (vreinterpretq_f64_p8): Likewise.
5524 (vreinterpretq_f64_p16): Likewise.
5525 (vreinterpretq_f64_s8): Likewise.
5526 (vreinterpretq_f64_s16): Likewise.
5527 (vreinterpretq_f64_s32): Likewise.
5528 (vreinterpretq_f64_s64): Likewise.
5529 (vreinterpretq_f64_u8): Likewise.
5530 (vreinterpretq_f64_u16): Likewise.
5531 (vreinterpretq_f64_u32): Likewise.
5532 (vreinterpretq_f64_u64): Likewise.
5533 (vreinterpret_s64_f64): Likewise.
5534 (vreinterpretq_s64_f64): Likewise.
5535 (vreinterpret_u64_f64): Likewise.
5536 (vreinterpretq_u64_f64): Likewise.
5537 (vreinterpret_s8_f64): Likewise.
5538 (vreinterpretq_s8_f64): Likewise.
5539 (vreinterpret_s16_f64): Likewise.
5540 (vreinterpretq_s16_f64): Likewise.
5541 (vreinterpret_s32_f64): Likewise.
5542 (vreinterpretq_s32_f64): Likewise.
5543 (vreinterpret_u8_f64): Likewise.
5544 (vreinterpretq_u8_f64): Likewise.
5545 (vreinterpret_u16_f64): Likewise.
5546 (vreinterpretq_u16_f64): Likewise.
5547 (vreinterpret_u32_f64): Likewise.
5548 (vreinterpretq_u32_f64): Likewise.
5550 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5552 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
5553 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
5554 (vreinterpret_p8_s8): Likewise.
5555 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
5556 (vreinterpret_p8_s16): Likewise.
5557 (vreinterpret_p8_s32): Likewise.
5558 (vreinterpret_p8_s64): Likewise.
5559 (vreinterpret_p8_f32): Likewise.
5560 (vreinterpret_p8_u8): Likewise.
5561 (vreinterpret_p8_u16): Likewise.
5562 (vreinterpret_p8_u32): Likewise.
5563 (vreinterpret_p8_u64): Likewise.
5564 (vreinterpret_p8_p16): Likewise.
5565 (vreinterpretq_p8_s8): Likewise.
5566 (vreinterpretq_p8_s16): Likewise.
5567 (vreinterpretq_p8_s32): Likewise.
5568 (vreinterpretq_p8_s64): Likewise.
5569 (vreinterpretq_p8_f32): Likewise.
5570 (vreinterpretq_p8_u8): Likewise.
5571 (vreinterpretq_p8_u16): Likewise.
5572 (vreinterpretq_p8_u32): Likewise.
5573 (vreinterpretq_p8_u64): Likewise.
5574 (vreinterpretq_p8_p16): Likewise.
5575 (vreinterpret_p16_s8): Likewise.
5576 (vreinterpret_p16_s16): Likewise.
5577 (vreinterpret_p16_s32): Likewise.
5578 (vreinterpret_p16_s64): Likewise.
5579 (vreinterpret_p16_f32): Likewise.
5580 (vreinterpret_p16_u8): Likewise.
5581 (vreinterpret_p16_u16): Likewise.
5582 (vreinterpret_p16_u32): Likewise.
5583 (vreinterpret_p16_u64): Likewise.
5584 (vreinterpret_p16_p8): Likewise.
5585 (vreinterpretq_p16_s8): Likewise.
5586 (vreinterpretq_p16_s16): Likewise.
5587 (vreinterpretq_p16_s32): Likewise.
5588 (vreinterpretq_p16_s64): Likewise.
5589 (vreinterpretq_p16_f32): Likewise.
5590 (vreinterpretq_p16_u8): Likewise.
5591 (vreinterpretq_p16_u16): Likewise.
5592 (vreinterpretq_p16_u32): Likewise.
5593 (vreinterpretq_p16_u64): Likewise.
5594 (vreinterpretq_p16_p8): Likewise.
5595 (vreinterpret_f32_s8): Likewise.
5596 (vreinterpret_f32_s16): Likewise.
5597 (vreinterpret_f32_s32): Likewise.
5598 (vreinterpret_f32_s64): Likewise.
5599 (vreinterpret_f32_u8): Likewise.
5600 (vreinterpret_f32_u16): Likewise.
5601 (vreinterpret_f32_u32): Likewise.
5602 (vreinterpret_f32_u64): Likewise.
5603 (vreinterpret_f32_p8): Likewise.
5604 (vreinterpret_f32_p16): Likewise.
5605 (vreinterpretq_f32_s8): Likewise.
5606 (vreinterpretq_f32_s16): Likewise.
5607 (vreinterpretq_f32_s32): Likewise.
5608 (vreinterpretq_f32_s64): Likewise.
5609 (vreinterpretq_f32_u8): Likewise.
5610 (vreinterpretq_f32_u16): Likewise.
5611 (vreinterpretq_f32_u32): Likewise.
5612 (vreinterpretq_f32_u64): Likewise.
5613 (vreinterpretq_f32_p8): Likewise.
5614 (vreinterpretq_f32_p16): Likewise.
5615 (vreinterpret_s64_s8): Likewise.
5616 (vreinterpret_s64_s16): Likewise.
5617 (vreinterpret_s64_s32): Likewise.
5618 (vreinterpret_s64_f32): Likewise.
5619 (vreinterpret_s64_u8): Likewise.
5620 (vreinterpret_s64_u16): Likewise.
5621 (vreinterpret_s64_u32): Likewise.
5622 (vreinterpret_s64_u64): Likewise.
5623 (vreinterpret_s64_p8): Likewise.
5624 (vreinterpret_s64_p16): Likewise.
5625 (vreinterpretq_s64_s8): Likewise.
5626 (vreinterpretq_s64_s16): Likewise.
5627 (vreinterpretq_s64_s32): Likewise.
5628 (vreinterpretq_s64_f32): Likewise.
5629 (vreinterpretq_s64_u8): Likewise.
5630 (vreinterpretq_s64_u16): Likewise.
5631 (vreinterpretq_s64_u32): Likewise.
5632 (vreinterpretq_s64_u64): Likewise.
5633 (vreinterpretq_s64_p8): Likewise.
5634 (vreinterpretq_s64_p16): Likewise.
5635 (vreinterpret_u64_s8): Likewise.
5636 (vreinterpret_u64_s16): Likewise.
5637 (vreinterpret_u64_s32): Likewise.
5638 (vreinterpret_u64_s64): Likewise.
5639 (vreinterpret_u64_f32): Likewise.
5640 (vreinterpret_u64_u8): Likewise.
5641 (vreinterpret_u64_u16): Likewise.
5642 (vreinterpret_u64_u32): Likewise.
5643 (vreinterpret_u64_p8): Likewise.
5644 (vreinterpret_u64_p16): Likewise.
5645 (vreinterpretq_u64_s8): Likewise.
5646 (vreinterpretq_u64_s16): Likewise.
5647 (vreinterpretq_u64_s32): Likewise.
5648 (vreinterpretq_u64_s64): Likewise.
5649 (vreinterpretq_u64_f32): Likewise.
5650 (vreinterpretq_u64_u8): Likewise.
5651 (vreinterpretq_u64_u16): Likewise.
5652 (vreinterpretq_u64_u32): Likewise.
5653 (vreinterpretq_u64_p8): Likewise.
5654 (vreinterpretq_u64_p16): Likewise.
5655 (vreinterpret_s8_s16): Likewise.
5656 (vreinterpret_s8_s32): Likewise.
5657 (vreinterpret_s8_s64): Likewise.
5658 (vreinterpret_s8_f32): Likewise.
5659 (vreinterpret_s8_u8): Likewise.
5660 (vreinterpret_s8_u16): Likewise.
5661 (vreinterpret_s8_u32): Likewise.
5662 (vreinterpret_s8_u64): Likewise.
5663 (vreinterpret_s8_p8): Likewise.
5664 (vreinterpret_s8_p16): Likewise.
5665 (vreinterpretq_s8_s16): Likewise.
5666 (vreinterpretq_s8_s32): Likewise.
5667 (vreinterpretq_s8_s64): Likewise.
5668 (vreinterpretq_s8_f32): Likewise.
5669 (vreinterpretq_s8_u8): Likewise.
5670 (vreinterpretq_s8_u16): Likewise.
5671 (vreinterpretq_s8_u32): Likewise.
5672 (vreinterpretq_s8_u64): Likewise.
5673 (vreinterpretq_s8_p8): Likewise.
5674 (vreinterpretq_s8_p16): Likewise.
5675 (vreinterpret_s16_s8): Likewise.
5676 (vreinterpret_s16_s32): Likewise.
5677 (vreinterpret_s16_s64): Likewise.
5678 (vreinterpret_s16_f32): Likewise.
5679 (vreinterpret_s16_u8): Likewise.
5680 (vreinterpret_s16_u16): Likewise.
5681 (vreinterpret_s16_u32): Likewise.
5682 (vreinterpret_s16_u64): Likewise.
5683 (vreinterpret_s16_p8): Likewise.
5684 (vreinterpret_s16_p16): Likewise.
5685 (vreinterpretq_s16_s8): Likewise.
5686 (vreinterpretq_s16_s32): Likewise.
5687 (vreinterpretq_s16_s64): Likewise.
5688 (vreinterpretq_s16_f32): Likewise.
5689 (vreinterpretq_s16_u8): Likewise.
5690 (vreinterpretq_s16_u16): Likewise.
5691 (vreinterpretq_s16_u32): Likewise.
5692 (vreinterpretq_s16_u64): Likewise.
5693 (vreinterpretq_s16_p8): Likewise.
5694 (vreinterpretq_s16_p16): Likewise.
5695 (vreinterpret_s32_s8): Likewise.
5696 (vreinterpret_s32_s16): Likewise.
5697 (vreinterpret_s32_s64): Likewise.
5698 (vreinterpret_s32_f32): Likewise.
5699 (vreinterpret_s32_u8): Likewise.
5700 (vreinterpret_s32_u16): Likewise.
5701 (vreinterpret_s32_u32): Likewise.
5702 (vreinterpret_s32_u64): Likewise.
5703 (vreinterpret_s32_p8): Likewise.
5704 (vreinterpret_s32_p16): Likewise.
5705 (vreinterpretq_s32_s8): Likewise.
5706 (vreinterpretq_s32_s16): Likewise.
5707 (vreinterpretq_s32_s64): Likewise.
5708 (vreinterpretq_s32_f32): Likewise.
5709 (vreinterpretq_s32_u8): Likewise.
5710 (vreinterpretq_s32_u16): Likewise.
5711 (vreinterpretq_s32_u32): Likewise.
5712 (vreinterpretq_s32_u64): Likewise.
5713 (vreinterpretq_s32_p8): Likewise.
5714 (vreinterpretq_s32_p16): Likewise.
5715 (vreinterpret_u8_s8): Likewise.
5716 (vreinterpret_u8_s16): Likewise.
5717 (vreinterpret_u8_s32): Likewise.
5718 (vreinterpret_u8_s64): Likewise.
5719 (vreinterpret_u8_f32): Likewise.
5720 (vreinterpret_u8_u16): Likewise.
5721 (vreinterpret_u8_u32): Likewise.
5722 (vreinterpret_u8_u64): Likewise.
5723 (vreinterpret_u8_p8): Likewise.
5724 (vreinterpret_u8_p16): Likewise.
5725 (vreinterpretq_u8_s8): Likewise.
5726 (vreinterpretq_u8_s16): Likewise.
5727 (vreinterpretq_u8_s32): Likewise.
5728 (vreinterpretq_u8_s64): Likewise.
5729 (vreinterpretq_u8_f32): Likewise.
5730 (vreinterpretq_u8_u16): Likewise.
5731 (vreinterpretq_u8_u32): Likewise.
5732 (vreinterpretq_u8_u64): Likewise.
5733 (vreinterpretq_u8_p8): Likewise.
5734 (vreinterpretq_u8_p16): Likewise.
5735 (vreinterpret_u16_s8): Likewise.
5736 (vreinterpret_u16_s16): Likewise.
5737 (vreinterpret_u16_s32): Likewise.
5738 (vreinterpret_u16_s64): Likewise.
5739 (vreinterpret_u16_f32): Likewise.
5740 (vreinterpret_u16_u8): Likewise.
5741 (vreinterpret_u16_u32): Likewise.
5742 (vreinterpret_u16_u64): Likewise.
5743 (vreinterpret_u16_p8): Likewise.
5744 (vreinterpret_u16_p16): Likewise.
5745 (vreinterpretq_u16_s8): Likewise.
5746 (vreinterpretq_u16_s16): Likewise.
5747 (vreinterpretq_u16_s32): Likewise.
5748 (vreinterpretq_u16_s64): Likewise.
5749 (vreinterpretq_u16_f32): Likewise.
5750 (vreinterpretq_u16_u8): Likewise.
5751 (vreinterpretq_u16_u32): Likewise.
5752 (vreinterpretq_u16_u64): Likewise.
5753 (vreinterpretq_u16_p8): Likewise.
5754 (vreinterpretq_u16_p16): Likewise.
5755 (vreinterpret_u32_s8): Likewise.
5756 (vreinterpret_u32_s16): Likewise.
5757 (vreinterpret_u32_s32): Likewise.
5758 (vreinterpret_u32_s64): Likewise.
5759 (vreinterpret_u32_f32): Likewise.
5760 (vreinterpret_u32_u8): Likewise.
5761 (vreinterpret_u32_u16): Likewise.
5762 (vreinterpret_u32_u64): Likewise.
5763 (vreinterpret_u32_p8): Likewise.
5764 (vreinterpret_u32_p16): Likewise.
5765 (vreinterpretq_u32_s8): Likewise.
5766 (vreinterpretq_u32_s16): Likewise.
5767 (vreinterpretq_u32_s32): Likewise.
5768 (vreinterpretq_u32_s64): Likewise.
5769 (vreinterpretq_u32_f32): Likewise.
5770 (vreinterpretq_u32_u8): Likewise.
5771 (vreinterpretq_u32_u16): Likewise.
5772 (vreinterpretq_u32_u64): Likewise.
5773 (vreinterpretq_u32_p8): Likewise.
5774 (vreinterpretq_u32_p16): Likewise.
5776 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5778 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
5780 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator extended.
5782 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
5783 (vqnegd_s64): Likewise.
5784 (vqabs_s64): Likewise.
5785 (vqabsd_s64): Likewise.
5787 2014-04-22 Richard Henderson <rth@redhat.com>
5789 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
5790 computation to the top of the loop.
5792 2014-04-22 Renlin <renlin.li@arm.com>
5793 Jiong Wang <jiong.wang@arm.com>
5795 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
5796 * config/aarch64/aarch64.c (aarch64_layout_frame)
5797 (aarch64_initial_elimination_offset): Likewise.
5799 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
5801 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
5804 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
5806 * machmode.h (bitwise_mode_for_mode): Declare.
5807 * stor-layout.h (bitwise_type_for_mode): Likewise.
5808 * stor-layout.c (bitwise_mode_for_mode): New function.
5809 (bitwise_type_for_mode): Likewise.
5810 * builtins.c (fold_builtin_memory_op): Use it instead of
5811 int_mode_for_mode and build_nonstandard_integer_type.
5813 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5815 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
5816 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
5817 (*-*-solaris2*): Simplify.
5818 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
5819 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
5820 *-*-solaris2.9* handling.
5822 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
5824 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
5825 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
5827 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
5828 * configure: Regenerate.
5830 * config/i386/sol2-9.h: Remove.
5832 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
5833 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
5834 Remove Solaris 9 references.
5836 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
5838 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
5839 (floatuns<GPI:mode><GPF:mode>2): Remove.
5840 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
5841 and floatuns conversions.
5842 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
5843 and floatuns conversions.
5844 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
5845 (w1,w2): New mode attributes for inequal width conversions.
5847 2014-04-22 Renlin Li <Renlin.Li@arm.com>
5849 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
5850 the output asm format.
5852 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
5854 * config/aarch64/aarch64-simd.md
5855 (aarch64_cm<optab>di): Always split.
5856 (*aarch64_cm<optab>di): New.
5857 (aarch64_cmtstdi): Always split.
5858 (*aarch64_cmtstdi): New.
5860 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5862 PR tree-optimization/60823
5863 * omp-low.c (ipa_simd_modify_function_body): Go through
5864 all SSA_NAMEs and for those refering to vector arguments
5865 which are going to be replaced adjust SSA_NAME_VAR and,
5866 if it is a default definition, change it into a non-default
5867 definition assigned at the beginning of function from new_decl.
5868 (ipa_simd_modify_stmt_ops): Rewritten.
5869 * tree-dfa.c (set_ssa_default_def): When removing default def,
5870 check for NULL loc instead of NULL *loc.
5872 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5874 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
5875 restrictions on core registers for DImode values in Thumb2.
5877 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5879 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
5880 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
5882 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5884 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
5885 (*iordi_notzesidi_di): Likewise.
5886 (*iordi_notsesidi_di): Likewise.
5888 2014-04-22 Ian Bolton <ian.bolton@arm.com>
5890 * config/arm/arm-protos.h (tune_params): New struct members.
5891 * config/arm/arm.c: Initialise tune_params per processor.
5892 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
5893 for speed, based on new tune_params.
5895 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
5897 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro added.
5898 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added macro.
5899 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment corrected.
5900 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
5901 * config/aarch64/arm_neon.h (vrnd_f64): Added.
5902 (vrnda_f64): Likewise.
5903 (vrndi_f64): Likewise.
5904 (vrndm_f64): Likewise.
5905 (vrndn_f64): Likewise.
5906 (vrndp_f64): Likewise.
5907 (vrndx_f64): Likewise.
5909 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5911 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
5912 GET_MODE_SIZE argument is enum machine_mode.
5914 2014-04-22 Jakub Jelinek <jakub@redhat.com>
5917 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
5918 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
5920 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
5923 * asan.c (asan_emit_stack_protection): Force the base to align to
5924 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
5925 appropriate bits if STRICT_ALIGNMENT.
5926 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
5928 (expand_used_vars): Leave a space in the stack frame for alignment
5929 if STRICT_ALIGNMENT.
5931 2014-04-21 David Malcolm <dmalcolm@redhat.com>
5933 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
5935 (gimple_store_p): Likewise.
5936 (gimple_assign_load_p): Likewise.
5937 (gimple_assign_cast_p): Likewise.
5938 (gimple_clobber_p): Likewise.
5940 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
5941 rather than a gimple.
5942 (gimple_assign_cast_p): Likewise.
5944 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
5947 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
5948 If mode is DDmode and TARGET_E500_DOUBLE allow move.
5950 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
5951 more debug information for E500 if -mdebug=reg.
5953 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
5956 * config/i386/i386.c (ix86_expand_builtin)
5957 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
5958 register for target RTX.
5959 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
5961 2014-04-18 Cong Hou <congh@google.com>
5963 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
5964 the widen-mult pattern by handling two operands with different sizes,
5965 and operands whose size is smaller than half of the result type.
5967 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5969 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
5970 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
5971 (do_estimate_edge_time): Compute it.
5972 * ipa-inline.c (want_inline_small_function_p): Bypass
5973 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
5975 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
5977 * ipa-inline.c (spec_rem): New static variable.
5978 (dump_overall_stats): New function.
5979 (dump_inline_stats): New function.
5981 2014-04-18 Richard Henderson <rth@redhat.com>
5983 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
5984 to GET_MODE_SIZE, not a reg_class_t.
5986 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5988 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
5989 (vsx_xxmrglw_<mode>): Likewise.
5991 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
5994 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
5995 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
5996 (rs6000_init_hard_regno_mode_ok): Likewise.
5998 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
6000 * ipa-inline.c (inline_small_functions): Account only non-cold
6002 * doc/invoke.texi (inline-unit-growth): Update documentation.
6004 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
6006 * config/rs6000/rs6000.md (addti3, subti3): New.
6008 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
6011 * config/i386/i386.c (ix86_expand_clear): Remove outdated
6012 comment. Check optimize_insn_for_size_p instead of
6013 optimize_insn_for_speed_p.
6015 2014-04-17 Martin Jambor <mjambor@suse.cz>
6017 * gimple-iterator.c (gsi_start_edge): New function.
6018 * gimple-iterator.h (gsi_start_edge): Declare.
6019 * tree-sra.c (single_non_eh_succ): New function.
6020 (disqualify_ops_if_throwing_stmt): Renamed to
6021 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
6022 having one non-EH successor BB.
6023 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
6024 generate loads into replacements.
6025 (sra_modify_assign): Likewise and and also use the simple path for
6027 (sra_modify_function_body): Commit statements on edges.
6029 2014-04-17 Richard Biener <rguenther@suse.de>
6032 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
6033 comparison results and add clarifying comment.
6035 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6037 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
6038 (blank_mode): Initialize it.
6039 (emit_mode_size_inline, emit_mode_nunits_inline,
6040 emit_mode_inner_inline): New functions.
6041 (emit_insn_modes_h): Call them and surround their output with
6042 #if GCC_VERSION >= 4001 ... #endif.
6043 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
6044 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
6045 mode_* arrays if the argument is __builtin_constant_p.
6046 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
6047 is enum machine_mode.
6049 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6051 * passes.c (opt_pass::execute): Adjust.
6052 (pass_manager::execute_pass_mode_switching): Likewise.
6053 (early_local_passes::execute): Likewise.
6054 (execute_one_pass): Pass cfun to the pass's execute method.
6055 * tree-pass.h (opt_pass::execute): Add function * argument.
6056 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6057 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6058 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6059 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6060 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6061 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
6062 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
6063 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6064 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6065 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
6066 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
6067 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
6068 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
6069 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
6070 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6071 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6072 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6073 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
6074 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
6075 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6076 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6077 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6078 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6079 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6080 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6081 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6082 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6083 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6084 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6087 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6089 * passes.c (opt_pass::gate): Take function * argument.
6090 (gate_all_early_local_passes): Merge into
6091 (early_local_passes::gate): this.
6092 (gate_all_early_optimizations): Merge into
6093 (all_early_optimizations::gate): this.
6094 (gate_all_optimizations): Mege into
6095 (all_optimizations::gate): this.
6096 (gate_all_optimizations_g): Merge into
6097 (all_optimizations_g::gate): this.
6098 (gate_rest_of_compilation): Mege into
6099 (rest_of_compilation::gate): this.
6100 (gate_postreload): Merge into
6101 (postreload::gate): this.
6102 (dump_one_pass): Pass cfun to the pass's gate method.
6103 (execute_ipa_summary_passes): Likewise.
6104 (execute_one_pass): Likewise.
6105 (ipa_write_summaries_2): Likewise.
6106 (ipa_write_optimization_summaries_1): Likewise.
6107 (ipa_read_summaries_1): Likewise.
6108 (ipa_read_optimization_summaries_1): Likewise.
6109 (execute_ipa_stmt_fixups): Likewise.
6110 * tree-pass.h (opt_pass::gate): Add function * argument.
6111 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
6112 combine-stack-adj.c, combine.c, compare-elim.c,
6113 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6114 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
6115 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
6116 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
6117 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
6118 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
6119 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6120 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
6121 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
6122 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
6123 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
6124 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
6125 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
6126 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6127 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6128 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6129 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6130 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6131 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6132 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6133 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6134 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6135 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6136 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
6137 var-tracking.c, vtable-verify.c, web.c: Adjust.
6139 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6141 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
6142 * configure: Regenerate.
6144 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6146 * passes.c (dump_one_pass): don't check pass->has_gate.
6147 (execute_ipa_summary_passes): Likewise.
6148 (execute_one_pass): Likewise.
6149 (ipa_write_summaries_2): Likewise.
6150 (ipa_write_optimization_summaries_1): Likewise.
6151 (ipa_read_optimization_summaries_1): Likewise.
6152 (execute_ipa_stmt_fixups): Likewise.
6153 * tree-pass.h (pass_data::has_gate): Remove.
6154 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
6155 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
6156 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
6157 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
6158 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
6159 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
6160 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
6161 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
6162 gimple-low.c, gimple-ssa-isolate-paths.c,
6163 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
6164 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6165 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
6166 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
6167 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
6168 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
6169 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
6170 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6171 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
6172 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
6173 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
6174 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6175 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6176 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
6177 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
6178 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
6179 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6180 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6181 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
6182 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
6183 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
6184 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
6187 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6189 * pass_manager.h (pass_manager::register_dump_files_1): Remove
6191 * passes.c (pass_manager::register_dump_files_1): Merge into
6192 (pass_manager::register_dump_files): this, and remove its handling of
6193 properties since the pass always has the properties anyway.
6194 (pass_manager::pass_manager): Adjust.
6196 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
6198 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
6199 * passes.c (pass_manager::register_dump_files_1): Remove dead code
6200 dealing with properties.
6201 (pass_manager::register_dump_files): Adjust.
6203 2014-03-20 Mark Wielaard <mjw@redhat.com>
6205 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
6206 then represent the bound as normal constant value.
6208 2014-04-17 Jakub Jelinek <jakub@redhat.com>
6211 Forward port from 4.8 branch
6212 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
6214 * config/i386/bmiintrin.h (_blsi_u32): New.
6218 (_blsmsk_u32): Ditto.
6219 (_blsmsk_u64): Ditto.
6220 (_tzcnt_u32): Ditto.
6221 (_tzcnt_u64): Ditto.
6223 2014-04-17 Kito Cheng <kito@0xlab.org>
6225 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
6227 2014-04-17 Richard Biener <rguenther@suse.de>
6230 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
6231 boolean results for comparisons.
6233 2014-04-17 Richard Biener <rguenther@suse.de>
6235 PR tree-optimization/60836
6236 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
6237 initial PHI args to be gimple values.
6239 2014-04-17 Richard Biener <rguenther@suse.de>
6241 PR tree-optimization/60841
6242 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
6243 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
6244 of stmts to SLP build.
6245 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
6246 (vect_analyze_slp): Likewise.
6247 (vect_analyze_slp_instance): Likewise.
6248 (vect_build_slp_tree): Limit overall SLP tree growth.
6249 * tree-vectorizer.h (vect_analyze_data_refs,
6250 vect_analyze_slp): Adjust prototypes.
6252 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6254 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
6257 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6259 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
6260 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
6261 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
6262 for TARGET_SLOW_PSHUFB
6264 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
6266 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
6267 * config/i386/i386.c (intel_cost): Ditto.
6269 2014-04-17 Joey Ye <joey.ye@arm.com>
6271 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
6273 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6275 * opts.c (common_handle_option): Disable -fipa-reference coorectly
6278 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6280 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
6281 (type_all_derivations_known_p): New predicate.
6282 (type_all_ctors_visible_p): New predicate.
6283 (type_possibly_instantiated_p): New predicate.
6284 (get_odr_type): Compute all_derivations_known.
6285 (dump_odr_type): Dump the flag.
6286 (maybe_record_type): Cleanup.
6287 (record_target_from_binfo): Add bases_to_consider array;
6288 record bases for types w/o instances and skip CXX destructor.
6289 (possible_polymorphic_call_targets_1): Add bases_to_consider
6290 and consider_construction parameters; check if type may have instance.
6291 (get_polymorphic_call_info): Set maybe_in_construction to true
6292 when we know nothing.
6293 (record_targets_from_bases): Skip CXX destructors; they are
6294 never called for types in construction.
6295 (possible_polymorphic_call_targets): Do not record target when
6296 type may not have instance.
6298 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
6301 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
6302 external aliases alive, too.
6304 2014-04-16 Andrew Pinski <apinski@cavium.com>
6306 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
6309 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6311 * final.c (compute_alignments): Do not apply loop alignment to a block
6312 falling through to the exit.
6314 2014-04-16 Catherine Moore <clm@codesourcery.com>
6316 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
6317 Adjust constraints for microMIPS store patterns.
6319 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6321 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
6323 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
6325 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
6326 (append_use): Run at -O0.
6327 (append_vdef): Likewise.
6328 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
6329 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
6331 2014-04-16 Jakub Jelinek <jakub@redhat.com>
6333 PR tree-optimization/60844
6334 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
6335 (propagate_op_to_single_use, remove_visited_stmt_chain,
6336 linearize_expr, repropagate_negates, reassociate_bb): Use it
6337 instead of gsi_remove.
6339 2014-04-16 Martin Jambor <mjambor@suse.cz>
6341 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
6342 ipa_transforms_to_apply.
6343 (cgraph_function_versioning): Assert that old_node has empty
6344 ipa_transforms_to_apply.
6345 * trans-mem.c (ipa_tm_create_version): Likewise.
6346 * tree-inline.c (tree_function_versioning): Do not duplicate
6347 ipa_transforms_to_apply.
6349 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6352 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
6354 Pass necessary as flags on 64-bit Solaris/x86.
6355 Use lowercase relocs for x86_64-*-*.
6356 * configure: Regenerate.
6358 2014-04-15 Jan Hubicka <jh@suse.cz>
6360 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
6361 (maybe_record_node, likely_target_p): Use it.
6363 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6366 Revert following patch
6368 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6371 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
6372 software floating point or no floating point registers, do not
6373 allow any type in the FPRs. Eliminate a test for SPE SIMD types
6374 in GPRs that occurs after we tested for GPRs that would never be
6377 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
6378 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
6379 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
6380 specifically allow DDmode, since that does not use the SPE SIMD
6383 2014-03-21 Mark Wielaard <mjw@redhat.com>
6385 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
6386 as unsigned or int depending on type and value used.
6388 2014-04-15 Richard Biener <rguenther@suse.de>
6390 PR rtl-optimization/56965
6391 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
6392 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
6394 * alias.c (true_dependence_1): Do not call
6395 nonoverlapping_component_refs_p.
6396 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
6397 nonoverlapping_component_refs_p.
6398 (indirect_refs_may_alias_p): Likewise.
6400 2014-04-15 Teresa Johnson <tejohnson@google.com>
6402 * cfg.c (dump_bb_info): Fix flags check.
6403 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
6405 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6407 PR rtl-optimization/60663
6408 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
6411 2014-04-15 Richard Biener <rguenther@suse.de>
6413 * lto-streamer.h (LTO_major_version): Bump to 4.
6415 2014-04-15 Richard Biener <rguenther@suse.de>
6417 * common.opt (lto_partition_model): New enum.
6418 (flto-partition=): Merge separate options with a single with argument,
6419 add -flto-partition=one support.
6420 * flag-types.h (enum lto_partition_model): Declare.
6421 * opts.c (finish_options): Remove duplicate -flto-partition=
6423 * lto-wrapper.c (run_gcc): Adjust.
6425 2014-04-15 Richard Biener <rguenther@suse.de>
6427 * alias.c (ncr_compar): New function.
6428 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
6430 2014-04-15 Richard Biener <rguenther@suse.de>
6432 * alias.c (record_component_aliases): Do not walk BINFOs.
6434 2014-04-15 Richard Biener <rguenther@suse.de>
6436 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
6437 Add struct function argument and adjust.
6438 (find_func_aliases_for_call): Likewise.
6439 (find_func_aliases): Likewise.
6440 (find_func_clobbers): Likewise.
6441 (intra_create_variable_infos): Likewise.
6442 (compute_points_to_sets): Likewise.
6443 (ipa_pta_execute): Adjust. Do not push/pop cfun.
6445 2014-04-15 Richard Biener <rguenther@suse.de>
6447 * tree.c (iterative_hash_expr): Use enum tree_code_class
6448 to store TREE_CODE_CLASS.
6449 (tree_block): Likewise.
6450 (tree_set_block): Likewise.
6451 * tree.h (fold_build_pointer_plus_loc): Use
6452 convert_to_ptrofftype_loc.
6454 2014-04-15 Jakub Jelinek <jakub@redhat.com>
6457 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
6460 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
6462 * cfgloop.h (struct loop): Move force_vectorize down.
6463 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
6464 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
6465 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
6466 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
6467 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
6468 * tree-core.h (enum annot_expr_kind): Add new kind values.
6469 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
6470 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
6472 * tree.def (ANNOTATE_EXPR): Tweak comment.
6474 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6476 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
6479 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
6481 * tree.h (TYPE_IDENTIFIER): Declare.
6482 * tree.c (subrange_type_for_debug_p): Use it.
6483 * godump.c (go_format_type): Likewise.
6484 * dwarf2out.c (is_cxx_auto, modified_type_die,
6485 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
6486 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
6488 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6491 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
6493 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
6495 * config/i386/i386.c (examine_argument): Return bool. Return true if
6496 parameter should be passed in memory.
6497 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
6498 (construct_container): Update calls to examine_argument.
6499 (function_arg_advance_64): Ditto.
6500 (return_in_memory_32): Merge with ix86_return_in_memory.
6501 (return_in_memory_64): Ditto.
6502 (return_in_memory_ms_64): Ditto.
6504 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
6506 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
6507 * coverage.c (coverage_compute_profile_id): Handle externally visible
6510 2014-04-14 Martin Jambor <mjambor@suse.cz>
6512 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
6513 DECL_DISREGARD_INLINE_LIMITS functions.
6515 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6518 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
6520 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
6523 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
6524 optimize_insn_for_speed_p instead of
6525 optimize_function_for_speed_p.
6527 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
6529 * doc/invoke.texi (free): Document AArch64.
6531 2014-04-14 Richard Biener <rguenther@suse.de>
6533 PR tree-optimization/60042
6534 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
6535 (insert_into_preds_of_block): Do not prevent PHI insertion
6536 for REFERENCE exprs here ...
6537 (eliminate_dom_walker::before_dom_children): ... but prevent
6538 their use here under similar conditions when applied to the
6539 IL after PRE optimizations.
6541 2014-04-14 Richard Biener <rguenther@suse.de>
6543 * passes.def: Move early points-to after early SRA.
6545 2014-04-14 Richard Biener <rguenther@suse.de>
6547 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
6548 check for which sign-changes we allow when forwarding
6549 a converted value into a switch.
6551 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6553 * stor-layout.c (place_field): Finalize non-constant offset for the
6556 2014-04-14 Richard Biener <rguenther@suse.de>
6558 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
6560 (expand_switch_using_bit_tests_p): Likewise.
6561 (process_switch): Compute and pass on speed_p based on the
6563 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
6564 optimize_bb_for_speed_p.
6566 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
6568 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
6569 * function.h (struct function): Rename has_force_vect_loops into
6570 has_force_vectorize_loops.
6571 * lto-streamer-in.c (input_cfg): Adjust for renaming.
6572 (input_struct_function_base): Likewise.
6573 * lto-streamer-out.c (output_cfg): Likewise.
6574 (output_struct_function_base): Likewise.
6575 * omp-low.c (expand_omp_simd): Likewise.
6576 * tree-cfg.c (move_sese_region_to_fn): Likewise.
6577 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
6578 (version_loop_for_if_conversion): Likewise.
6579 (tree_if_conversion): Likewise.
6580 (main_tree_if_conversion): Likewise.
6581 (gate_tree_if_conversion): Likewise.
6582 * tree-inline.c (copy_loops): Likewise.
6583 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
6584 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
6585 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
6586 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
6587 * tree-vectorizer.c (vectorize_loops): Likewise.
6588 * tree-vectorizer.h (unlimited_cost_model): Likewise.
6590 2014-04-14 Richard Biener <rguenther@suse.de>
6593 * lto-streamer-out.c (wrap_refs): New function.
6594 (lto_output): Wrap symbol references in global initializes in
6595 type-preserving MEM_REFs.
6597 2014-04-14 Christian Bruel <christian.bruel@st.com>
6599 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
6601 2014-04-14 Christian Bruel <christian.bruel@st.com>
6603 * config/sh/sh.md (setmemqi): New expand pattern.
6604 * config/sh/sh.h (CLEAR_RATIO): Define.
6605 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
6606 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
6608 2014-04-14 Richard Biener <rguenther@suse.de>
6611 * fold-const.c (negate_expr_p): Don't negate directional rounding
6613 (fold_negate_expr): Likewise.
6615 2014-04-14 Richard Biener <rguenther@suse.de>
6617 PR tree-optimization/59817
6618 PR tree-optimization/60453
6619 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
6620 recursion to catch all CHRECs in the scalar evolution and restrict
6621 the predicate for the remains appropriately.
6623 2014-04-12 Catherine Moore <clm@codesourcery.com>
6625 * config/mips/constraints.md: Add new register constraint "kb".
6626 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
6627 (*movhi_internal): Likewise.
6628 (*movqi_internal): Likewise.
6629 * config/mips/mips.h (M16_STORE_REGS): New register class.
6630 (REG_CLASS_NAMES): Add M16_STORE_REGS.
6631 (REG_CLASS_CONTENTS): Likewise.
6632 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
6634 2014-04-11 Tobias Burnus <burnus@net-b.de>
6637 * doc/invoke.texi (-Wformat-signedness): Document it.
6638 (Wformat=2): Mention that this enables -Wformat-signedness.
6640 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6642 * common/config/epiphany/epiphany-common.c
6643 (epiphany_option_optimization_table): Enable section anchors by
6644 default at -O1 or higher.
6645 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
6646 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
6647 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
6648 carries no extra cost.
6649 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
6650 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
6651 * config/epiphany/predicates.md (memclob_operand): New predicate.
6652 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
6653 Use memclob_operand predicate and X constraint for operand 3.
6655 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6657 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
6658 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
6661 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
6663 PR rtl-optimization/60651
6664 * mode-switching.c (optimize_mode_switching): Make sure to emit
6665 sets of a lower numbered entity before sets of a higher numbered
6666 entity to a mode of the same or lower priority.
6667 When creating a seginfo for a basic block that starts with a code
6668 label, move the insertion point past the code label.
6669 (new_seginfo): Document and enforce requirement that
6670 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
6671 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
6672 * doc/tm.texi: Regenerate.
6674 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
6677 * config/arc/arc.c (arc_save_restore): Fix assert typo.
6679 2013-04-11 Jakub Jelinek <jakub@redhat.com>
6681 * BASE-VER: Set to 4.10.0.
6683 2014-04-11 Tobias Burnus <burnus@net-b.de>
6686 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
6687 * doc/gcc.texi (Service): Update description in the @menu
6688 * doc/invoke.texi (Option Summary): Remove misplaced and
6691 2014-04-11 Steve Ellcey <sellcey@mips.com>
6692 Jakub Jelinek <jakub@redhat.com>
6695 * expr.c (convert_move): Use emit_store_flag_force instead of
6696 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
6699 2014-04-11 Richard Biener <rguenther@suse.de>
6702 * varasm.c (assemble_alias): Avoid endless error reporting
6703 recursion by setting TREE_ASM_WRITTEN.
6705 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6707 * config/s390/s390.md: Add a splitter for NOT rtx.
6709 2014-04-11 Jakub Jelinek <jakub@redhat.com>
6711 PR rtl-optimization/60663
6712 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
6714 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
6715 Jakub Jelinek <jakub@redhat.com>
6718 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
6719 flag from decl_node to node.
6721 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6724 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
6725 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
6726 ameliorating the cases where it can be.
6728 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
6731 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6733 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6734 (loadsync_<mode>): Change mode.
6735 (load_quadpti, store_quadpti): New.
6736 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6737 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6738 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
6740 2014-04-09 Cong Hou <congh@google.com>
6743 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
6746 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6748 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
6749 instead of vnor to exploit possible fusion opportunity in the
6751 (altivec_expand_vec_perm_const_le): Likewise.
6753 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
6755 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
6756 (loadsync_<mode>): Change mode.
6757 (load_quadpti, store_quadpti): New.
6758 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
6759 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
6761 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
6764 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
6765 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
6766 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
6768 2014-04-08 Richard Biener <rguenther@suse.de>
6771 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
6772 a 64bit widest int print double-int similar to on HWI64 hosts.
6774 2014-04-08 Richard Biener <rguenther@suse.de>
6776 PR tree-optimization/60785
6777 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
6778 default defs properly.
6780 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
6782 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
6783 (Weffc++): Likewise.
6785 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
6787 * ipa-devirt.c (maybe_record_node): When node is not recorded,
6788 set completep to false rather than true.
6790 2014-04-07 Douglas B Rupp <rupp@adacore.com>
6793 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
6794 ARM_TARGET2_DWARF_FORMAT.
6796 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
6799 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
6800 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
6803 2014-04-07 Richard Biener <rguenther@suse.de>
6805 PR tree-optimization/60766
6806 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
6807 (may_eliminate_iv): Convert cand_value_at result to desired type.
6809 2014-04-07 Jason Merrill <jason@redhat.com>
6812 * common.opt (-fno-gnu-unique): Add.
6813 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
6815 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6817 * haifa-sched.c: Fix outdated function reference and minor
6818 grammar errors in introductory comment.
6820 2014-04-07 Richard Biener <rguenther@suse.de>
6823 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
6825 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
6827 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
6830 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
6831 size accounting for thunks.
6832 (pa_asm_output_mi_thunk): Use final_start_function() and
6833 final_end_function() to output function start and end directives.
6835 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
6837 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
6838 device specific ISA/ feature information. Remove short_sp and
6839 errata_skip ds. Add avr_device_specific_features enum to have device
6841 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
6842 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
6843 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
6844 updated device specific info.
6845 * config/avr/avr-mcus.def: Merge device specific details to
6846 dev_attribute field.
6847 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
6849 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
6850 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
6851 assembler if RMW isa supported by current device.
6852 * config/avr/genmultilib.awk: Update as device info structure changed.
6853 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
6855 2014-04-04 Cong Hou <congh@google.com>
6857 PR tree-optimization/60656
6858 * tree-vect-stmts.c (supportable_widening_operation):
6859 Fix a bug that elements in a vector with vect_used_by_reduction
6860 property are incorrectly reordered when the operation on it is not
6861 consistant with the one in reduction operation.
6863 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
6865 PR rtl-optimization/60155
6866 * gcse.c (record_set_data): New function.
6867 (single_set_gcse): New function.
6868 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
6869 (hoist_code): Likewise.
6870 (get_pressure_class_and_nregs): Likewise.
6872 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
6874 * explow.c (probe_stack_range): Emit a final optimization blockage.
6876 2014-04-04 Anthony Green <green@moxielogic.com>
6878 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
6881 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
6884 * lto-cgraph.c (input_overwrite_node): Check that partitioning
6885 flags are set only during streaming.
6886 * ipa.c (process_references, walk_polymorphic_call_targets,
6887 symtab_remove_unreachable_nodes): Drop bodies of always inline
6888 after early inlining.
6889 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
6891 2014-04-04 Jakub Jelinek <jakub@redhat.com>
6892 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6895 * dwarf2out.c (const_ok_for_output_1): Reject expressions
6898 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6901 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
6903 (cortex_a53_fdivd): Likewise.
6905 2014-04-04 Martin Jambor <mjambor@suse.cz>
6908 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
6910 * cgraph.c (clone_of_p): Also return true if thunks match.
6911 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
6912 cgraph_function_or_thunk_node and an obsolete comment.
6913 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
6915 (build_function_decl_skip_args): Likewise.
6916 (set_new_clone_decl_and_node_flags): New function.
6917 (duplicate_thunk_for_node): Likewise.
6918 (redirect_edge_duplicating_thunks): Likewise.
6919 (cgraph_clone_node): New parameter args_to_skip, pass it to
6920 redirect_edge_duplicating_thunks which is called instead of
6921 cgraph_redirect_edge_callee.
6922 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
6923 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
6925 2014-04-04 Jeff Law <law@redhat.com>
6928 * config/arm/predicates.md (const_int_I_operand): New predicate.
6929 (const_int_M_operand): Similarly.
6930 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
6932 (insv_t2, extv_reg, extzv_t2): Likewise.
6933 (load_multiple_with_writeback): Similarly for const_int_I_operand.
6934 (pop_multiple_with_writeback_and_return): Likewise.
6935 (vfp_pop_multiple_with_writeback): Likewise
6937 2014-04-04 Richard Biener <rguenther@suse.de>
6940 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
6941 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
6943 * gimplify.h (gimple_add_tmp_var_fn): Declare.
6944 * gimplify.c (gimple_add_tmp_var_fn): New function.
6945 * gimple-expr.h (create_tmp_reg_fn): Declare.
6946 * gimple-expr.c (create_tmp_reg_fn): New function.
6947 * gimple-low.c (record_vars_into): Don't change cfun.
6948 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
6949 code generation without cfun.
6951 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
6954 * Makefile.in (install-driver): Fix shell scripting.
6956 2014-04-03 Cong Hou <congh@google.com>
6958 PR tree-optimization/60505
6959 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
6960 threshold of number of iterations below which no vectorization
6962 * tree-vect-loop.c (new_loop_vec_info):
6963 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
6964 * tree-vect-loop.c (vect_analyze_loop_operations):
6965 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
6966 * tree-vect-loop.c (vect_transform_loop):
6967 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
6968 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
6969 of iterations of the loop and see if we should build the epilogue.
6971 2014-04-03 Richard Biener <rguenther@suse.de>
6973 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
6974 (streamer_tree_cache_create): Adjust.
6975 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
6976 to allow optional nodes array.
6977 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
6978 (streamer_tree_cache_append): Likewise.
6979 (streamer_tree_cache_create): Create nodes array optionally
6980 as specified by parameter.
6981 * lto-streamer-out.c (create_output_block): Avoid maintaining
6982 the node array in the writer cache.
6983 (DFS_write_tree): Remove assertion.
6984 (produce_asm_for_decls): Free the out decl state hash table early.
6985 * lto-streamer-in.c (lto_data_in_create): Adjust for
6986 streamer_tree_cache_create prototype change.
6988 2014-04-03 Richard Biener <rguenther@suse.de>
6990 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
6991 set TREE_CHAIN to NULL_TREE.
6993 2014-04-03 Richard Biener <rguenther@suse.de>
6995 PR tree-optimization/60740
6996 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
6997 over all GIMPLE_COND operands.
6999 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
7001 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
7002 (Weffc++): Remove Scott's numbering, merge lists and reference
7005 2014-04-03 Nick Clifton <nickc@redhat.com>
7007 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
7010 2014-04-03 Martin Jambor <mjambor@suse.cz>
7012 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
7013 mention gcc_unreachable before failing.
7014 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
7017 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
7020 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
7021 inconsistent code and instead mark the context inconsistent.
7022 (possible_polymorphic_call_targets): For inconsistent contexts
7023 return empty complete list.
7025 2014-04-02 Anthony Green <green@moxielogic.com>
7027 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
7028 (extendqisi2, extendhisi2): Define.
7029 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
7030 (WCHAR_TYPE): Change to unsigned int.
7032 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7034 PR tree-optimization/60733
7035 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
7036 insertion point for PHI candidates to be the end of the feeding
7037 block for the PHI argument.
7039 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
7041 PR rtl-optimization/60650
7042 * lra-constraints.c (process_alt_operands): Decrease reject for
7043 earlyclobber matching.
7045 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7047 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
7049 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7051 * config/spu/spu.c (pad_bb): Do not crash when the last
7052 insn is CODE_FOR_blockage.
7054 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7056 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
7057 lies outside the target mode.
7059 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
7062 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
7063 software floating point or no floating point registers, do not
7064 allow any type in the FPRs. Eliminate a test for SPE SIMD types
7065 in GPRs that occurs after we tested for GPRs that would never be
7068 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
7069 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
7070 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
7071 specifically allow DDmode, since that does not use the SPE SIMD
7074 2014-04-02 Richard Biener <rguenther@suse.de>
7077 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
7078 MODE_INTs. Properly use negv_optab.
7079 (expand_abs): Likewise.
7081 2014-04-02 Richard Biener <rguenther@suse.de>
7084 * Makefile.in (install-driver): Guard extra installs with special
7087 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
7089 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7092 2014-04-01 Richard Henderson <rth@redhat.com>
7095 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
7096 alternative enabled before register allocation.
7098 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
7100 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
7101 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
7103 (nios2_large_got_address): Remove unneeded 'sym' parameter.
7104 (nios2_got_address): Update nios2_large_got_address call site.
7105 (nios2_delegitimize_address): New function.
7106 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
7107 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
7108 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
7110 2014-04-01 Martin Husemann <martin@duskware.de>
7112 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
7115 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
7117 PR rtl-optimization/60604
7118 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
7119 check from register_operand.
7120 (register_operand): Redefine in terms of general_operand.
7121 (nonmemory_operand): Use register_operand for the non-constant cases.
7123 2014-04-01 Richard Biener <rguenther@suse.de>
7125 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
7127 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
7129 * doc/invoke.texi (mapp-regs): Clarify.
7131 2014-03-31 Ulrich Drepper <drepper@gmail.com>
7133 * config/i386/avx512fintrin.h (__v32hi): Define type.
7134 (__v64qi): Likewise.
7135 (_mm512_set1_epi8): Define.
7136 (_mm512_set1_epi16): Define.
7137 (_mm512_set4_epi32): Define.
7138 (_mm512_set4_epi64): Define.
7139 (_mm512_set4_pd): Define.
7140 (_mm512_set4_ps): Define.
7141 (_mm512_setr4_epi64): Define.
7142 (_mm512_setr4_epi32): Define.
7143 (_mm512_setr4_pd): Define.
7144 (_mm512_setr4_ps): Define.
7145 (_mm512_setzero_epi32): Define.
7147 2014-03-31 Martin Jambor <mjambor@suse.cz>
7150 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
7151 callsite_arguments_match_p. Updated all callers. Also check types of
7152 corresponding formal parameters and actual arguments.
7153 (not_all_callers_have_enough_arguments_p) Renamed to
7154 some_callers_have_mismatched_arguments_p.
7156 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
7158 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
7160 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
7163 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
7166 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
7168 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
7169 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
7171 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
7172 Use FMAMODE_NOVF512 mode iterator.
7173 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
7174 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
7175 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
7176 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
7178 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
7179 Use VF_128_256 mode iterator.
7180 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
7183 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7185 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
7186 static chain if needed.
7188 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7191 * lra-constraints.c (index_part_to_reg): New.
7192 (process_address): Use it.
7194 2014-03-27 Jeff Law <law@redhat.com>
7195 Jakub Jelinek <jakub@redhat.com>
7198 * expr.c (do_tablejump): Use simplify_gen_binary rather than
7199 gen_rtx_{PLUS,MULT} to build up the address expression.
7201 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
7202 creating non-canonical RTL.
7204 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7207 * ipa-inline.c (want_inline_small_function_p): Short circuit large
7208 functions; reorganize to make cheap checks first.
7209 (inline_small_functions): Do not estimate growth when dumping;
7211 * ipa-inline.h (inline_summary): Add min_size.
7212 (growth_likely_positive): New function.
7213 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
7214 (set_cond_stmt_execution_predicate): Cleanup.
7215 (estimate_edge_size_and_time): Compute min_size.
7216 (estimate_calls_size_and_time): Likewise.
7217 (estimate_node_size_and_time): Likewise.
7218 (inline_update_overall_summary): Update min_size.
7219 (do_estimate_edge_time): Likewise.
7220 (do_estimate_edge_size): Update.
7221 (do_estimate_edge_hints): Update.
7222 (growth_likely_positive): New function.
7224 2014-03-28 Jakub Jelinek <jakub@redhat.com>
7227 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
7228 also if addr has VOIDmode.
7230 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7232 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
7233 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
7235 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
7236 instructions as well as AdvancedSIMD loads.
7238 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7240 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
7241 Use crypto_aese type.
7242 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
7243 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
7244 crypto_aese, crypto_aesmc. Move to types.md.
7245 * config/arm/types.md (crypto_aes): Split into crypto_aese,
7247 * config/arm/iterators.md (crypto_type): Likewise.
7249 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
7251 * cgraph.c: Include expr.h and tree-dfa.h.
7252 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
7255 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
7258 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
7259 regs from checking multi-reg pseudos.
7261 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7263 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
7265 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7267 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
7268 if it would clobber the stack pointer, even temporarily.
7270 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
7272 * mode-switching.c: Make small adjustments to the top comment.
7274 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7276 * config/rs6000/constraints.md (wD constraint): New constraint to
7277 match the constant integer to get the top DImode/DFmode out of a
7278 vector in a VSX register.
7280 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
7281 match the constant integer to get the top DImode/DFmode out of a
7282 vector in a VSX register.
7284 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
7287 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
7290 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
7291 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
7293 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
7294 Optimize vec_extract of 64-bit values, where the value being
7295 extracted is in the top word, where we can use scalar
7296 instructions. Add direct move and store support. Combine the big
7297 endian/little endian vector select load support into a single insn.
7298 (vsx_extract_<mode>_internal1): Likewise.
7299 (vsx_extract_<mode>_internal2): Likewise.
7300 (vsx_extract_<mode>_load): Likewise.
7301 (vsx_extract_<mode>_store): Likewise.
7302 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
7303 combined into vsx_extract_<mode>_load.
7304 (vsx_extract_<mode>_one_le): Likewise.
7306 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
7307 define the top 64-bit vector element.
7309 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
7312 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7313 Document vec_vbpermq builtin.
7316 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
7317 enable use of xxsldwi and xxpermdi builtin functions.
7318 (vec_xxpermdi): Likewise.
7320 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
7321 Document use of vec_xxsldwi and vec_xxpermdi builtins.
7323 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
7325 PR rtl-optimization/60650
7326 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
7328 (find_spills_for): New.
7329 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
7330 Spill all pseudos on the second iteration.
7332 2014-03-27 Marek Polacek <polacek@redhat.com>
7335 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
7338 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7340 * config/s390/s390.c (s390_can_use_return_insn): Check for
7341 call-saved FPRs on 31 bit.
7343 2014-03-27 Jakub Jelinek <jakub@redhat.com>
7346 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
7347 if they need regimplification, just drop them instead of
7348 calling gimple_regimplify_operands on them.
7350 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
7353 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
7354 (aarch64_frame_pointer_required): Adjust logic.
7355 (aarch64_can_eliminate): Adjust logic.
7356 (aarch64_override_options_after_change): Adjust logic.
7358 2014-03-27 Dehao Chen <dehao@google.com>
7360 * ipa-inline.c (early_inliner): Update node's inline info.
7362 2014-03-26 Dehao Chen <dehao@google.com>
7364 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
7365 compiler inserted conditional jumps for NAN float check.
7367 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7369 * ubsan.h (ubsan_create_data): Change second argument's type
7370 to const location_t *.
7371 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
7373 (ubsan_create_data): Change second argument to const location_t *PLOC.
7374 Create Loc field whenever PLOC is non-NULL.
7375 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
7376 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
7380 * real.c (real_to_integer2): Change type of low to UHWI.
7382 2014-03-26 Tobias Burnus <burnus@net-b.de>
7384 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
7385 (CILK_SELF_SPECS): New define.
7386 (driver_self_specs): Use it.
7388 2014-03-26 Richard Biener <rguenther@suse.de>
7390 * tree-pretty-print.c (percent_K_format): Implement special
7391 case for LTO and its stripped down BLOCK tree.
7393 2014-03-26 Jakub Jelinek <jakub@redhat.com>
7396 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
7398 * tree-vrp.c (simplify_internal_call_using_ranges): If only
7399 one range is range_int_cst_p, but not both, at least optimize
7400 addition/subtraction of 0 and multiplication by 0 or 1.
7401 * gimple-fold.c (gimple_fold_call): Fold
7402 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
7403 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
7404 INTEGER_CSTs, try to fold at least x * 0 and y - y.
7406 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
7408 PR rtl-optimization/60452
7409 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
7410 <case REG>: Return 1 for invalid offsets from the frame pointer.
7412 2014-03-26 Marek Polacek <polacek@redhat.com>
7415 * doc/extend.texi (C Extensions): Mention variable-length arrays in
7418 2014-03-26 Marek Polacek <polacek@redhat.com>
7421 * doc/extend.texi (Designated Inits): Describe what happens to omitted
7424 2014-03-26 Marek Polacek <polacek@redhat.com>
7427 * ira-color.c (update_conflict_hard_regno_costs): Perform the
7428 multiplication in unsigned type.
7430 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7432 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
7434 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
7436 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
7438 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7441 * cif-code.def (UNREACHABLE) New code.
7442 * ipa-inline.c (inline_small_functions): Skip edges to
7443 __builtlin_unreachable.
7444 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
7445 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
7446 predicate to __bulitin_unreachable.
7447 (set_cond_stmt_execution_predicate): Fix issue when
7448 invert_tree_comparison returns ERROR_MARK.
7449 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
7450 propagate to inline clones.
7451 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
7453 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
7454 * cgraphclones.c (cgraph_clone_node): If call destination is already
7455 ureachable, do not redirect it back.
7456 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
7459 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
7461 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
7462 Do not modify inline clones.
7464 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7466 * config/i386/i386.md (general_sext_operand): New mode attr.
7467 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
7468 don't generate (sign_extend (const_int)).
7469 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
7470 operands[2]. Use We constraint instead of <i> and
7471 <general_sext_operand> predicate instead of <general_operand>.
7472 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
7473 * config/i386/constraints.md (We): New constraint.
7474 * config/i386/predicates.md (x86_64_sext_operand,
7475 sext_operand): New predicates.
7477 2014-03-25 Martin Jambor <mjambor@suse.cz>
7480 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
7481 inconsistent devirtualizations to __builtin_unreachable.
7483 2014-03-25 Marek Polacek <polacek@redhat.com>
7486 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
7488 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
7490 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
7491 order of elements for big-endian.
7493 2014-03-25 Richard Biener <rguenther@suse.de>
7496 * gimplify-me.c (gimple_regimplify_operands): Update the
7499 2014-03-25 Martin Jambor <mjambor@suse.cz>
7502 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
7503 (lto_output_varpool_node): Likewise.
7504 (input_overwrite_node): Likewise.
7505 (input_varpool_node): Likewise.
7507 2014-03-25 Richard Biener <rguenther@suse.de>
7509 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
7510 (run_gcc): Likewise.
7512 2014-03-25 Jakub Jelinek <jakub@redhat.com>
7514 * combine.c (simplify_compare_const): Add MODE argument.
7515 Handle mode_width 0 as very large mode_width.
7516 (try_combine, simplify_comparison): Adjust callers.
7518 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
7519 type to avoid signed integer overflow.
7520 * explow.c (plus_constant): Likewise.
7522 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
7524 * doc/generic.texi: Correct typos.
7526 2014-03-24 Tobias Burnus <burnus@net-b.de>
7528 * doc/invoke.texi (-flto): Expand section about
7529 using static libraries with LTO.
7531 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7533 PR rtl-optimization/60501
7534 * optabs.def (addptr3_optab): New optab.
7535 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
7536 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
7537 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
7539 * lra.c (emit_add3_insn): Use the addptr pattern if available.
7541 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
7543 2014-03-24 Ulrich Drepper <drepper@gmail.com>
7545 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
7548 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
7549 (_mm256_undefined_ps): Define.
7550 (_mm256_undefined_pd): Define.
7551 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
7552 (_mm_undefined_pd): Define.
7553 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
7554 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
7555 (_mm512_undefined_ps): Define.
7556 (_mm512_undefined_pd): Define.
7557 Use _mm*_undefined_*.
7558 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
7560 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
7562 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
7563 (lshr_simd): DI mode added.
7564 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
7565 (aarch64_ushr_simddi): Likewise.
7566 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
7567 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
7568 (vshrd_n_u64): Likewise.
7570 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7572 * Makefile.in (s-macro_list): Depend on cc1.
7574 2014-03-23 Teresa Johnson <tejohnson@google.com>
7576 * ipa-utils.c (ipa_print_order): Use specified dump file.
7578 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
7580 PR rtl-optimization/60601
7581 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
7583 * gcc.c (eval_spec_function): Initialize save_growing_value.
7585 2014-03-22 Jakub Jelinek <jakub@redhat.com>
7588 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
7589 code == MINUS_EXPR, never swap op0 with op1.
7591 * toplev.c (init_local_tick): Avoid signed integer multiplication
7593 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
7594 shift by first operand's bitsize.
7596 2014-03-21 Jakub Jelinek <jakub@redhat.com>
7599 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
7601 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
7602 TARGET_ISA_64BIT_P(x).
7604 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7606 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
7607 pattern for vector nor instead of subtract from splat(-1).
7608 (altivec_expand_vec_perm_const_le): Likewise.
7610 2014-03-21 Richard Henderson <rth@twiddle.net>
7613 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
7614 related insns after epilogue_completed.
7616 2014-03-21 Martin Jambor <mjambor@suse.cz>
7619 * cgraph.h (symtab_node): New flag body_removed.
7620 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
7621 when removing bodies.
7622 * symtab.c (dump_symtab_base): Dump body_removed flag.
7623 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
7624 had their bodies removed.
7626 2014-03-21 Martin Jambor <mjambor@suse.cz>
7629 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
7632 2014-03-21 Richard Biener <rguenther@suse.de>
7634 PR tree-optimization/60577
7635 * tree-core.h (struct tree_base): Document nothrow_flag use
7636 in VAR_DECL_NONALIASED.
7637 * tree.h (VAR_DECL_NONALIASED): New.
7638 (may_be_aliased): Adjust.
7639 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
7641 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7643 * expr.c (expand_expr_real_1): Remove outdated comment.
7645 2014-03-20 Jakub Jelinek <jakub@redhat.com>
7648 * ira.c (adjust_cleared_regs): Call copy_rtx on
7649 *reg_equiv[REGNO (loc)].src_p before passing it to
7650 simplify_replace_fn_rtx.
7653 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
7654 into CONST, put pic register as first operand of PLUS. Use
7655 gen_const_mem for both 32-bit and 64-bit PIC got loads.
7657 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
7659 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
7661 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7663 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
7664 around for store forwarding issue in the FPU on the UT699.
7665 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
7666 loads and operations if -mfix-ut699 is specified.
7667 (divtf3_hq): Tweak attribute.
7668 (sqrttf2_hq): Likewise.
7670 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
7672 * calls.c (store_one_arg): Remove incorrect const qualification on the
7673 type of the temporary.
7674 * cfgexpand.c (expand_return): Likewise.
7675 * expr.c (expand_constructor): Likewise.
7676 (expand_expr_real_1): Likewise.
7678 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
7680 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
7683 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
7686 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
7688 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
7690 * config/arm/aarch-common-protos.h
7691 (alu_cost_table): Fix spelling of "extend".
7692 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
7694 2014-03-19 Richard Biener <rguenther@suse.de>
7697 * tree-core.h (tree_type_common): Re-order pointer members
7698 to reduce recursion depth during GC walks.
7700 2014-03-19 Marek Polacek <polacek@redhat.com>
7703 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
7704 before accessing it.
7706 2014-03-19 Richard Biener <rguenther@suse.de>
7709 * lto-streamer-in.c (input_function): In WPA stage do not drop
7712 2014-03-19 Jakub Jelinek <jakub@redhat.com>
7714 PR tree-optimization/60559
7715 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
7716 with build_zero_cst assignment.
7718 2014-03-18 Kai Tietz <ktietz@redhat.com>
7720 PR rtl-optimization/56356
7721 * sdbout.c (sdbout_parms): Verify that parms'
7722 incoming argument is valid.
7723 (sdbout_reg_parms): Likewise.
7725 2014-03-18 Richard Henderson <rth@redhat.com>
7728 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
7729 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
7730 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
7732 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
7734 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
7735 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
7736 Italicize plugin event names in description. Explain that
7737 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
7738 Remind that no GCC functions should be called after PLUGIN_FINISH.
7739 Explain what pragmas with expansion are.
7741 2014-03-18 Martin Liska <mliska@suse.cz>
7743 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
7744 gimple call statement is update.
7745 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
7746 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
7748 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7751 * ubsan.c (ubsan_instrument_unreachable): Call
7752 initialize_sanitizer_builtins.
7753 (ubsan_pass): Likewise.
7756 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
7757 varpool_finalize_decl instead of rest_of_decl_compilation.
7759 2014-03-18 Richard Biener <rguenther@suse.de>
7761 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
7762 by using bitmap_and_compl instead of bitmap_and_compl_into.
7763 (df_rd_transfer_function): Likewise.
7765 2014-03-18 Richard Biener <rguenther@suse.de>
7767 * doc/lto.texi (fresolution): Fix typo.
7769 2014-03-18 Richard Biener <rguenther@suse.de>
7771 * doc/invoke.texi (flto): Update for changes in 4.9.
7773 2014-03-18 Richard Biener <rguenther@suse.de>
7775 * doc/loop.texi: Remove section on the removed lambda framework.
7776 Update loop docs with recent changes in preserving loop structure.
7778 2014-03-18 Richard Biener <rguenther@suse.de>
7780 * doc/lto.texi (-fresolution): Document.
7782 2014-03-18 Richard Biener <rguenther@suse.de>
7784 * doc/contrib.texi: Adjust my name.
7786 2014-03-18 Jakub Jelinek <jakub@redhat.com>
7789 * internal-fn.c: Include diagnostic-core.h.
7790 (expand_BUILTIN_EXPECT): New function.
7791 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
7792 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
7793 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
7794 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
7796 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
7797 Revert 3 argument __builtin_expect code.
7798 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
7799 * gimple-fold.c (gimple_fold_call): Likewise.
7800 * tree.h (fold_builtin_expect): New prototype.
7801 * builtins.c (build_builtin_expect_predicate): Add predictor
7802 argument, if non-NULL, create 3 argument __builtin_expect.
7803 (fold_builtin_expect): No longer static. Add ARG2 argument,
7804 pass it through to build_builtin_expect_predicate.
7805 (fold_builtin_2): Adjust caller.
7806 (fold_builtin_3): Handle BUILT_IN_EXPECT.
7807 * internal-fn.def (BUILTIN_EXPECT): New.
7809 2014-03-18 Tobias Burnus <burnus@net-b.de>
7812 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
7813 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
7814 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
7816 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
7819 * predict.c (combine_predictions_for_bb): Fix up formatting.
7820 (expr_expected_value_1, expr_expected_value): Add predictor argument,
7821 fill what it points to if non-NULL.
7822 (tree_predict_by_opcode): Adjust caller, use the predictor.
7823 * predict.def (PRED_COMPARE_AND_SWAP): Add.
7825 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
7827 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
7828 proper constant for the store mode.
7830 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
7832 * symtab.c (change_decl_assembler_name): Fix transparent alias
7835 2014-03-16 Renlin Li <Renlin.Li@arm.com>
7837 * config/aarch64/aarch64.c: Correct the comments about the
7838 aarch64 stack layout.
7840 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
7842 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
7843 check for GF_OMP_FOR_KIND_FOR.
7845 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
7847 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
7848 ymm and zmm register names.
7850 2014-03-17 Jakub Jelinek <jakub@redhat.com>
7853 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
7854 note creation for the 2010-08-31 changes.
7856 2014-03-17 Marek Polacek <polacek@redhat.com>
7859 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
7860 as -fno-tree-loop-vectorize.
7861 (expand_omp_simd): Likewise.
7863 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
7865 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
7866 (eligible_for_call_delay): New prototype.
7867 * config/sparc/sparc.c (tls_call_delay): Rename into...
7868 (eligible_for_call_delay): ...this. Return false if the instruction
7869 cannot be put in the delay slot of a branch.
7870 (eligible_for_restore_insn): Simplify.
7871 (eligible_for_return_delay): Return false if the instruction cannot be
7872 put in the delay slot of a branch and simplify.
7873 (eligible_for_sibcall_delay): Return false if the instruction cannot be
7874 put in the delay slot of a branch.
7875 * config/sparc/sparc.md (fix_ut699): New attribute.
7876 (tls_call_delay): Delete.
7877 (in_call_delay): Reimplement.
7878 (eligible_for_sibcall_delay): Rename into...
7879 (in_sibcall_delay): ...this.
7880 (eligible_for_return_delay): Rename into...
7881 (in_return_delay): ...this.
7882 (in_branch_delay): Reimplement.
7883 (in_uncond_branch_delay): Delete.
7884 (in_annul_branch_delay): Delete.
7886 2014-03-14 Richard Henderson <rth@redhat.com>
7889 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
7890 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
7891 (*floathi<X87MODEF>2_i387_with_temp): Remove.
7892 (floathi splitters): Remove.
7893 (float<SWI48x>xf2): New pattern.
7894 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
7895 code that tried to handle DImode for 32-bit, but which was excluded
7896 by the pattern's condition. Drop allocation of stack temporary.
7897 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
7898 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
7899 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
7900 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
7901 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
7902 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
7903 (*float<SWI48><MODEF>2_sse_interunit): Remove.
7904 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
7905 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
7906 (*float<SWI48x><X87MODEF>2_i387): Remove.
7907 (all float _with_temp splitters): Remove.
7908 (*float<SWI48x><MODEF>2_i387): New pattern.
7909 (*float<SWI48><MODEF>2_sse): New pattern.
7910 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
7911 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
7913 2014-03-14 Jakub Jelinek <jakub@redhat.com>
7914 Marek Polacek <polacek@redhat.com>
7917 * common.opt (dump_base_name_prefixed): New Variable.
7918 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
7919 if x_dump_base_name_prefixed is already set, set it at the end.
7921 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
7923 PR rtl-optimization/60508
7924 * lra-constraints.c (get_reload_reg): Add new parameter
7926 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
7927 Pass the new parameter values.
7929 2014-03-14 Richard Biener <rguenther@suse.de>
7931 * common.opt: Revert unintented changes from r205065.
7934 2014-03-14 Richard Biener <rguenther@suse.de>
7937 * cfghooks.c (split_block): Properly adjust all loops the
7938 block was a latch of.
7940 2014-03-14 Martin Jambor <mjambor@suse.cz>
7943 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
7946 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
7949 * config/avr/avr.c (avr_set_current_function): Pass function name
7950 through default_strip_name_encoding before sanity checking instead
7951 of skipping the first char of the assembler name.
7953 2014-03-13 Richard Henderson <rth@redhat.com>
7956 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
7957 (ix86_force_to_memory, ix86_free_from_memory): Remove.
7958 * config/i386/i386-protos.h: Likewise.
7959 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
7960 in the expander instead of a splitter.
7961 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
7962 any possibility of requiring a memory.
7963 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
7964 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
7965 (fp branch splitters): Update for ix86_split_fp_branch.
7966 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
7967 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
7968 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
7969 (*fop_<MODEF>_2_i387): Remove f/r alternative.
7970 (*fop_<MODEF>_3_i387): Likewise.
7971 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
7972 (splitters for the fop_* register patterns): Remove.
7973 (fscalexf4_i387): Rename from *fscalexf4_i387.
7974 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
7976 2014-03-13 Jakub Jelinek <jakub@redhat.com>
7978 PR tree-optimization/59779
7979 * tree-dfa.c (get_ref_base_and_extent): Use double_int
7980 type for bitsize and maxsize instead of HOST_WIDE_INT.
7982 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
7984 PR rtl-optimization/57320
7985 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
7986 the CFG after thread_prologue_and_epilogue_insns.
7988 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
7990 PR rtl-optimization/57189
7991 * lra-constraints.c (process_alt_operands): Disfavor spilling
7994 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
7996 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
7998 2014-03-13 Jakub Jelinek <jakub@redhat.com>
8000 PR tree-optimization/59025
8002 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
8003 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
8005 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
8008 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
8009 calls of avr_out_plus_1.
8011 2014-03-13 Bin Cheng <bin.cheng@arm.com>
8013 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
8014 BB's single pred and update the father loop's latch info later.
8016 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
8018 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
8022 (VEC_base): Likewise.
8023 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
8024 registers, we need to swap double words in little endian mode.
8026 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
8027 to be a container mode for 128-bit integer operations added in ISA
8028 2.07. Unlike TImode and PTImode, the preferred register set is
8029 the Altivec/VMX registers for the 128-bit operations.
8031 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
8033 (rs6000_split_128bit_ok_p): Likewise.
8035 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
8036 macros for creating ISA 2.07 normal and overloaded builtin
8037 functions with 3 arguments.
8038 (BU_P8V_OVERLOAD_3): Likewise.
8039 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
8040 for use as overloaded functions.
8041 (VPERM_1TI_UNS): Likewise.
8042 (VSEL_1TI): Likewise.
8043 (VSEL_1TI_UNS): Likewise.
8044 (ST_INTERNAL_1ti): Likewise.
8045 (LD_INTERNAL_1ti): Likewise.
8046 (XXSEL_1TI): Likewise.
8047 (XXSEL_1TI_UNS): Likewise.
8048 (VPERM_1TI): Likewise.
8049 (VPERM_1TI_UNS): Likewise.
8050 (XXPERMDI_1TI): Likewise.
8051 (SET_1TI): Likewise.
8052 (LXVD2X_V1TI): Likewise.
8053 (STXVD2X_V1TI): Likewise.
8054 (VEC_INIT_V1TI): Likewise.
8055 (VEC_SET_V1TI): Likewise.
8056 (VEC_EXT_V1TI): Likewise.
8057 (EQV_V1TI): Likewise.
8058 (NAND_V1TI): Likewise.
8059 (ORC_V1TI): Likewise.
8060 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
8061 added in ISA 2.07. Add both normal 'altivec' builtins, and the
8063 (VADDUQM): Likewise.
8064 (VSUBCUQ): Likewise.
8065 (VADDEUQM): Likewise.
8066 (VADDECUQ): Likewise.
8067 (VSUBEUQM): Likewise.
8068 (VSUBECUQ): Likewise.
8070 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
8071 __int128_t and __uint128_t types.
8072 (__uint128_type): Likewise.
8073 (altivec_categorize_keyword): Add support for vector __int128_t,
8074 vector __uint128_t, vector __int128, and vector unsigned __int128
8075 as a container type for TImode operations that need to be done in
8076 VSX/Altivec registers.
8077 (rs6000_macro_to_expand): Likewise.
8078 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
8079 to support 128-bit integer instructions vaddcuq, vadduqm,
8080 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
8081 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
8083 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
8084 for V1TImode, and set up preferences to use VSX/Altivec registers.
8085 Setup VSX reload handlers.
8086 (rs6000_debug_reg_global): Likewise.
8087 (rs6000_init_hard_regno_mode_ok): Likewise.
8088 (rs6000_preferred_simd_mode): Likewise.
8089 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
8090 (easy_altivec_constant): Likewise.
8091 (output_vec_const_move): Likewise.
8092 (rs6000_expand_vector_set): Convert V1TImode set and extract to
8094 (rs6000_expand_vector_extract): Likewise.
8095 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
8097 (rs6000_const_vec): Add support for V1TImode.
8098 (rs6000_emit_le_vsx_load): Swap double words when loading or
8099 storing TImode/V1TImode.
8100 (rs6000_emit_le_vsx_store): Likewise.
8101 (rs6000_emit_le_vsx_move): Likewise.
8102 (rs6000_emit_move): Add support for V1TImode.
8103 (altivec_expand_ld_builtin): Likewise.
8104 (altivec_expand_st_builtin): Likewise.
8105 (altivec_expand_vec_init_builtin): Likewise.
8106 (altivec_expand_builtin): Likewise.
8107 (rs6000_init_builtins): Add support for V1TImode type. Add
8108 support for ISA 2.07 128-bit integer builtins. Define type names
8109 for the VSX/Altivec vector types.
8110 (altivec_init_builtins): Add support for overloaded vector
8111 functions with V1TImode type.
8112 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
8113 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
8115 (rs6000_split_128bit_ok_p): Likewise.
8116 (rs6000_handle_altivec_attribute): Create V1TImode from vector
8117 __int128_t and vector __uint128_t.
8119 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
8120 and mode attributes.
8127 (VS_scalar): Likewise.
8128 (VS_double): Likewise.
8129 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
8131 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
8132 we support the ISA 2.07 128-bit integer arithmetic instructions.
8133 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
8134 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
8135 and TImode types for use with the builtin functions.
8136 (V1TI_type_node): Likewise.
8137 (unsigned_V1TI_type_node): Likewise.
8138 (intTI_type_internal_node): Likewise.
8139 (uintTI_type_internal_node): Likewise.
8141 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
8142 128-bit builtin functions.
8143 (UNSPEC_VADDEUQM): Likewise.
8144 (UNSPEC_VADDECUQ): Likewise.
8145 (UNSPEC_VSUBCUQ): Likewise.
8146 (UNSPEC_VSUBEUQM): Likewise.
8147 (UNSPEC_VSUBECUQ): Likewise.
8148 (VM): Add V1TImode to vector mode iterators.
8150 (VI_unit): Likewise.
8151 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
8152 (altivec_vaddcuq): Likewise.
8153 (altivec_vsubuqm): Likewise.
8154 (altivec_vsubcuq): Likewise.
8155 (altivec_vaddeuqm): Likewise.
8156 (altivec_vaddecuq): Likewise.
8157 (altivec_vsubeuqm): Likewise.
8158 (altivec_vsubecuq): Likewise.
8160 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
8162 (BOOL_128): Likewise.
8163 (BOOL_REGS_OUTPUT): Likewise.
8164 (BOOL_REGS_OP1): Likewise.
8165 (BOOL_REGS_OP2): Likewise.
8166 (BOOL_REGS_UNARY): Likewise.
8167 (BOOL_REGS_AND_CR0): Likewise.
8169 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
8170 128-bit integer builtin support.
8171 (vec_vadduqm): Likewise.
8172 (vec_vaddecuq): Likewise.
8173 (vec_vaddeuqm): Likewise.
8174 (vec_vsubecuq): Likewise.
8175 (vec_vsubeuqm): Likewise.
8176 (vec_vsubcuq): Likewise.
8177 (vec_vsubuqm): Likewise.
8179 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
8180 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
8181 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
8182 128-bit integer add/subtract to ISA 2.07.
8184 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
8186 * config/arc/arc.c (arc_predicate_delay_insns):
8187 Fix third argument passed to conditionalize_nonjump.
8189 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
8191 * config/aarch64/aarch64-builtins.c
8192 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
8193 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
8194 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
8195 instead of __builtin_lfloor.
8196 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
8198 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8200 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
8201 (tree_ssa_ifcombine_bb_1): New function.
8202 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
8203 is an empty forwarder block to then_bb or vice versa and then_bb
8204 and else_bb are effectively swapped.
8206 2014-03-12 Christian Bruel <christian.bruel@st.com>
8209 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
8210 REG_CFA_DEF_CFA note.
8211 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
8212 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
8214 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
8216 PR tree-optimization/60454
8217 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
8219 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8221 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
8222 Do not define target_cpu_default2 to generic.
8223 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
8224 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
8225 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
8227 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8228 Marc Glisse <marc.glisse@inria.fr>
8230 PR tree-optimization/60502
8231 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
8232 instead of build_low_bits_mask.
8234 2014-03-12 Jakub Jelinek <jakub@redhat.com>
8237 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
8238 if there are multiple uses, but op doesn't live on E edge.
8239 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
8240 clobber stmts before __builtin_unreachable.
8242 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
8244 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
8245 hard_frame_pointer_rtx.
8246 * cse.c (cse_insn): Remove volatile check.
8247 * cselib.c (cselib_process_insn): Likewise.
8248 * dse.c (scan_insn): Likewise.
8250 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8252 * config/arc/arc.c (conditionalize_nonjump): New function,
8254 (arc_ifcvt): ... this.
8255 (arc_predicate_delay_insns): Use it.
8257 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
8259 * config/arc/predicates.md (extend_operand): During/after reload,
8260 allow const_int_operand.
8261 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
8262 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
8263 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
8265 (umulsi3_highpart_i): Likewise.
8267 2014-03-11 Richard Biener <rguenther@suse.de>
8269 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
8270 Add asserts to guard possible wrong-code bugs.
8272 2014-03-11 Richard Biener <rguenther@suse.de>
8274 PR tree-optimization/60429
8275 PR tree-optimization/60485
8276 * tree-ssa-structalias.c (set_union_with_increment): Properly
8277 take into account all fields that overlap the shifted vars.
8278 (do_sd_constraint): Likewise.
8279 (do_ds_constraint): Likewise.
8280 (get_constraint_for_ptr_offset): Likewise.
8282 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
8284 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
8285 (nios2_compute_frame_layout):
8286 Add calculation of cfun->machine->fp_save_offset.
8287 (nios2_expand_prologue): Correct setting of frame pointer register
8289 (nios2_expand_epilogue): Update recovery of stack pointer from
8290 frame pointer accordingly.
8291 (nios2_initial_elimination_offset): Update calculation of offset
8292 for eliminating to HARD_FRAME_POINTER_REGNUM.
8294 2014-03-10 Jakub Jelinek <jakub@redhat.com>
8297 * ipa.c (symtab_remove_unreachable_nodes): Don't call
8298 cgraph_get_create_node on VAR_DECLs.
8300 2014-03-10 Richard Biener <rguenther@suse.de>
8303 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
8305 2014-03-08 Douglas B Rupp <rupp@gnat.com>
8307 * config/vms/vms.opt (vms_float_format): New variable.
8309 2014-03-08 Tobias Burnus <burnus@net-b.de>
8311 * doc/invoke.texi (-fcilkplus): Update implementation status.
8313 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
8314 Richard Biener <rguenther@suse.de>
8316 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
8317 consistently accross all TUs.
8318 (run_gcc): Enable -fshort-double automatically at link at link-time
8319 and disallow override.
8321 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
8324 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
8325 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
8326 if they can't be used.
8328 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8330 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
8331 for Solaris 11/x86 ld.
8332 * configure: Regenerate.
8334 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8336 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
8337 (LIB_TLS_SPEC): Save as ld_tls_libs.
8338 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
8339 (HAVE_AS_IX86_TLSLDM): New test.
8340 * configure, config.in: Regenerate.
8341 * config/i386/i386.c (legitimize_tls_address): Fall back to
8342 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
8343 cannot support TLS_MODEL_LOCAL_DYNAMIC.
8344 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
8345 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
8347 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
8349 * common.opt (fira-loop-pressure): Mark as optimization.
8351 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
8353 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
8354 an OpenMP mappable type.
8356 2014-03-06 Matthias Klose <doko@ubuntu.com>
8358 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
8359 MULTILIB_OSDIRNAMES is not defined.
8361 2014-03-06 Jakub Jelinek <jakub@redhat.com>
8362 Meador Inge <meadori@codesourcery.com>
8365 * config/arm/arm.c (arm_tls_symbol_p): Remove.
8366 (arm_legitimize_address): Call legitimize_tls_address for any
8367 arm_tls_referenced_p expression, handle constant addend. Call it
8368 before testing for !TARGET_ARM.
8369 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
8371 2014-03-06 Richard Biener <rguenther@suse.de>
8377 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8379 * tree-streamer.c (record_common_node): Assert we don't record
8380 nodes with type double.
8381 (preload_common_node): Skip type double, complex double and double
8382 pointer since it is now frontend dependent due to fshort-double option.
8384 2014-03-06 Richard Biener <rguenther@suse.de>
8386 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
8387 or -fno-lto is specified and the linker has full plugin support.
8388 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
8389 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
8390 * lto-wrapper.c (merge_and_complain): Merge compile-time
8391 optimization levels.
8392 (run_gcc): And pass it through to the link options.
8394 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
8398 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8400 * cselib.c (remove_useless_values): Skip to avoid quadratic
8401 behavior if the condition moved from...
8402 (cselib_process_insn): ... here holds.
8404 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8407 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
8408 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
8411 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
8412 (TM_H): Add x86-tune.def.
8414 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8416 * config/aarch64/aarch64.c (generic_tunings):
8417 Use cortexa57_extra_costs.
8419 2014-03-05 Jakub Jelinek <jakub@redhat.com>
8422 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
8423 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
8424 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
8427 2014-03-04 Heiher <r@hev.cc>
8429 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
8430 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
8432 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
8434 * config/i386/predicates.md (const2356_operand): Change to ...
8435 (const2367_operand): ... this.
8436 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
8438 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8439 (*avx512pf_scatterpf<mode>sf): Ditto.
8440 (avx512pf_scatterpf<mode>df): Ditto.
8441 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8442 (*avx512pf_scatterpf<mode>df): Ditto.
8443 * config/i386/i386.c (ix86_expand_builtin): Update
8444 incorrect hint operand error message.
8446 2014-03-04 Richard Biener <rguenther@suse.de>
8448 * lto-section-in.c (lto_get_section_data): Fix const cast.
8450 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
8452 * tree-streamer.c (record_common_node): Assert we don't record
8453 nodes with type double.
8454 (preload_common_node): Skip type double, complex double and double
8455 pointer since it is now frontend dependent due to fshort-double option.
8457 2014-03-04 Richard Biener <rguenther@suse.de>
8460 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
8461 (lto_input_toplevel_asms): Likewise.
8462 * lto-section-in.c (lto_get_section_data): Instead do it here
8465 2014-03-04 Richard Biener <rguenther@suse.de>
8467 PR tree-optimization/60382
8468 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
8469 dead PHIs a reduction.
8471 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
8473 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
8475 (_mm_prefetch): Move out of GCC target("sse") pragma.
8476 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
8477 GCC target("prfchw") pragma.
8478 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
8480 * config/i386/i386.c (ix86_option_override_internal): Enable
8481 -mprfchw with -mprefetchwt1.
8483 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8485 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
8488 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
8490 * opts.h (CL_PCH_IGNORE): Define.
8491 * targhooks.c (option_affects_pch_p):
8492 Return false for options that have CL_PCH_IGNORE set.
8493 * opt-functions.awk: Process PchIgnore.
8494 * doc/options.texi: Document PchIgnore.
8496 * config/arc/arc.opt (misize): Add PchIgnore property.
8498 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8500 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
8501 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
8502 constraint on constants to permit them being loaded into
8503 GENERAL_REGS or BASE_REGS.
8505 2014-03-03 Nick Clifton <nickc@redhat.com>
8507 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
8508 anti-cacnonical alternatives.
8509 (negandhi3_real): New pattern.
8510 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
8512 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
8514 * config/avr/avr-mcus.def: Remove atxmega16x1.
8515 * config/avr/avr-tables.opt: Regenerate.
8516 * config/avr/t-multilib: Regenerate.
8517 * doc/avr-mmcu.texi: Regenerate.
8519 2014-03-03 Tobias Grosser <tobias@grosser.es>
8520 Mircea Namolaru <mircea.namolaru@inria.fr>
8522 PR tree-optimization/58028
8523 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
8526 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
8528 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
8529 not handled by recognizers.
8531 2014-03-03 Jakub Jelinek <jakub@redhat.com>
8534 * function.c (expand_function_end): Don't emit
8535 clobber_return_register sequence if clobber_after is a BARRIER.
8536 * cfgexpand.c (construct_exit_block): Append instructions before
8537 return_label to prev_bb.
8539 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8541 * config/rs6000/constraints.md: Document reserved use of "wc".
8543 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8546 * ipa.c (function_and_variable_visibility): When dissolving comdat
8547 group, also set all symbols to local.
8549 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
8554 2013-12-14 Jan Hubicka <jh@suse.cz>
8556 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
8558 2014-03-02 Jon Beniston <jon@beniston.com>
8564 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
8565 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
8566 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
8567 (simple_return, *simple_return): New patterns
8568 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
8569 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
8571 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
8573 * dwarf2out.c (gen_subprogram_die): Tidy.
8575 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
8578 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
8579 (*mov_t_msb_neg_negc): ... this new insn.
8581 2014-02-28 Jason Merrill <jason@redhat.com>
8584 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
8587 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
8590 * dwarf2out.c (decltype_auto_die): New static.
8591 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
8592 (gen_type_die_with_usage): Handle 'decltype(auto)'.
8593 (is_cxx_auto): Likewise.
8595 2014-02-28 Ian Bolton <ian.bolton@arm.com>
8597 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
8598 we are not using general regs only.
8600 2014-02-28 Richard Biener <rguenther@suse.de>
8603 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
8604 previous fix and only allow to remove trivial pre-headers
8605 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
8606 (remove_forwarder_block): Properly update the latch of a loop.
8608 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8611 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
8612 (cselib_preserved_hash_table): New.
8613 (preserve_constants_and_equivs): Move preserved vals to it.
8614 (cselib_find_slot): Look it up first.
8615 (cselib_init): Initialize it.
8616 (cselib_finish): Release it.
8617 (dump_cselib_table): Dump it.
8619 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8622 * cselib.c (remove_useless_values): Skip to avoid quadratic
8623 behavior if the condition moved from...
8624 (cselib_process_insn): ... here holds.
8626 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
8629 * var-tracking.c (vt_initialize): Apply the same condition to
8630 preserve the CFA base value.
8632 2014-02-28 Joey Ye <joey.ye@arm.com>
8635 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
8636 if reload in progress or completed.
8638 2014-02-28 Tobias Burnus <burnus@net-b.de>
8641 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
8644 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
8646 * doc/tm.texi.in (Condition Code Status): Update documention for
8647 relative locations of cc0-setter and cc0-user.
8649 2014-02-27 Jeff Law <law@redhat.com>
8651 PR rtl-optimization/52714
8652 * combine.c (try_combine): When splitting an unrecognized PARALLEL
8653 into two independent simple sets, if I3 is a jump, ensure the
8654 pattern we place into I3 is a (set (pc) ...).
8656 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
8657 Jeff Law <law@redhat.com>
8659 PR rtl-optimization/49847
8660 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
8661 are in different blocks.
8662 * doc/tm.texi (Condition Code Status): Update documention for
8663 relative locations of cc0-setter and cc0-user.
8665 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
8668 * lra.c (lra_emit_add): Check SUBREG too.
8670 2014-02-27 Andreas Schwab <schwab@suse.de>
8672 * config/m68k/m68k.c (m68k_option_override): Disable
8673 -flive-range-shrinkage for classic m68k.
8674 (m68k_override_options_after_change): Likewise.
8676 2014-02-27 Marek Polacek <polacek@redhat.com>
8679 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
8680 -Wmaybe-uninitialized.
8682 2014-02-27 Alan Modra <amodra@gmail.com>
8685 * reload1.c (emit_input_reload_insns): When reload_override_in,
8686 set old to rl->in_reg when rl->in_reg is a subreg.
8688 2014-02-26 Richard Biener <rguenther@suse.de>
8691 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
8693 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8695 * common/config/i386/predicates.md (const1256_operand): Remove.
8696 (const2356_operand): New.
8697 (const_1_to_2_operand): Remove.
8698 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
8699 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
8700 (*avx512pf_gatherpf<mode>sf): Ditto.
8701 (avx512pf_gatherpf<mode>df): Ditto.
8702 (*avx512pf_gatherpf<mode>df_mask): Ditto.
8703 (*avx512pf_gatherpf<mode>df): Ditto.
8704 (avx512pf_scatterpf<mode>sf): Ditto.
8705 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
8706 (*avx512pf_scatterpf<mode>sf): Ditto.
8707 (avx512pf_scatterpf<mode>df): Ditto.
8708 (*avx512pf_scatterpf<mode>df_mask): Ditto.
8709 (*avx512pf_scatterpf<mode>df): Ditto.
8710 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
8712 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
8714 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
8715 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
8716 (_mm512_mask_testn_epi64_mask): Move to ...
8717 * config/i386/avx512cdintrin.h: Here.
8718 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
8719 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
8720 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
8721 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
8722 TARGET_AVX512F from TARGET_AVX512CD.
8724 2014-02-26 Richard Biener <rguenther@suse.de>
8727 * ipa.c (walk_polymorphic_call_targets): Properly guard
8728 call to inline_update_overall_summary.
8730 2014-02-26 Bin Cheng <bin.cheng@arm.com>
8733 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
8734 and latches only if requested. Fix latch if it is removed.
8735 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
8736 LOOPS_HAVE_PREHEADERS.
8738 2014-02-25 Andrew Pinski <apinski@cavium.com>
8740 * builtins.c (expand_builtin_thread_pointer): Create a new target
8741 when the target is NULL.
8743 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
8745 PR rtl-optimization/60317
8746 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8747 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
8748 * lra-assigns.c: Include params.h.
8749 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
8750 other reload pseudos considerations.
8752 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8754 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
8755 to use canonical form for nor<mode>3.
8757 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8760 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
8763 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
8765 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
8766 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
8767 (ix86_handle_option): Handle OPT_mprefetchwt1.
8768 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
8769 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8771 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
8772 OPTION_MASK_ISA_PREFETCHWT1.
8773 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
8774 (PTA_PREFETCHWT1): New.
8775 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
8776 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
8777 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
8778 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
8779 (*prefetch_avx512pf_<mode>_: Change into ...
8780 (*prefetch_prefetchwt1_<mode>: This.
8781 * config/i386/i386.opt (mprefetchwt1): New.
8782 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
8783 (_mm_prefetch): Handle intent to write.
8784 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
8786 2014-02-25 Richard Biener <rguenther@suse.de>
8789 * emit-rtl.c (mem_attrs_htab): Remove.
8790 (mem_attrs_htab_hash): Likewise.
8791 (mem_attrs_htab_eq): Likewise.
8792 (set_mem_attrs): Always allocate new mem-attrs when something changed.
8793 (init_emit_once): Do not allocate mem_attrs_htab.
8795 2014-02-25 Richard Biener <rguenther@suse.de>
8798 * lto-opts.c (lto_write_options): Output non-explicit conservative
8799 -fwrapv, -fno-trapv and -fno-strict-overflow.
8800 * lto-wrapper.c (merge_and_complain): Handle merging those options.
8801 (run_gcc): And pass them through.
8803 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8805 * sel-sched.c (calculate_new_fences): New parameter ptime.
8806 Calculate it as a maximum over all fence cycles.
8807 (sel_sched_region_2): Adjust the call to calculate_new_fences.
8808 Print the final schedule timing when sched_verbose.
8810 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
8812 PR rtl-optimization/60292
8813 * sel-sched.c (fill_vec_av_set): Do not reset target availability
8814 bit fot the fence instruction.
8816 2014-02-24 Alangi Derick <alangiderick@gmail.com>
8818 * calls.h: Fix typo in comment.
8820 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
8822 * config/pa/pa.c (pa_output_move_double): Don't valididate when
8823 adjusting offsetable addresses.
8825 2014-02-24 Guozhi Wei <carrot@google.com>
8827 * sparseset.h (sparseset_pop): Fix the wrong index.
8829 2014-02-24 Walter Lee <walt@tilera.com>
8831 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
8832 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
8834 * common/config/tilegx/tilegx-common.c
8835 (TARGET_DEFAULT_TARGET_FLAGS): Define.
8836 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
8838 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
8839 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
8840 (tilegx_gimplify_va_arg_expr): Handle big endian.
8841 (tilegx_expand_unaligned_load): Ditto.
8842 (tilegx_expand_unaligned_store): Ditto.
8843 (TARGET_RETURN_IN_MSB): New.
8844 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
8845 (TARGET_ENDIAN_DEFAULT): New.
8846 (TARGET_BIG_ENDIAN): Handle big endian.
8847 (BYTES_BIG_ENDIAN): Ditto.
8848 (WORDS_BIG_ENDIAN): Ditto.
8849 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
8852 * config/tilegx/tilegx.md (extv): Handle big endian.
8854 (insn_st<n>): Ditto.
8855 (insn_st<n>_add<bitsuffix>): Ditto.
8856 (insn_stnt<n>): Ditto.
8857 (insn_stnt<n>_add<bitsuffix>):Ditto.
8858 (vec_interleave_highv8qi): Handle big endian.
8859 (vec_interleave_highv8qi_be): New.
8860 (vec_interleave_highv8qi_le): New.
8861 (insn_v1int_h): Handle big endian.
8862 (vec_interleave_lowv8qi): Handle big endian.
8863 (vec_interleave_lowv8qi_be): New.
8864 (vec_interleave_lowv8qi_le): New.
8865 (insn_v1int_l): Handle big endian.
8866 (vec_interleave_highv4hi): Handle big endian.
8867 (vec_interleave_highv4hi_be): New.
8868 (vec_interleave_highv4hi_le): New.
8869 (insn_v2int_h): Handle big endian.
8870 (vec_interleave_lowv4hi): Handle big endian.
8871 (vec_interleave_lowv4hi_be): New.
8872 (vec_interleave_lowv4hi_le): New.
8873 (insn_v2int_l): Handle big endian.
8874 (vec_interleave_highv2si): Handle big endian.
8875 (vec_interleave_highv2si_be): New.
8876 (vec_interleave_highv2si_le): New.
8877 (insn_v4int_h): Handle big endian.
8878 (vec_interleave_lowv2si): Handle big endian.
8879 (vec_interleave_lowv2si_be): New.
8880 (vec_interleave_lowv2si_le): New.
8881 (insn_v4int_l): Handle big endian.
8882 * config/tilegx/tilegx.opt (mbig-endian): New option.
8883 (mlittle-endian): New option.
8884 * doc/install.texi: Document tilegxbe-linux.
8885 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
8887 2014-02-24 Martin Jambor <mjambor@suse.cz>
8890 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
8891 there are no parameter descriptors.
8893 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
8895 PR rtl-optimization/60268
8896 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
8897 initialization to ...
8898 (sched_rgn_init): ... here.
8899 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
8901 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8903 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
8906 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8908 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
8911 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8913 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
8914 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
8916 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
8918 * config/microblaze/predicates.md: Add cmp_op predicate.
8919 * config/microblaze/microblaze.md: Add branch_compare instruction
8920 which uses cmp_op predicate and emits cmp insn before branch.
8921 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
8922 to microblaze_expand_conditional_branch and consolidate logic.
8923 (microblaze_expand_conditional_branch): emit branch_compare
8924 insn instead of handling cmp op separate from branch insn.
8926 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8928 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
8931 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8933 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
8934 define_insn with define_expand and new define_insn
8935 *altivec_lve<VI_char>x_internal.
8936 (altivec_stve<VI_char>x): Replace define_insn with define_expand
8937 and new define_insn *altivec_stve<VI_char>x_internal.
8938 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
8940 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
8942 (altivec_expand_stvex_be): New function.
8944 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
8946 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
8947 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8948 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
8949 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
8951 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
8954 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
8955 instead of emit_move_insn.
8957 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8959 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
8961 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
8962 gen_altivec_vsumsws.
8964 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8966 * config/rs6000/altivec.md (altivec_lvxl): Rename as
8967 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
8968 (altivec_lvxl_<mode>): New define_expand incorporating
8969 -maltivec=be semantics where needed.
8970 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
8971 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
8972 semantics where needed.
8973 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
8974 (altivec_stvx_<mode>): New define_expand incorporating
8975 -maltivec=be semantics where needed.
8976 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
8977 VM2 iterator instead of V4SI.
8978 (altivec_stvxl_<mode>): New define_expand incorporating
8979 -maltivec=be semantics where needed.
8980 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
8981 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
8982 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
8983 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
8984 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
8985 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
8986 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
8987 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
8988 ALTIVEC_BUILTIN_STVXL.
8989 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
8990 (altivec_expand_stvx_be): Likewise.
8991 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
8992 (altivec_expand_lvx_be): Likewise.
8993 (altivec_expand_stvx_be): Likewise.
8994 (altivec_expand_builtin): Add cases for
8995 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
8996 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
8997 (altivec_init_builtins): Add definitions for
8998 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
8999 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
9001 2014-02-21 Catherine Moore <clm@codesourcery.com>
9003 * doc/invoke.texi (mvirt, mno-virt): Document.
9004 * config/mips/mips.opt (mvirt): New option.
9005 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
9007 2014-02-21 Richard Biener <rguenther@suse.de>
9009 PR tree-optimization/60276
9010 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
9011 (STMT_VINFO_MIN_NEG_DIST): New macro.
9012 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
9013 STMT_VINFO_MIN_NEG_DIST.
9014 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
9015 made for negative dependence distances still hold.
9017 2014-02-21 Richard Biener <rguenther@suse.de>
9020 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
9021 DECL_INITIAL for globals not in the current function context.
9023 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9025 PR tree-optimization/56490
9026 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
9027 * tree-ssa-uninit.c: Include params.h.
9028 (compute_control_dep_chain): Add num_calls argument, return false
9029 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
9030 num_calls to recursive call.
9031 (find_predicates): Change dep_chain into normal array,
9032 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
9033 variable and adjust compute_control_dep_chain caller.
9034 (find_def_preds): Likewise.
9036 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
9038 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
9039 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
9041 2014-02-21 Nick Clifton <nickc@redhat.com>
9043 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
9044 (pushhi1): Likewise.
9045 (popqi1): Add mode to pre_dec.
9048 2014-02-21 Jakub Jelinek <jakub@redhat.com>
9050 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
9051 mode for mask of V8SFmode permutation.
9053 2014-02-20 Richard Henderson <rth@redhat.com>
9056 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
9057 a new pseudo for OLDVAL.
9059 2014-02-20 Jakub Jelinek <jakub@redhat.com>
9062 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
9063 gen_reg_rtx if d->testing_p.
9064 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
9065 if d->testing_p and we will certainly return true.
9066 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
9069 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
9071 * emit-rtl.c (gen_reg_rtx): Assert that
9072 crtl->emit.regno_pointer_align_length is non-zero.
9074 2014-02-20 Richard Henderson <rth@redhat.com>
9077 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
9078 on failure the store back into EXPECT.
9080 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
9081 Sandra Loosemore <sandra@codesourcery.com>
9083 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
9084 * config/nios2/nios2.c (nios2_function_profiler): Add
9085 -fPIC (flag_pic == 2) support.
9086 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
9087 (nios2_large_offset_p): New function.
9088 (nios2_unspec_reloc_p): Move up position, update to use
9089 nios2_large_offset_p.
9090 (nios2_unspec_address): Remove function.
9091 (nios2_unspec_offset): New function.
9092 (nios2_large_got_address): New function.
9093 (nios2_got_address): Add large offset support.
9094 (nios2_legitimize_tls_address): Update usage of removed and new
9096 (nios2_symbol_binds_local_p): New function.
9097 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
9098 (nios2_legitimize_address): Update to use nios2_large_offset_p.
9099 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
9100 (nios2_print_operand): Merge H/L processing, add hiadj/lo
9101 processing for (const (unspec ...)).
9102 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
9104 2014-02-20 Richard Biener <rguenther@suse.de>
9106 * tree-cfg.c (replace_uses_by): Mark altered BBs before
9107 doing the substitution.
9108 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
9110 2014-02-20 Martin Jambor <mjambor@suse.cz>
9113 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
9114 info when checking whether lattices are bottom.
9116 2014-02-20 Richard Biener <rguenther@suse.de>
9119 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
9122 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
9125 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
9126 parameter specifying the scaling.
9127 (inline_call): Update.
9128 (want_inline_recursively): Guard division by zero.
9129 (recursive_inlining): Update.
9130 * ipa-inline.h (clone_inlined_nodes): Update.
9132 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9135 * config/i386/i386.c (classify_argument): Pass structures of size
9136 64 bytes or less in register.
9138 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
9139 Kirill Yukhin <kirill.yukhin@intel.com>
9141 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
9142 (_mm_rcp28_round_ss): Ditto.
9143 (_mm_rsqrt28_round_sd): Ditto.
9144 (_mm_rsqrt28_round_ss): Ditto.
9145 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
9146 (_mm_rcp14_round_ss): Ditto.
9147 (_mm_rsqrt14_round_sd): Ditto.
9148 (_mm_rsqrt14_round_ss): Ditto.
9149 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
9150 the first input operand, get rid of match_dup.
9151 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
9153 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
9155 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
9156 operand as the first input operand, set type attribute.
9157 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
9159 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
9160 operand as the first input operand, set type attribute.
9162 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9164 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
9167 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
9170 * config/i386/i386.c (construct_container): Remove TFmode check
9171 for X86_64_INTEGER_CLASS.
9173 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
9176 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
9177 only when -Wpsabi is enabled.
9179 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
9182 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
9183 passing arrays in registers are the same as for structs, so remove the
9184 special case for them.
9186 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
9188 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
9189 destination type, extract only the valid bits if the source type is not
9190 integral and has a different mode.
9192 2014-02-19 Richard Biener <rguenther@suse.de>
9195 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
9198 2014-02-19 Richard Biener <rguenther@suse.de>
9201 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
9202 (ipa_modify_call_arguments): Emit an argument load explicitely and
9203 preserve virtual SSA form there and for the replacement call.
9204 Do not update SSA form nor free dominance info.
9206 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9208 * ipa.c (function_and_variable_visibility): Also clear WEAK
9209 flag when disolving COMDAT_GROUP.
9211 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9213 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
9214 * ipa-prop.c (ipa_set_jf_known_type): Return early when
9216 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
9217 do more sanity checks.
9218 (detect_type_change): Return true when giving up early.
9219 (compute_complex_assign_jump_func): Fix type parameter of
9220 ipa_set_ancestor_jf.
9221 (compute_complex_ancestor_jump_func): Likewise.
9222 (update_jump_functions_after_inlining): Fix updating of
9224 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
9226 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
9228 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
9229 inline clones when edge disappears.
9231 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
9234 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
9235 Split 64-bit moves into 2 patterns. Do not allow the use of
9236 direct move for TDmode in little endian, since the decimal value
9237 has little endian bytes within a word, but the 64-bit pieces are
9238 ordered in a big endian fashion, and normal subreg's of TDmode are
9240 (mov<mode>_64bit_dm): Likewise.
9241 (movtd_64bit_nodm): Likewise.
9243 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9245 PR tree-optimization/60174
9246 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
9247 statement of an SSA_NAME that occurs in an abnormal PHI node.
9249 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9252 * final.c (SEEN_BB): Remove.
9253 (SEEN_NOTE, SEEN_EMITTED): Renumber.
9254 (final_scan_insn): Don't force_source_line on second
9255 NOTE_INSN_BASIC_BLOCK.
9257 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
9260 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
9261 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
9262 (type_natural_mode): Warn ABI change when %zmm register is not
9263 available for AVX512F vector value passing.
9265 2014-02-18 Kai Tietz <ktietz@redhat.com>
9268 * config/i386/i386.c (ix86_expand_prologue): Use value in
9269 rax register as displacement when restoring %r10 or %rax.
9270 Fix wrong offset when restoring both registers.
9272 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
9274 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
9275 assertion with conditional return.
9277 2014-02-18 Jakub Jelinek <jakub@redhat.com>
9278 Uros Bizjak <ubizjak@gmail.com>
9281 * config/i386/driver-i386.c (host_detect_local_cpu): If
9282 YMM state is not saved by the OS, also clear has_f16c. Move
9283 CPUID 0x80000001 handling before YMM state saving checking.
9285 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
9287 PR rtl-optimization/58960
9288 * haifa-sched.c (alloc_global_sched_pressure_data): New,
9289 factored out from ...
9290 (sched_init): ... here.
9291 (free_global_sched_pressure_data): New, factored out from ...
9292 (sched_finish): ... here.
9293 * sched-int.h (free_global_sched_pressure_data): Declare.
9294 * sched-rgn.c (nr_regions_initial): New static global.
9295 (haifa_find_rgns): Initialize it.
9296 (schedule_region): Disable sched-pressure for the newly
9299 2014-02-17 Richard Biener <rguenther@suse.de>
9301 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
9302 release SSA defs of pattern stmts.
9304 2014-02-17 Richard Biener <rguenther@suse.de>
9306 * tree-inline.c (expand_call_inline): Release the virtual
9307 operand defined by the call we are about to inline.
9309 2014-02-17 Richard Biener <rguenther@suse.de>
9311 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
9313 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
9314 Ilya Tocar <ilya.tocar@intel.com>
9316 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
9317 arguments order in builtin.
9318 (_mm512_permutexvar_epi64): Ditto.
9319 (_mm512_mask_permutexvar_epi64): Ditto
9320 (_mm512_maskz_permutexvar_epi32): Ditto
9321 (_mm512_permutexvar_epi32): Ditto
9322 (_mm512_mask_permutexvar_epi32): Ditto
9324 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9326 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
9327 (p8_vmrgow): Likewise.
9329 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9331 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
9334 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
9337 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
9338 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
9339 into 64-bit and 32-bit moves. On 64-bit moves, add support for
9340 using direct move instructions on ISA 2.07. Also adjust
9341 instruction length for 64-bit.
9342 (mov<mode>_64bit, TFmode/TDmode): Likewise.
9343 (mov<mode>_32bit, TFmode/TDmode): Likewise.
9345 2014-02-15 Alan Modra <amodra@gmail.com>
9349 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
9350 find_replacement on parts of insn rtl that might be reloaded.
9352 2014-02-15 Richard Biener <rguenther@suse.de>
9354 PR tree-optimization/60183
9355 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
9356 (tree_ssa_phiprop): Calculate and free post-dominators.
9358 2014-02-14 Jeff Law <law@redhat.com>
9360 PR rtl-optimization/60131
9361 * ree.c (get_extended_src_reg): New function.
9362 (combine_reaching_defs): Use it rather than assuming location of REG.
9363 (find_and_remove_re): Verify first operand of extension is
9364 a REG before adding the insns to the copy list.
9366 2014-02-14 Roland McGrath <mcgrathr@google.com>
9368 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
9369 * configure: Regenerated.
9370 * config.in: Regenerated.
9371 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
9372 instead of ASM_SHORT.
9374 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
9375 Richard Earnshaw <rearnsha@arm.com>
9377 PR rtl-optimization/59535
9378 * lra-constraints.c (process_alt_operands): Encourage alternative
9379 when unassigned pseudo class is superset of the alternative class.
9380 (inherit_reload_reg): Don't inherit when optimizing for code size.
9381 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
9382 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
9383 modes not less than 4 for Thumb1.
9385 2014-02-14 Kyle McMartin <kyle@redhat.com>
9388 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
9390 2014-02-14 Richard Biener <rguenther@suse.de>
9392 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
9393 (get_frame_arg): Drop the assert with langhook types_compatible_p.
9394 Do not strip INDIRECT_REFs.
9396 2014-02-14 Richard Biener <rguenther@suse.de>
9399 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
9400 DECL_FUNCTION_SPECIFIC_TARGET.
9401 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
9402 * tree-streamer-out.c (pack_ts_target_option): Remove.
9403 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
9404 (write_ts_function_decl_tree_pointers): Do not stream
9405 DECL_FUNCTION_SPECIFIC_TARGET.
9406 * tree-streamer-in.c (unpack_ts_target_option): Remove.
9407 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
9408 (lto_input_ts_function_decl_tree_pointers): Do not stream
9409 DECL_FUNCTION_SPECIFIC_TARGET.
9411 2014-02-14 Jakub Jelinek <jakub@redhat.com>
9413 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
9414 (get_initial_def_for_induction, vectorizable_induction): Ignore
9415 debug stmts when looking for exit_phi.
9416 (vectorizable_live_operation): Fix up condition.
9418 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9420 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
9421 nreverse() because it changes the content of original tree list.
9423 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9425 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
9426 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
9428 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
9430 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
9431 GNU coding standards.
9433 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9436 * dwarf2out.c (gen_subprogram_die): Don't call
9437 add_calling_convention_attribute if subr_die is old_die.
9439 2014-02-13 Sharad Singhai <singhai@google.com>
9441 * doc/optinfo.texi: Fix order of nodes.
9443 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
9445 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
9446 operands[2], not operands[3].
9448 2014-02-13 Richard Biener <rguenther@suse.de>
9451 * doc/install.texi (ISL): Update recommended version to 0.12.2,
9452 mention the possibility of an in-tree build.
9453 (CLooG): Update recommended version to 0.18.1, mention the
9454 possibility of an in-tree build and clarify that the ISL
9455 bundled with CLooG does not work.
9457 2014-02-13 Jakub Jelinek <jakub@redhat.com>
9460 * expr.c (compress_float_constant): If x is a hard register,
9461 extend into a pseudo and then move to x.
9463 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
9465 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
9466 caused by bad second argument to warning_at() with -mhotpatch and
9467 nested functions (e.g. with gfortran).
9469 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
9471 * opts.c (option_name): Remove "enabled by default" rider.
9473 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
9475 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
9477 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
9478 Uros Bizjak <ubizjak@gmail.com>
9481 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
9482 * configure: Regenerated.
9484 2014-02-12 Richard Biener <rguenther@suse.de>
9486 * vec.c (vec_prefix::calculate_allocation): Move as
9487 inline variant to vec.h.
9488 (vec_prefix::calculate_allocation_1): New out-of-line version.
9489 * vec.h (vec_prefix::calculate_allocation_1): Declare.
9490 (vec_prefix::m_has_auto_buf): Rename to ...
9491 (vec_prefix::m_using_auto_storage): ... this.
9492 (vec_prefix::calculate_allocation): Inline the easy cases
9493 and dispatch to calculate_allocation_1 which doesn't need the
9495 (va_heap::reserve): Use gcc_checking_assert.
9496 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
9497 m_using_auto_storage.
9498 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
9500 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
9501 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
9502 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
9504 2014-02-12 Richard Biener <rguenther@suse.de>
9506 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
9507 when we found a dependence.
9509 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
9511 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
9513 (maybe_fold_stmt): ... into this new function.
9514 * omp-low.c (lower_omp): Update comment.
9516 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
9519 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
9522 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
9524 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
9525 identifiers in comments.
9526 (cortexa53_extra_costs): Likewise.
9527 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
9528 (cortexa7_extra_costs): Likewise.
9529 (cortexa12_extra_costs): Likewise.
9530 (cortexa15_extra_costs): Likewise.
9531 (v7m_extra_costs): Likewise.
9533 2014-02-12 Richard Biener <rguenther@suse.de>
9536 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
9537 of posix_memalign being successful.
9538 (lower_stmt): Restrict lowering of posix_memalign to when
9539 -ftree-bit-ccp is enabled.
9541 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
9543 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
9545 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
9547 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
9549 PR rtl-optimization/60116
9550 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
9551 other_insn once the combination has been validated.
9553 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
9556 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
9558 * ipa-devirt.c: Include demangle.h
9559 (odr_violation_reported): New static variable.
9560 (add_type_duplicate): Update odr_violations.
9561 (maybe_record_node): Add completep parameter; update it.
9562 (record_target_from_binfo): Add COMPLETEP parameter;
9563 update it as needed.
9564 (possible_polymorphic_call_targets_1): Likewise.
9565 (struct polymorphic_call_target_d): Add nonconstruction_targets;
9566 rename FINAL to COMPLETE.
9567 (record_targets_from_bases): Sanity check we found the binfo;
9568 fix COMPLETEP updating.
9569 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
9570 parameter, fix computing of COMPLETEP.
9571 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
9572 at LTO time do demangling.
9573 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
9574 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
9576 (gimple_get_virt_method_for_binfo): Likewise.
9577 * gimple-fold.h (gimple_get_virt_method_for_binfo,
9578 gimple_get_virt_method_for_vtable): Update prototypes.
9580 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
9583 * genautomata.c (add_presence_absence): Fix typo with
9584 {final_}presence_list.
9586 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9589 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
9590 for VSX/Altivec vectors that land in GPR registers.
9592 2014-02-11 Richard Henderson <rth@redhat.com>
9593 Jakub Jelinek <jakub@redhat.com>
9596 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
9597 around drhs if type conversion to lacc->type is not useless.
9599 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9601 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
9603 (cortex-a57.cortex-a53): Likewise.
9604 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
9606 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9608 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
9611 2014-02-11 Renlin Li <Renlin.Li@arm.com>
9613 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
9614 add_options_for_arm_vfp3.
9616 2014-02-11 Jeff Law <law@redhat.com>
9619 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
9620 object with an undesirable mode.
9622 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9625 * config/i386/sol2-9.h: New file.
9626 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
9627 *-*-solaris2.9*): Use it.
9629 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9631 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
9632 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
9634 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
9636 * config/microblaze/microblaze.c: Extend mcpu version format
9638 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9640 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
9642 2014-02-10 Richard Henderson <rth@redhat.com>
9645 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
9646 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
9647 ms-abi vs -mno-accumulate-outgoing-args.
9648 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
9649 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
9652 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
9655 * cfgexpand.c (expand_asm_operands): Attach source location to
9656 ASM_INPUT rtx objects.
9657 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
9659 2014-02-10 Nick Clifton <nickc@redhat.com>
9661 * config/mn10300/mn10300.c (popcount): New function.
9662 (mn10300_expand_prologue): Include saved registers in stack usage
9665 2014-02-10 Jeff Law <law@redhat.com>
9668 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
9669 when changing the SET_DEST of a prior insn to avoid an input reload.
9671 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
9673 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
9674 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
9675 -mcall-openbsd, or -mcall-linux.
9676 (CC1_ENDIAN_BIG_SPEC): Remove.
9677 (CC1_ENDIAN_LITTLE_SPEC): Remove.
9678 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9679 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
9680 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
9681 and %cc1_endian_default.
9682 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
9684 2014-02-10 Richard Biener <rguenther@suse.de>
9686 PR tree-optimization/60115
9687 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
9688 MEM_REF handling. Properly verify that the accesses are not
9689 out of the objects bound.
9691 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9693 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
9696 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
9698 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
9699 proper constants and fix formatting.
9700 (possible_polymorphic_call_targets): Fix formatting.
9702 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
9703 Ilya Tocar <ilya.tocar@intel.com>
9705 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
9706 (_mm512_loadu_epi32): Renamed into...
9707 (_mm512_loadu_si512): This.
9708 (_mm512_storeu_epi32): Renamed into...
9709 (_mm512_storeu_si512): This.
9710 (_mm512_maskz_ceil_ps): Removed.
9711 (_mm512_maskz_ceil_pd): Ditto.
9712 (_mm512_maskz_floor_ps): Ditto.
9713 (_mm512_maskz_floor_pd): Ditto.
9714 (_mm512_floor_round_ps): Ditto.
9715 (_mm512_floor_round_pd): Ditto.
9716 (_mm512_ceil_round_ps): Ditto.
9717 (_mm512_ceil_round_pd): Ditto.
9718 (_mm512_mask_floor_round_ps): Ditto.
9719 (_mm512_mask_floor_round_pd): Ditto.
9720 (_mm512_mask_ceil_round_ps): Ditto.
9721 (_mm512_mask_ceil_round_pd): Ditto.
9722 (_mm512_maskz_floor_round_ps): Ditto.
9723 (_mm512_maskz_floor_round_pd): Ditto.
9724 (_mm512_maskz_ceil_round_ps): Ditto.
9725 (_mm512_maskz_ceil_round_pd): Ditto.
9726 (_mm512_expand_pd): Ditto.
9727 (_mm512_expand_ps): Ditto.
9728 * config/i386/i386.c (ix86_builtins): Remove
9729 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
9730 (bdesc_args): Ditto.
9731 * config/i386/predicates.md (const1256_operand): New.
9732 (const_1_to_2_operand): Ditto.
9733 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
9734 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
9735 (*avx512pf_gatherpf<mode>sf): Ditto.
9736 (avx512pf_gatherpf<mode>df): Ditto.
9737 (*avx512pf_gatherpf<mode>df_mask): Ditto.
9738 (*avx512pf_gatherpf<mode>df): Ditto.
9739 (avx512pf_scatterpf<mode>sf): Ditto.
9740 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
9741 (*avx512pf_scatterpf<mode>sf): Ditto.
9742 (avx512pf_scatterpf<mode>df): Ditto.
9743 (*avx512pf_scatterpf<mode>df_mask): Ditto.
9744 (*avx512pf_scatterpf<mode>df): Ditto.
9745 (avx512f_expand<mode>): Removed.
9746 (<shift_insn><mode>3<mask_name>): Change predicate type.
9748 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9750 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
9751 not at the end of datarefs vector use ordered_remove to avoid
9752 reordering datarefs vector.
9755 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
9756 mark local addressable non-static vars as GOVD_PRIVATE
9757 instead of GOVD_LOCAL.
9758 * omp-low.c (lower_omp_for): Move gimple_bind_vars
9759 and BLOCK_VARS of gimple_bind_block to new_stmt rather
9763 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
9764 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
9765 assume_aligned or alloc_align attributes.
9766 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
9767 arguments. Handle also assume_aligned and alloc_align attributes.
9768 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
9769 calls to functions with assume_aligned or alloc_align attributes.
9770 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
9772 2014-02-08 Terry Guo <terry.guo@arm.com>
9774 * doc/invoke.texi: Document ARM -march=armv7e-m.
9776 2014-02-08 Jakub Jelinek <jakub@redhat.com>
9778 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
9779 flag on __cilkrts_rethrow builtin.
9782 * ipa-cp.c (determine_versionability): Fail at -O0
9783 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
9784 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
9787 2014-02-04 Jakub Jelinek <jakub@redhat.com>
9790 * tree-inline.c (copy_forbidden): Fail for
9791 __attribute__((optimize (0))) functions.
9793 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9795 * varpool.c: Include pointer-set.h.
9796 (varpool_remove_unreferenced_decls): Variables in other partitions
9797 will not be output; be however careful to not lose information
9800 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
9802 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
9803 lookup in the vtable constructor.
9805 2014-02-07 Jeff Law <law@redhat.com>
9808 * config/m68k/m68k.md (ashldi_extsi): Turn into a
9809 define_insn_and_split.
9811 * ipa-inline.c (inline_small_functions): Fix typos.
9813 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9815 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
9816 (s390_can_use_return_insn): Declare.
9817 * config/s390/s390.h (EPILOGUE_USES): Define.
9818 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
9820 (s390_chunkify_start): Handle return JUMP_LABELs.
9821 (s390_early_mach): Emit a main_pool instruction on the entry edge.
9822 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
9823 (s390_can_use_return_insn): New functions.
9824 (s390_fix_long_loop_prediction): Handle conditional returns.
9825 (TARGET_SET_UP_BY_PROLOGUE): Define.
9826 * config/s390/s390.md (ANY_RETURN): New code iterator.
9827 (*creturn, *csimple_return, return, simple_return): New patterns.
9829 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9831 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
9832 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
9833 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
9834 REG_CFA_RESTORE list when deciding not to restore a register.
9836 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9838 * config/s390/s390.c: Include tree-pass.h and context.h.
9839 (s390_early_mach): New function, split out from...
9840 (s390_emit_prologue): ...here.
9841 (pass_data_s390_early_mach): New pass structure.
9842 (pass_s390_early_mach): New class.
9843 (s390_option_override): Create and register early_mach pass.
9844 Move to end of file.
9846 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
9848 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
9849 to match for the exit block.
9851 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9853 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
9854 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
9855 Reject misaligned operands.
9857 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9859 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
9861 2014-02-07 Richard Biener <rguenther@suse.de>
9864 * gimple-low.c (lower_builtin_posix_memalign): New function.
9865 (lower_stmt): Call it to lower posix_memalign in a way
9866 to make alignment info accessible.
9868 2014-02-07 Jakub Jelinek <jakub@redhat.com>
9871 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
9872 __builtin_setjmp_receiver.
9874 2014-02-07 Richard Biener <rguenther@suse.de>
9877 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
9878 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
9879 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
9880 Handle BUILT_IN_POSIX_MEMALIGN.
9881 (find_func_clobbers): Likewise.
9882 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
9883 (call_may_clobber_ref_p_1): Likewise.
9885 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9888 * ipa-devirt.c (record_target_from_binfo): Remove overactive
9891 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9894 * lto-cgraph.c (lto_output_node): Use
9895 symtab_get_symbol_partitioning_class.
9896 (lto_output_varpool_node): likewise.
9897 (symtab_get_symbol_partitioning_class): Move here from
9899 * cgraph.h (symbol_partitioning_class): Likewise.
9900 (symtab_get_symbol_partitioning_class): Declare.
9902 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9904 * ggc.h (ggc_internal_cleared_alloc): New macro.
9905 * vec.h (vec_safe_copy): Handle memory stats.
9906 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
9907 * target-globals.c (save_target_globals): Likewise.
9909 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
9912 * expr.c (emit_move_resolve_push): Export; be bit more selective
9913 on when to clear alias set.
9914 * expr.h (emit_move_resolve_push): Declare.
9915 * function.h (struct function): Add tail_call_marked.
9916 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
9917 * config/i386/i386-protos.h (ix86_expand_push): Remove.
9918 * config/i386/i386.md (TImode move expander): De not call
9920 (FP push expanders): Preserve memory attributes.
9921 * config/i386/sse.md (push<mode>1): Remove.
9922 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
9923 (ix86_expand_push): Remove.
9924 * config/i386/mmx.md (push<mode>1): Remove.
9926 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9928 PR rtl-optimization/60030
9929 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
9930 lopart with paradoxical subreg before shifting it up by hprec.
9932 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9934 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
9935 Remove extra newline at end of file.
9936 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
9937 (arm_issue_rate): Handle cortexa57.
9938 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
9939 (cortex-a57.cortex-a53): Likewise.
9941 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9944 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
9945 don't record in REG_FRAME_RELATED_EXPR registers not set in that
9947 (arm_expand_prologue): Adjust all callers.
9948 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
9949 info, registers also at the lowest numbered registers side. Use
9950 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
9954 * var-tracking.c (adjust_mems): Before adding a SET to
9955 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
9957 2014-02-06 Alan Modra <amodra@gmail.com>
9960 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
9961 change SDmode to DDmode when lra_in_progress.
9963 2014-02-06 Jakub Jelinek <jakub@redhat.com>
9966 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
9967 free_data_ref on the dr first, and before goto again also set dr
9968 to the next dr. For simd_lane_access, free old datarefs[i] before
9969 overwriting it. For get_vectype_for_scalar_type failure, don't
9970 free_data_ref if simd_lane_access.
9972 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
9975 * tree.h (opts_for_fn): New inline function.
9976 (opt_for_fn): Define.
9977 * config/i386/i386.c (ix86_function_regparm): Use
9978 opt_for_fn (decl, optimize) instead of optimize.
9980 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
9982 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
9983 for SYMBOL_REF in large memory model.
9985 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9987 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
9989 (cortex-a57): Likewise.
9990 (cortex-a57.cortex-a53): Likewise.
9992 2014-02-06 Yury Gribov <y.gribov@samsung.com>
9993 Kugan Vivekanandarajah <kuganv@linaro.org>
9995 * config/arm/arm.c (arm_vector_alignment_reachable): Check
9997 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
9999 2014-02-06 Richard Biener <rguenther@suse.de>
10001 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
10002 set_loop_copy and initialize_original_copy_tables.
10004 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
10006 * config/aarch64/aarch64-simd.md
10007 (aarch64_ashr_simddi): Change QI to SI.
10009 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10010 Jakub Jelinek <jakub@redhat.com>
10012 PR middle-end/60013
10013 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
10016 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10018 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
10019 CODE_FOR_altivec_vpku[hw]um to
10020 CODE_FOR_altivec_vpku[hw]um_direct.
10021 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
10022 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
10023 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
10024 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
10026 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10028 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
10029 generation for -maltivec=be.
10030 (altivec_vsumsws): Simplify redundant test.
10032 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10034 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
10035 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
10036 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
10037 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
10038 gen_altivec_vpkuwum.
10039 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
10041 (altivec_vpks<VI_char>ss): Likewise.
10042 (altivec_vpks<VI_char>us): Likewise.
10043 (altivec_vpku<VI_char>us): Likewise.
10044 (altivec_vpku<VI_char>um): Likewise.
10045 (altivec_vpku<VI_char>um_direct): New (copy of
10046 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
10048 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
10049 target is little endian and -maltivec=be is not specified.
10050 (*altivec_vupkhs<VU_char>_direct): New (copy of
10051 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
10052 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
10053 target is little endian and -maltivec=be is not specified.
10054 (*altivec_vupkls<VU_char>_direct): New (copy of
10055 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
10056 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
10057 little endian and -maltivec=be is not specified.
10058 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
10059 little endian and -maltivec=be is not specified.
10061 2014-02-05 Richard Henderson <rth@redhat.com>
10064 * combine-stack-adj.c: Revert r206943.
10065 * sched-int.h (struct deps_desc): Add last_args_size.
10066 * sched-deps.c (init_deps): Initialize it.
10067 (sched_analyze_insn): Add OUTPUT dependencies between insns that
10068 contain REG_ARGS_SIZE notes.
10070 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
10072 * lto-cgraph.c (asm_nodes_output): Make global.
10073 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
10074 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
10075 (driver_handle_option): Handle OPT_fwpa.
10077 2014-02-05 Jakub Jelinek <jakub@redhat.com>
10080 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
10081 a comment typo and formatting issue. If odr_hash hasn't been
10082 created, return vNULL and set *completep to false.
10084 PR middle-end/57499
10085 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
10086 bb with no successors.
10088 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
10091 * doc/invoke.texi (-march): Clarify documentation for ARM.
10092 (-mtune): Likewise.
10095 2014-02-05 Richard Biener <rguenther@suse.de>
10097 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
10098 when not vectorizing because of too many alias checks.
10099 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
10100 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
10102 2014-02-05 Nick Clifton <nickc@redhat.com>
10104 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
10105 accept extended registers in any mode when compiling for the MN10300.
10107 2014-02-05 Yury Gribov <y.gribov@samsung.com>
10109 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
10110 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
10111 sanitization attributes.
10112 (can_inline_edge_p): Likewise.
10113 (sanitize_attrs_match_for_inline_p): New function.
10115 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10117 * ipa-prop.c (detect_type_change): Shor circuit testing of
10118 type changes on THIS pointer.
10120 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
10123 * config/pa/pa.c (legitimize_tls_address): Return original address
10124 if not passed a SYMBOL_REF rtx.
10125 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
10127 (pa_emit_move_sequence): Simplify TLS source operands.
10128 (pa_legitimate_constant_p): Reject all TLS constants.
10129 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
10130 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
10132 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10134 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
10135 groups when we know they are controlled by LTO.
10136 * varasm.c (default_binds_local_p_1): If object is in other partition,
10137 it will be resolved locally.
10139 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10141 * config/host-linux.c (linux_gt_pch_use_address): Don't
10142 use SSIZE_MAX because it is not always defined.
10144 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
10147 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
10148 threshold for pseudo splitting.
10149 (update_ebb_live_info): Process call argument hard registers and
10150 hard registers from insn definition too.
10151 (max_small_class_regs_num): New constant.
10152 (inherit_in_ebb): Update live hard regs through EBBs. Update
10153 reloads_num only for small register classes. Don't split for
10156 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
10159 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
10162 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
10164 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
10165 visibility is safe.
10167 2014-02-04 Marek Polacek <polacek@redhat.com>
10169 * gdbinit.in (pel): Define.
10171 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
10173 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
10176 2014-02-04 Richard Biener <rguenther@suse.de>
10179 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
10180 in function context local.
10181 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
10182 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
10183 similar to LTO_imported_decl_ref.
10185 2014-02-04 Jakub Jelinek <jakub@redhat.com>
10187 PR tree-optimization/60002
10188 * cgraphclones.c (build_function_decl_skip_args): Clear
10189 DECL_LANG_SPECIFIC.
10191 PR tree-optimization/60023
10192 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
10193 false to gsi_replace.
10194 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
10195 has been in some EH region and vec_stmt could throw, add
10196 vec_stmt into the same EH region.
10197 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
10198 has no lhs, ignore it.
10199 * internal-fn.c (expand_MASK_LOAD): Likewise.
10202 * tree-inline.c (copy_forbidden): Fail for
10203 __attribute__((optimize (0))) functions.
10206 * omp-low.c (simd_clone_struct_copy): If from->inbranch
10207 is set, copy one less argument.
10208 (expand_simd_clones): Don't subtract clone_info->inbranch
10209 from simd_clone_struct_alloc argument.
10211 PR rtl-optimization/57915
10212 * recog.c (simplify_while_replacing): If all unary/binary/relational
10213 operation arguments are constant, attempt to simplify those.
10215 PR middle-end/59261
10216 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
10217 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
10219 2014-02-04 Richard Biener <rguenther@suse.de>
10221 PR tree-optimization/60012
10222 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
10223 TBAA disambiguation to all DDRs.
10225 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
10228 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
10229 (LINK_SPEC): Use it for -shared, -shared-libgcc.
10231 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10234 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
10236 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10238 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
10239 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
10241 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10244 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
10245 to figure out targets of polymorphic calls with known decl.
10246 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10247 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
10248 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
10249 (get_polymorphic_call_info): ... here.
10250 (get_polymorphic_call_info_from_invariant): New function.
10252 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10254 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
10255 lookup via vtable pointer; check for type consistency
10256 and turn inconsitent facts into UNREACHABLE.
10257 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10258 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
10259 type inconsistent querries; return UNREACHABLE instead.
10261 2014-02-03 Richard Henderson <rth@twiddle.net>
10264 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
10265 already processed this node.
10266 (normalize_one_pred_1): Pass along mark_set.
10267 (normalize_one_pred): Create and destroy a pointer_set_t.
10268 (normalize_one_pred_chain): Likewise.
10270 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
10272 PR gcov-profile/58602
10273 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
10275 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
10278 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
10279 -fno-devirtualize; try to devirtualize by the knowledge of
10280 virtual table pointer given by aggregate propagation.
10281 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
10282 (ipa_print_node_jump_functions): Dump also offset that
10283 is relevant for polymorphic calls.
10284 (determine_known_aggregate_parts): Add arg_type parameter; use it
10285 instead of determining the type from pointer type.
10286 (ipa_compute_jump_functions_for_edge): Update call of
10287 determine_known_aggregate_parts.
10288 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
10289 (gimple_get_virt_method_for_binfo): ... here; simplify using
10290 vtable_pointer_value_to_vtable.
10291 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
10292 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
10293 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
10294 (vtable_pointer_value_to_vtable): Break out from ...; handle also
10296 (vtable_pointer_value_to_binfo): ... here.
10297 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
10299 2014-02-03 Teresa Johnson <tejohnson@google.com>
10301 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
10302 redef of outer loop index variable.
10304 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
10308 * doc/extend.texi (Function Attributes): Typo.
10310 2014-02-03 Cong Hou <congh@google.com>
10312 PR tree-optimization/60000
10313 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
10314 if the vectorized statement is a store. A store statement can only
10315 appear at the end of pattern statements.
10317 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10319 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
10320 (ix86_option_override_internal): Default long double to 64-bit for
10321 32-bit Bionic and to 128-bit for 64-bit Bionic.
10323 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
10324 TARGET_LONG_DOUBLE_128 is true.
10325 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
10327 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
10328 (mlong-double-64): Negate -mlong-double-128.
10329 (mlong-double-128): New option.
10331 * config/i386/i386-c.c (ix86_target_macros): Define
10332 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
10334 * doc/invoke.texi: Document -mlong-double-128.
10336 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
10338 PR rtl-optimization/60024
10339 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
10341 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
10343 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
10345 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10347 PR rtl-optimization/57662
10348 * sel-sched.c (code_motion_path_driver): Do not mark already not
10349 existing blocks in the visiting bitmap.
10351 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
10353 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
10354 on the insn being emitted.
10356 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
10357 Will Deacon <will.deacon@arm.com>
10359 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
10361 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10363 * config/arm/arm-tables.opt: Regenerate.
10365 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10367 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
10368 for vector types other than V16QImode.
10369 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
10370 define_expand, and call altivec_expand_vec_perm_le when producing
10371 code with little endian element order.
10372 (*altivec_vperm_<mode>_internal): New insn having previous
10373 behavior of altivec_vperm_<mode>.
10374 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
10375 altivec_expand_vec_perm_le when producing code with little endian
10377 (*altivec_vperm_<mode>_uns_internal): New insn having previous
10378 behavior of altivec_vperm_<mode>_uns.
10380 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10382 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
10383 (altivec_vsumsws): Add handling for -maltivec=be with a little
10385 (altivec_vsumsws_direct): New.
10386 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
10387 gen_altivec_vsumsws.
10389 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10391 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
10392 vtable_pointer_value_to_binfo): New functions.
10393 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
10394 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
10396 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
10398 * config/nios2/nios2.md (load_got_register): Initialize GOT
10399 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
10400 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
10402 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
10404 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
10405 preserverd by passthrough, do not propagate the type.
10407 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10409 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
10410 (mips_atomic_assign_expand_fenv): New function.
10411 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
10413 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10415 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
10416 (__builtin_mips_set_fcsr): Likewise.
10417 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
10418 MIPS_USI_FTYPE_VOID.
10419 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
10420 (mips16_expand_set_fcsr): Likewise.
10421 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
10422 (mips16_set_fcsr_stub): Likewise.
10423 (mips16_get_fcsr_one_only_stub): New class.
10424 (mips16_set_fcsr_one_only_stub): Likewise.
10425 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
10426 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
10427 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
10428 (hard_float): New availability predicate.
10429 (mips_builtins): Add get_fcsr and set_fcsr.
10430 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
10431 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
10432 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
10433 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
10434 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
10437 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
10439 * config/mips/mips.c (mips_one_only_stub): New class.
10440 (mips_need_mips16_rdhwr_p): Replace with...
10441 (mips16_rdhwr_stub): ...this new variable.
10442 (mips16_stub_call_address): New function.
10443 (mips16_rdhwr_one_only_stub): New class.
10444 (mips_expand_thread_pointer): Use mips16_stub_call_address.
10445 (mips_output_mips16_rdhwr): Delete.
10446 (mips_finish_stub): New function.
10447 (mips_code_end): Use it to handle rdhwr stubs.
10449 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
10452 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
10453 when calculating size of integer atomic types.
10455 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
10457 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
10459 2014-02-01 Jakub Jelinek <jakub@redhat.com>
10461 PR tree-optimization/60003
10462 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
10463 * profile.c (branch_prob): Use gimple_call_builtin_p
10464 to check for BUILT_IN_SETJMP_RECEIVER.
10465 * tree-inline.c (copy_bb): Call notice_special_calls.
10467 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
10470 * lra-constraints.c (process_alt_operands): Update reload_sum only
10473 2014-01-31 Richard Henderson <rth@redhat.com>
10475 PR middle-end/60004
10476 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
10477 until after else_eh is processed.
10479 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10481 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
10482 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
10483 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
10484 in smmintrin.h, remove them.
10485 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
10486 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
10487 * config/i386/i386.md (ROUND_SAE): Fix value.
10488 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
10489 (const48_operand): New.
10490 * config/i386/subst.md (round), (round_expand): Use
10491 const_4_or_8_to_11_operand.
10492 (round_saeonly), (round_saeonly_expand): Use const48_operand.
10494 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
10496 * config/i386/constraints.md (Yk): Swap meaning with k.
10497 * config/i386/i386.md (movhi_internal): Change Yk to k.
10498 (movqi_internal): Ditto.
10499 (*k<logic><mode>): Ditto.
10502 (kandn<mode>): Ditto.
10503 (*<code>hi_1): Ditto.
10504 (*<code>qi_1): Ditto.
10505 (kxnor<mode>): Ditto.
10506 (kortestzhi): Ditto.
10507 (kortestchi): Ditto.
10509 (*one_cmplhi2_1): Ditto.
10510 (*one_cmplqi2_1): Ditto.
10511 * config/i386/sse.md (): Change k to Yk.
10512 (avx512f_load<mode>_mask): Ditto.
10513 (avx512f_blendm<mode>): Ditto.
10514 (avx512f_store<mode>_mask): Ditto.
10515 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
10516 (avx512f_storedqu<mode>_mask): Ditto.
10517 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
10519 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
10520 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
10521 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
10522 (avx512f_maskcmp<mode>3): Ditto.
10523 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
10524 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
10525 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
10526 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
10527 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
10528 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
10529 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
10530 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
10531 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
10532 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
10533 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
10534 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
10535 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
10536 (vec_extract_lo_<mode>_maskm): Ditto.
10537 (vec_extract_hi_<mode>_maskm): Ditto.
10538 (avx512f_vternlog<mode>_mask): Ditto.
10539 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
10540 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
10541 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
10542 (avx512f_<code>v8div16qi2_mask): Ditto.
10543 (avx512f_<code>v8div16qi2_mask_store): Ditto.
10544 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
10545 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
10546 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
10547 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
10548 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
10549 (*avx512pf_gatherpf<mode>df_mask): Ditto.
10550 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
10551 (*avx512pf_scatterpf<mode>df_mask): Ditto.
10552 (avx512cd_maskb_vec_dupv8di): Ditto.
10553 (avx512cd_maskw_vec_dupv16si): Ditto.
10554 (avx512f_vpermi2var<mode>3_maskz): Ditto.
10555 (avx512f_vpermi2var<mode>3_mask): Ditto.
10556 (avx512f_vpermi2var<mode>3_mask): Ditto.
10557 (avx512f_vpermt2var<mode>3_maskz): Ditto.
10558 (*avx512f_gathersi<mode>): Ditto.
10559 (*avx512f_gathersi<mode>_2): Ditto.
10560 (*avx512f_gatherdi<mode>): Ditto.
10561 (*avx512f_gatherdi<mode>_2): Ditto.
10562 (*avx512f_scattersi<mode>): Ditto.
10563 (*avx512f_scatterdi<mode>): Ditto.
10564 (avx512f_compress<mode>_mask): Ditto.
10565 (avx512f_compressstore<mode>_mask): Ditto.
10566 (avx512f_expand<mode>_mask): Ditto.
10567 * config/i386/subst.md (mask): Change k to Yk.
10568 (mask_scalar_merge): Ditto.
10571 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
10573 * doc/extend.texi (Vector Extensions): Document ?: in C++.
10575 2014-01-31 Richard Biener <rguenther@suse.de>
10577 PR middle-end/59990
10578 * builtins.c (fold_builtin_memory_op): Make sure to not
10579 use a floating-point mode or a boolean or enumeral type for
10580 the copy operation.
10582 2014-01-30 DJ Delorie <dj@redhat.com>
10584 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
10585 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
10586 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
10587 whenever main() has an epilogue.
10589 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10591 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
10592 unused variable "field".
10593 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
10594 (vsx_mergeh_<mode>): Likewise.
10595 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
10596 (altivec_vmrghh): Likewise.
10597 (altivec_vmrghw): Likewise.
10598 (altivec_vmrglb): Likewise.
10599 (altivec_vmrglh): Likewise.
10600 (altivec_vmrglw): Likewise.
10601 (altivec_vspltb): Add missing uses.
10602 (altivec_vsplth): Likewise.
10603 (altivec_vspltw): Likewise.
10604 (altivec_vspltsf): Likewise.
10606 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10609 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
10610 frame related instructions.
10612 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
10614 PR rtl-optimization/59959
10615 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
10616 any reload of register whose subreg is invalid.
10618 2014-01-30 Jakub Jelinek <jakub@redhat.com>
10620 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
10621 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
10622 Add missing return type - void.
10624 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10626 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
10627 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
10628 remove element index adjustment for endian (now handled in vsx.md
10630 (altivec_expand_vec_perm_const): Use
10631 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
10632 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
10633 (vsx_xxspltw_<mode>): Adjust element index for little endian.
10634 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
10635 define_expand and a new define_insn *altivec_vspltb_internal;
10636 adjust for -maltivec=be on a little endian target.
10637 (altivec_vspltb_direct): New.
10638 (altivec_vsplth): Divide into a define_expand and a new
10639 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
10640 little endian target.
10641 (altivec_vsplth_direct): New.
10642 (altivec_vspltw): Divide into a define_expand and a new
10643 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
10644 little endian target.
10645 (altivec_vspltw_direct): New.
10646 (altivec_vspltsf): Divide into a define_expand and a new
10647 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
10648 a little endian target.
10650 2014-01-30 Richard Biener <rguenther@suse.de>
10652 PR tree-optimization/59993
10653 * tree-ssa-forwprop.c (associate_pointerplus): Check we
10654 can propagate form the earlier stmt and avoid the transform
10655 when the intermediate result is needed.
10657 2014-01-30 Alangi Derick <alangiderick@gmail.com>
10659 * README.Portability: Fix typo.
10661 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
10663 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
10664 comparison_operator with ordered_comparison_operator.
10666 2014-01-30 Nick Clifton <nickc@redhat.com>
10668 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
10669 Rename to mn10300_store_multiple_regs.
10670 * config/mn10300/mn10300.c: Likewise.
10671 * config/mn10300/mn10300.md (store_movm): Fix typo: call
10672 store_multiple_regs.
10673 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
10674 Call mn10300_store_multiple_regs.
10676 2014-01-30 Nick Clifton <nickc@redhat.com>
10677 DJ Delorie <dj@redhat.com>
10679 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
10680 %fp 2 to keep registers after it properly word-aligned.
10681 (rl78_alloc_physical_registers_umul): Handle the case where both
10682 input operands are the same.
10684 2014-01-30 Richard Biener <rguenther@suse.de>
10686 PR tree-optimization/59903
10687 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
10690 2014-01-30 Jason Merrill <jason@redhat.com>
10693 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
10696 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
10698 2014-01-30 Richard Biener <rguenther@suse.de>
10700 PR tree-optimization/59951
10701 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
10703 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
10706 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
10707 SFmode to DFmode case.
10709 2014-01-29 DJ Delorie <dj@redhat.com>
10711 * config/msp430/msp430.opt (-minrt): New.
10712 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
10714 (ENDFILE_SPEC): Likewise.
10716 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
10718 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
10719 (estimate_function_body_sizes): Use it.
10721 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
10724 * dwarf2out.c (is_cxx_auto): New.
10725 (is_base_type): Use it.
10726 (gen_type_die_with_usage): Likewise.
10728 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10730 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
10731 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
10732 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
10733 -maltivec=be with LE targets.
10734 (vsx_mergeh_<mode>): Likewise.
10735 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
10736 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
10737 (altivec_vmrghb): Replace with define_expand and new
10738 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
10739 (altivec_vmrghb_direct): New define_insn.
10740 (altivec_vmrghh): Replace with define_expand and new
10741 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
10742 (altivec_vmrghh_direct): New define_insn.
10743 (altivec_vmrghw): Replace with define_expand and new
10744 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
10745 (altivec_vmrghw_direct): New define_insn.
10746 (*altivec_vmrghsf): Adjust for endianness.
10747 (altivec_vmrglb): Replace with define_expand and new
10748 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
10749 (altivec_vmrglb_direct): New define_insn.
10750 (altivec_vmrglh): Replace with define_expand and new
10751 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
10752 (altivec_vmrglh_direct): New define_insn.
10753 (altivec_vmrglw): Replace with define_expand and new
10754 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
10755 (altivec_vmrglw_direct): New define_insn.
10756 (*altivec_vmrglsf): Adjust for endianness.
10757 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10758 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10759 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
10760 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
10761 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10762 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10763 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
10764 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
10766 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
10768 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
10769 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
10772 2014-01-29 Richard Biener <rguenther@suse.de>
10774 PR tree-optimization/58742
10775 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
10776 associate_pointerplus_align.
10777 (associate_pointerplus_diff): New function.
10778 (associate_pointerplus): Likewise. Call associate_pointerplus_align
10779 and associate_pointerplus_diff.
10781 2014-01-29 Richard Biener <rguenther@suse.de>
10783 * lto-streamer.h (LTO_major_version): Bump to 3.
10784 (LTO_minor_version): Reset to 0.
10786 2014-01-29 Renlin Li <Renlin.Li@arm.com>
10788 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
10789 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
10790 (arm_file_start): Generate correct asm header for armv7ve.
10791 * config/arm/bpabi.h: Add multilib support for armv7ve.
10792 * config/arm/driver-arm.c: Change the architectures of cortex-a7
10793 and cortex-a15 to armv7ve.
10794 * config/arm/t-aprofile: Add multilib support for armv7ve.
10795 * doc/invoke.texi: Document -march=armv7ve.
10797 2014-01-29 Richard Biener <rguenther@suse.de>
10799 PR tree-optimization/58742
10800 * tree-ssa-forwprop.c (associate_plusminus): Return true
10801 if we changed sth, defer EH cleanup to ...
10802 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
10803 (simplify_mult): New function.
10805 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10807 PR middle-end/59917
10808 PR tree-optimization/59920
10809 * tree.c (build_common_builtin_nodes): Remove
10810 __builtin_setjmp_dispatcher initialization.
10811 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
10812 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
10813 instead of gsi_after_labels + manually skipping debug stmts.
10814 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
10815 ignore bbs with IFN_ABNORMAL_DISPATCHER.
10816 * tree-inline.c (copy_edges_for_bb): Remove
10817 can_make_abnormal_goto argument, instead add abnormal_goto_dest
10818 argument. Ignore computed_goto_p stmts. Don't call
10819 make_abnormal_goto_edges. If a call might need abnormal edges
10820 for non-local gotos, see if it already has an edge to
10821 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
10822 with true argument, don't do anything then, otherwise add
10823 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
10824 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
10826 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
10827 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
10828 (lower_stmt): Don't set data->calls_builtin_setjmp.
10829 (lower_builtin_setjmp): Adjust comment.
10830 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
10831 * tree-cfg.c (found_computed_goto): Remove.
10832 (factor_computed_gotos): Remove.
10833 (make_goto_expr_edges): Return bool, true for computed gotos.
10834 Don't call make_abnormal_goto_edges.
10835 (build_gimple_cfg): Don't set found_computed_goto, don't call
10836 factor_computed_gotos.
10837 (computed_goto_p): No longer static.
10838 (make_blocks): Don't set found_computed_goto.
10839 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
10840 (make_edges): If make_goto_expr_edges returns true, push bb
10841 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
10842 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
10843 vector. Record mapping between bbs and OpenMP regions if there
10844 are any, adjust make_gimple_omp_edges caller. Call
10845 handle_abnormal_edges.
10846 (make_abnormal_goto_edges): Remove.
10847 * tree-cfg.h (make_abnormal_goto_edges): Remove.
10848 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
10849 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
10850 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
10851 * internal-fn.def (ABNORMAL_DISPATCHER): New.
10852 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
10853 filling *region also set *region_idx to (*region)->entry->index.
10856 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
10857 For REGs set ORIGINAL_REGNO.
10859 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
10861 * doc/md.texi: Mention that a target shouldn't implement
10862 vec_widen_(s|u)mul_even/odd pair if it is less efficient
10865 2014-01-29 Jakub Jelinek <jakub@redhat.com>
10867 PR tree-optimization/59594
10868 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
10869 a copy of the datarefs vector rather than the vector itself.
10871 2014-01-28 Jason Merrill <jason@redhat.com>
10874 * dwarf2out.c (auto_die): New static.
10875 (gen_type_die_with_usage): Handle C++1y 'auto'.
10876 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
10879 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
10882 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
10883 (SPEC_X32): Likewise.
10884 (SPEC_64): Likewise.
10885 * config/i386/i386.c (ix86_option_override_internal): Turn off
10886 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
10888 (x86_file_start): Output .code16gcc for TARGET_16BIT.
10889 * config/i386/i386.h (TARGET_16BIT): New macro.
10890 (TARGET_16BIT_P): Likewise.
10891 * config/i386/i386.opt: Add m16.
10892 * doc/invoke.texi: Document -m16.
10894 2014-01-28 Jakub Jelinek <jakub@redhat.com>
10896 PR preprocessor/59935
10897 * input.c (location_get_source_line): Bail out on when line number
10898 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
10900 2014-01-28 Richard Biener <rguenther@suse.de>
10902 PR tree-optimization/58742
10903 * tree-ssa-forwprop.c (associate_plusminus): Handle
10904 pointer subtraction of the form (T)(P + A) - (T)P.
10906 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10908 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
10911 2014-01-28 Richard Biener <rguenther@suse.de>
10914 2014-01-28 Richard Biener <rguenther@suse.de>
10916 PR rtl-optimization/45364
10917 PR rtl-optimization/59890
10918 * var-tracking.c (local_get_addr_clear_given_value): Handle
10919 already cleared slot.
10920 (val_reset): Handle not allocated local_get_addr_cache.
10921 (vt_find_locations): Use post-order on the inverted CFG.
10923 2014-01-28 Richard Biener <rguenther@suse.de>
10925 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
10927 2014-01-28 Richard Biener <rguenther@suse.de>
10929 PR rtl-optimization/45364
10930 PR rtl-optimization/59890
10931 * var-tracking.c (local_get_addr_clear_given_value): Handle
10932 already cleared slot.
10933 (val_reset): Handle not allocated local_get_addr_cache.
10934 (vt_find_locations): Use post-order on the inverted CFG.
10936 2014-01-28 Alan Modra <amodra@gmail.com>
10938 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
10939 * configure.ac <recursive call for build != host>: Define
10940 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
10941 and LD_FOR_BUILD too.
10942 * configure: Regenerate.
10944 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
10946 * config/i386/i386.c (get_builtin_code_for_version): Separate
10947 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
10948 Broadwell from Haswell.
10950 2014-01-27 Steve Ellcey <sellcey@mips.com>
10952 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
10953 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
10954 * config/mips/mips.c (mips_option_override): Change setting
10956 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
10957 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
10958 Change from Mask to Var.
10960 2014-01-27 Jeff Law <law@redhat.com>
10962 * ipa-inline.c (inline_small_functions): Fix typo.
10964 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
10966 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
10967 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
10968 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
10969 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
10970 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
10971 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
10972 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
10973 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
10974 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
10975 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
10976 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
10977 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
10978 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
10979 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
10980 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
10981 (_mm512_storeu_epi64): Ditto.
10982 (_mm512_cmpge_epi32_mask): Ditto.
10983 (_mm512_cmpge_epu32_mask): Ditto.
10984 (_mm512_cmpge_epi64_mask): Ditto.
10985 (_mm512_cmpge_epu64_mask): Ditto.
10986 (_mm512_cmple_epi32_mask): Ditto.
10987 (_mm512_cmple_epu32_mask): Ditto.
10988 (_mm512_cmple_epi64_mask): Ditto.
10989 (_mm512_cmple_epu64_mask): Ditto.
10990 (_mm512_cmplt_epi32_mask): Ditto.
10991 (_mm512_cmplt_epu32_mask): Ditto.
10992 (_mm512_cmplt_epi64_mask): Ditto.
10993 (_mm512_cmplt_epu64_mask): Ditto.
10994 (_mm512_cmpneq_epi32_mask): Ditto.
10995 (_mm512_cmpneq_epu32_mask): Ditto.
10996 (_mm512_cmpneq_epi64_mask): Ditto.
10997 (_mm512_cmpneq_epu64_mask): Ditto.
10998 (_mm512_expand_pd): Ditto.
10999 (_mm512_expand_ps): Ditto.
11000 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
11001 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
11002 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
11003 * config/i386/i386.c (ix86_builtins): Add
11004 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
11005 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
11006 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
11007 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
11008 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
11009 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
11010 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
11011 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
11012 IX86_BUILTIN_PMOVUSQW512_MEM.
11013 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
11014 __builtin_ia32_pmovsqd512mem_mask,
11015 __builtin_ia32_pmovqd512mem_mask,
11016 __builtin_ia32_pmovusqw512mem_mask,
11017 __builtin_ia32_pmovsqw512mem_mask,
11018 __builtin_ia32_pmovqw512mem_mask,
11019 __builtin_ia32_pmovusdw512mem_mask,
11020 __builtin_ia32_pmovsdw512mem_mask,
11021 __builtin_ia32_pmovdw512mem_mask,
11022 __builtin_ia32_pmovqb512mem_mask,
11023 __builtin_ia32_pmovusqb512mem_mask,
11024 __builtin_ia32_pmovsqb512mem_mask,
11025 __builtin_ia32_pmovusdb512mem_mask,
11026 __builtin_ia32_pmovsdb512mem_mask,
11027 __builtin_ia32_pmovdb512mem_mask.
11028 (bdesc_args): Add __builtin_ia32_expanddf512,
11029 __builtin_ia32_expandsf512.
11030 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
11031 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
11032 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
11033 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
11034 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
11035 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
11036 (avx512f_<code>v8div16qi2_mask_store): This.
11037 (avx512f_expand<mode>): New.
11039 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
11041 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
11043 (_mm512_mask_prefetch_i64gather_pd): Ditto.
11044 (_mm512_prefetch_i32scatter_pd): Ditto.
11045 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
11046 (_mm512_prefetch_i64scatter_pd): Ditto.
11047 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
11048 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
11049 (_mm512_mask_prefetch_i64gather_ps): Ditto.
11050 (_mm512_prefetch_i32scatter_ps): Ditto.
11051 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
11052 (_mm512_prefetch_i64scatter_ps): Ditto.
11053 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
11054 * config/i386/i386-builtin-types.def: Define
11055 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
11056 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
11057 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
11058 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
11059 IX86_BUILTIN_SCATTERPFQPD.
11060 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
11061 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
11062 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
11063 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
11064 __builtin_ia32_scatterpfqps.
11065 (ix86_expand_builtin): Expand new built-ins.
11066 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
11067 fix memory access data type.
11068 (*avx512pf_gatherpf<mode>_mask): Ditto.
11069 (*avx512pf_gatherpf<mode>): Ditto.
11070 (avx512pf_scatterpf<mode>): Ditto.
11071 (*avx512pf_scatterpf<mode>_mask): Ditto.
11072 (*avx512pf_scatterpf<mode>): Ditto.
11073 (GATHER_SCATTER_SF_MEM_MODE): New.
11074 (avx512pf_gatherpf<mode>df): Ditto.
11075 (*avx512pf_gatherpf<mode>df_mask): Ditto.
11076 (*avx512pf_scatterpf<mode>df): Ditto.
11078 2014-01-27 Jakub Jelinek <jakub@redhat.com>
11081 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
11082 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
11085 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11087 * common/config/arm/arm-common.c
11088 (arm_rewrite_mcpu): Handle multiple names.
11090 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11092 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
11094 * gimple-builder.h (create_gimple_tmp): Delete.
11096 2014-01-27 Christian Bruel <christian.bruel@st.com>
11098 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
11101 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
11103 * config/pa/pa.md (call): Generate indirect long calls to non-local
11104 functions when outputing 32-bit code.
11105 (call_value): Likewise except for special call to buggy powf function.
11107 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
11108 portable runtime and PIC indirect calls.
11109 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
11110 and PIC call sequences. Use ldo instead of blr to set return register
11111 in PIC call sequence.
11113 2014-01-25 Walter Lee <walt@tilera.com>
11115 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
11116 avoid clobbering a live register.
11118 2014-01-25 Walter Lee <walt@tilera.com>
11120 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
11121 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
11122 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
11123 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
11125 2014-01-25 Walter Lee <walt@tilera.com>
11127 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
11128 arguments on even registers.
11129 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
11131 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
11132 (BIGGEST_ALIGNMENT): Ditto.
11133 (BIGGEST_FIELD_ALIGNMENT): Ditto.
11135 2014-01-25 Walter Lee <walt@tilera.com>
11137 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
11138 insns before bundling.
11139 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
11141 2014-01-25 Walter Lee <walt@tilera.com>
11143 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
11144 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
11145 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
11147 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11149 * config/mips/constraints.md (kl): Delete.
11150 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
11151 define expands, using...
11152 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
11153 instructions for MIPS16.
11154 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
11155 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
11157 2014-01-25 Walter Lee <walt@tilera.com>
11159 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
11163 2014-01-25 Walter Lee <walt@tilera.com>
11165 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
11166 (TARGET_EXPAND_TO_RTL_HOOK): Define.
11168 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
11170 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
11173 2014-01-25 Jakub Jelinek <jakub@redhat.com>
11175 * print-rtl.c (in_call_function_usage): New var.
11176 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
11177 EXPR_LIST mode as mode and not as reg note name.
11179 PR middle-end/59561
11180 * cfgloopmanip.c (copy_loop_info): If
11181 loop->warned_aggressive_loop_optimizations, make sure
11182 the flag is set in target loop too.
11184 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
11186 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
11188 * builtins.def: Likewise.
11189 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
11190 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
11191 * ira.c (ira_setup_eliminable_regset): Likewise.
11192 * omp-low.c (gate_expand_omp): Likewise.
11193 (execute_lower_omp): Likewise.
11194 (diagnose_sb_0): Likewise.
11195 (gate_diagnose_omp_blocks): Likewise.
11196 (simd_clone_clauses_extract): Likewise.
11199 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11201 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
11202 correction for little endian...
11203 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
11206 2014-01-24 Jeff Law <law@redhat.com>
11208 PR tree-optimization/59919
11209 * tree-vrp.c (find_assert_locations_1): Do not register asserts
11210 for non-returning calls.
11212 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
11214 * common/config/aarch64/aarch64-common.c
11215 (aarch64_rewrite_mcpu): Handle multiple names.
11216 * config/aarch64/aarch64.h
11217 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
11219 2014-01-24 Dodji Seketeli <dodji@redhat.com>
11221 * input.c (add_file_to_cache_tab): Handle the case where fopen
11224 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
11227 * config/i386/i386.md (pushsf splitter): Get stack adjustment
11228 from push operand if code of push isn't PRE_DEC.
11230 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11233 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
11234 -mquad-memory-atomic. Update -mquad-memory documentation to say
11235 it is only used for non-atomic loads/stores.
11237 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
11238 -mquad-memory or -mquad-memory-atomic switches.
11240 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
11241 -mquad-memory-atomic to ISA 2.07 support.
11243 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
11244 to separate support of normal quad word memory operations (ldq, stq)
11245 from the atomic quad word memory operations.
11247 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
11248 support to separate non-atomic quad word operations from atomic
11249 quad word operations. Disable non-atomic quad word operations in
11250 little endian mode so that we don't have to swap words after the
11251 load and before the store.
11252 (quad_load_store_p): Add comment about atomic quad word support.
11253 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
11254 options printed with -mdebug=reg.
11256 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
11257 -mquad-memory-atomic as the test for whether we have quad word
11258 atomic instructions.
11259 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
11260 or -mp8-vector are used, allow byte/half-word atomic operations.
11262 * config/rs6000/sync.md (load_lockedti): Insure that the address
11263 is a proper indexed or indirect address for the lqarx instruction.
11264 On little endian systems, swap the hi/lo registers after the lqarx
11266 (load_lockedpti): Use indexed_or_indirect_operand predicate to
11267 insure the address is valid for the lqarx instruction.
11268 (store_conditionalti): Insure that the address is a proper indexed
11269 or indirect address for the stqcrx. instruction. On little endian
11270 systems, swap the hi/lo registers before doing the stqcrx.
11272 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
11273 insure the address is valid for the stqcrx. instruction.
11275 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
11276 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
11277 type of quad memory support is available.
11279 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
11281 PR regression/59915
11282 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
11283 there is a danger of looping.
11285 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
11287 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
11288 force flag_ira_loop_pressure if set via command line.
11290 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11292 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
11293 (ashr_simd): New builtin handling DI mode.
11294 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
11295 (aarch64_sshr_simddi): New match pattern.
11296 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
11297 (vshrd_n_s64): Likewise.
11298 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
11300 2014-01-23 Nick Clifton <nickc@redhat.com>
11302 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
11303 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
11304 favour of mcu specific scripts.
11305 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
11308 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11309 Alex Velenko <Alex.Velenko@arm.com>
11311 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
11312 (vaddv_s16): Likewise.
11313 (vaddv_s32): Likewise.
11314 (vaddv_u8): Likewise.
11315 (vaddv_u16): Likewise.
11316 (vaddv_u32): Likewise.
11317 (vaddvq_s8): Likewise.
11318 (vaddvq_s16): Likewise.
11319 (vaddvq_s32): Likewise.
11320 (vaddvq_s64): Likewise.
11321 (vaddvq_u8): Likewise.
11322 (vaddvq_u16): Likewise.
11323 (vaddvq_u32): Likewise.
11324 (vaddvq_u64): Likewise.
11325 (vaddv_f32): Likewise.
11326 (vaddvq_f32): Likewise.
11327 (vaddvq_f64): Likewise.
11328 (vmaxv_f32): Likewise.
11329 (vmaxv_s8): Likewise.
11330 (vmaxv_s16): Likewise.
11331 (vmaxv_s32): Likewise.
11332 (vmaxv_u8): Likewise.
11333 (vmaxv_u16): Likewise.
11334 (vmaxv_u32): Likewise.
11335 (vmaxvq_f32): Likewise.
11336 (vmaxvq_f64): Likewise.
11337 (vmaxvq_s8): Likewise.
11338 (vmaxvq_s16): Likewise.
11339 (vmaxvq_s32): Likewise.
11340 (vmaxvq_u8): Likewise.
11341 (vmaxvq_u16): Likewise.
11342 (vmaxvq_u32): Likewise.
11343 (vmaxnmv_f32): Likewise.
11344 (vmaxnmvq_f32): Likewise.
11345 (vmaxnmvq_f64): Likewise.
11346 (vminv_f32): Likewise.
11347 (vminv_s8): Likewise.
11348 (vminv_s16): Likewise.
11349 (vminv_s32): Likewise.
11350 (vminv_u8): Likewise.
11351 (vminv_u16): Likewise.
11352 (vminv_u32): Likewise.
11353 (vminvq_f32): Likewise.
11354 (vminvq_f64): Likewise.
11355 (vminvq_s8): Likewise.
11356 (vminvq_s16): Likewise.
11357 (vminvq_s32): Likewise.
11358 (vminvq_u8): Likewise.
11359 (vminvq_u16): Likewise.
11360 (vminvq_u32): Likewise.
11361 (vminnmv_f32): Likewise.
11362 (vminnmvq_f32): Likewise.
11363 (vminnmvq_f64): Likewise.
11365 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
11367 * config/aarch64/aarch64-simd.md
11368 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
11369 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
11370 (*aarch64_mul3_elt<mode>): Likewise.
11371 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
11372 (*aarch64_mul3_elt_to_64v2df): Likewise.
11373 (*aarch64_mla_elt<mode>): Likewise.
11374 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
11375 (*aarch64_mls_elt<mode>): Likewise.
11376 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
11377 (*aarch64_fma4_elt<mode>): Likewise.
11378 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
11379 (*aarch64_fma4_elt_to_64v2df): Likewise.
11380 (*aarch64_fnma4_elt<mode>): Likewise.
11381 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
11382 (*aarch64_fnma4_elt_to_64v2df): Likewise.
11383 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
11384 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
11385 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11386 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
11387 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
11388 (aarch64_sqdmull_lane<mode>_internal): Likewise.
11389 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
11391 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
11393 * config/aarch64/aarch64-simd.md
11394 (aarch64_be_checked_get_lane<mode>): New define_expand.
11395 * config/aarch64/aarch64-simd-builtins.def
11396 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
11397 New builtin definition.
11398 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
11399 Use new safe be builtin.
11401 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
11403 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
11405 (aarch64_be_st1<mode>): Likewise.
11406 (aarch_ld1<VALL:mode>): Define_expand modified.
11407 (aarch_st1<VALL:mode>): Likewise.
11408 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
11409 (UNSPEC_ST1): Likewise.
11411 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
11413 * config/microblaze/microblaze.md: Add trap insn and attribute
11415 2014-01-23 Dodji Seketeli <dodji@redhat.com>
11417 PR preprocessor/58580
11418 * input.h (location_get_source_line): Take an additional line_size
11420 (void diagnostics_file_cache_fini): Declare new function.
11421 * input.c (struct fcache): New type.
11422 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
11423 New static constants.
11424 (diagnostic_file_cache_init, total_lines_num)
11425 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
11426 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
11427 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
11428 (get_next_line, read_next_line, goto_next_line, read_line_num):
11429 New static function definitions.
11430 (diagnostic_file_cache_fini): New function.
11431 (location_get_source_line): Take an additional output line_len
11432 parameter. Re-write using lookup_or_add_file_to_cache_tab and
11434 * diagnostic.c (diagnostic_finish): Call
11435 diagnostic_file_cache_fini.
11436 (adjust_line): Take an additional input parameter for the length
11437 of the line, rather than calculating it with strlen.
11438 (diagnostic_show_locus): Adjust the use of
11439 location_get_source_line and adjust_line with respect to their new
11440 signature. While displaying a line now, do not stop at the first
11441 null byte. Rather, display the zero byte as a space and keep
11442 going until we reach the size of the line.
11443 * Makefile.in: Add vec.o to OBJS-libcommon
11445 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11446 Ilya Tocar <ilya.tocar@intel.com>
11448 * config/i386/avx512fintrin.h (_mm512_kmov): New.
11449 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
11450 (__builtin_ia32_kmov16): Ditto.
11451 * config/i386/i386.md (UNSPEC_KMOV): New.
11454 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
11456 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
11457 (_mm512_storeu_si512): Ditto.
11459 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
11462 * rtl.h (get_referenced_operands): Declare.
11463 * recog.c (get_referenced_operands): New function.
11464 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
11465 operands have been referenced when recording LO_SUM references.
11467 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
11469 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
11471 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11473 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
11474 Enable for generic and recent AMD targets.
11476 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
11478 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
11479 ARG_SIZE note when adjustment was eliminated.
11481 2014-01-22 Jeff Law <law@redhat.com>
11483 PR tree-optimization/59597
11484 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
11485 in file. Accept new argument REGISTERING and use it to modify
11486 dump output appropriately.
11487 (register_jump_thread): Corresponding changes.
11488 (mark_threaded_blocks): Reinstate code to cancel unprofitable
11489 thread paths involving joiner blocks. Add code to dump cancelled
11490 jump threading paths.
11492 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
11494 PR rtl-optimization/59477
11495 * lra-constraints.c (inherit_in_ebb): Process call for living hard
11496 regs. Update reloads_num and potential_reload_hard_regs for all insns.
11498 2014-01-22 Tom Tromey <tromey@redhat.com>
11500 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
11502 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
11504 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11506 PR rtl-optimization/59896
11507 * lra-constraints.c (process_alt_operands): Check unused note for
11508 matched operands of insn with no output reloads.
11510 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
11512 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
11513 (mips_move_from_gpr_cost): Likewise.
11515 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
11517 PR rtl-optimization/59858
11518 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
11519 ira_class_hard_regs_num.
11520 (process_alt_operands): Increase reject for dying matched operand.
11522 2014-01-21 Jakub Jelinek <jakub@redhat.com>
11525 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
11526 smaller than size, perform several stores or loads and stores
11527 at dst + count - size to store or copy all of size bytes, rather
11528 than just last modesize bytes.
11530 2014-01-20 DJ Delorie <dj@redhat.com>
11532 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
11533 that CLOBBERs are REGs before propogating their values.
11535 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
11537 PR middle-end/59789
11538 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
11539 (cgraph_inline_failed_type): New function.
11540 * cgraph.h (DEFCIFCODE): Add type.
11541 (cgraph_inline_failed_type_t): New enum.
11542 (cgraph_inline_failed_type): New prototype.
11543 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
11544 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
11545 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
11546 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
11547 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
11548 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
11549 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
11550 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
11551 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
11552 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
11553 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
11554 OPTIMIZATION_MISMATCH.
11555 * tree-inline.c (expand_call_inline): Emit errors during
11556 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
11558 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11561 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
11562 mode attribute in insn output.
11564 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
11566 * output.h (output_constant): Delete.
11567 * varasm.c (output_constant): Make private.
11569 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
11571 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
11573 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11575 PR middle-end/59860
11576 * tree.h (fold_builtin_strcat): New prototype.
11577 * builtins.c (fold_builtin_strcat): No longer static. Add len
11578 argument, if non-NULL, don't call c_strlen. Optimize
11579 directly into __builtin_memcpy instead of __builtin_strcpy.
11580 (fold_builtin_2): Adjust fold_builtin_strcat caller.
11581 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
11583 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
11585 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11586 for SImode_address_operand operands, having only a REG argument.
11588 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
11590 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
11591 loader name using mbig-endian.
11592 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
11594 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11596 * doc/invoke.texi (-march): Clarify documentation for AArch64.
11597 (-mtune): Likewise.
11600 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
11602 * config/aarch64/aarch64-protos.h
11603 (aarch64_cannot_change_mode_class_ptr): Declare.
11604 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
11605 aarch64_cannot_change_mode_class_ptr): New.
11606 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
11607 backend hook aarch64_cannot_change_mode_class.
11609 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
11611 * common/config/aarch64/aarch64-common.c
11612 (aarch64_handle_option): Don't handle any option order logic here.
11613 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
11614 selected_cpu, warn on architecture version mismatch.
11615 (aarch64_override_options): Fix parsing order for option strings.
11617 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11618 Iain Sandoe <iain@codesourcery.com>
11621 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
11622 warning. Amend comment to reflect current functionality.
11624 2014-01-20 Richard Biener <rguenther@suse.de>
11626 PR middle-end/59860
11627 * builtins.c (fold_builtin_strcat): Remove case better handled
11628 by tree-ssa-strlen.c.
11630 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
11632 * config/aarch64/aarch64.opt
11633 (mcpu, march, mtune): Make case-insensitive.
11635 2014-01-20 Jakub Jelinek <jakub@redhat.com>
11638 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
11639 if operands[1] is a REG or ZERO_EXTEND of a REG.
11641 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
11643 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
11645 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
11647 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
11648 long non-pic millicode calls.
11650 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11652 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
11654 2014-01-19 Kito Cheng <kito@0xlab.org>
11656 * builtins.c (expand_movstr): Check movstr expand done or fail.
11658 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11659 H.J. Lu <hongjiu.lu@intel.com>
11662 * config/i386/i386.md (*lea<mode>): Zero-extend return register
11663 to DImode for zero-extended addresses.
11665 2014-01-19 Jakub Jelinek <jakub@redhat.com>
11667 PR rtl-optimization/57763
11668 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
11669 on the new indirect jump_insn and increment LABEL_NUSES (label).
11671 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
11675 * config.gcc (x86_archs): New variable.
11676 (x86_64_archs): Likewise.
11677 (x86_cpus): Likewise.
11678 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
11679 --with-arch/--with-cpu= options.
11680 Support --with-arch=/--with-cpu={nehalem,westmere,
11681 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
11683 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11685 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
11686 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
11688 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
11690 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
11692 2014-01-18 Jakub Jelinek <jakub@redhat.com>
11695 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
11696 clear cpp_get_options (parse_in)->warn_unused_macros for
11697 ix86_target_macros_internal with cpp_define.
11699 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
11701 * jump.c (delete_related_insns): Keep (use (insn))s.
11702 * reorg.c (redundant_insn): Check for barriers too.
11704 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11706 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
11708 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
11710 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
11711 call to $$dyncall when TARGET_LONG_CALLS is true.
11713 2014-01-17 Jeff Law <law@redhat.com>
11715 * ree.c (combine_set_extension): Temporarily disable test for
11716 changing number of hard registers.
11718 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11720 PR middle-end/58125
11721 * ipa-inline-analysis.c (inline_free_summary):
11722 Do not free summary of aliases.
11724 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11726 PR middle-end/59706
11727 * gimplify.c (gimplify_expr): Use create_tmp_var
11728 instead of create_tmp_var_raw. If cond doesn't have
11729 integral type, don't add the IFN_ANNOTATE builtin at all.
11731 2014-01-17 Martin Jambor <mjambor@suse.cz>
11734 * ipa-cp.c (prev_edge_clone): New variable.
11735 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
11736 Also resize prev_edge_clone vector.
11737 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
11738 (ipcp_edge_removal_hook): New function.
11739 (ipcp_driver): Register ipcp_edge_removal_hook.
11741 2014-01-17 Andrew Pinski <apinski@cavium.com>
11742 Steve Ellcey <sellcey@mips.com>
11745 * config/mips/mips.c (mips_print_operand): Check operand mode instead
11748 2014-01-17 Jeff Law <law@redhat.com>
11750 PR middle-end/57904
11751 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
11752 so that pass_ccp runs first.
11754 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11756 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
11757 (ix86_adjust_cost): Use !TARGET_XXX.
11758 (do_reorder_for_imul): Likewise.
11759 (swap_top_of_ready_list): Likewise.
11760 (ix86_sched_reorder): Likewise.
11762 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
11764 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
11765 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
11766 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
11767 (intel_memset): New. Duplicate slm_memset.
11768 (intel_cost): New. Duplicate slm_cost.
11769 (m_INTEL): New macro.
11770 (processor_target_table): Add "intel".
11771 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
11772 with PROCESSOR_INTEL for "intel".
11773 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
11774 PROCESSOR_SILVERMONT.
11775 (ix86_issue_rate): Likewise.
11776 (ix86_adjust_cost): Likewise.
11777 (ia32_multipass_dfa_lookahead): Likewise.
11778 (swap_top_of_ready_list): Likewise.
11779 (ix86_sched_reorder): Likewise.
11780 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
11781 instead of TARGET_OPT_AGU.
11782 * config/i386/i386.h (TARGET_INTEL): New.
11783 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
11784 (processor_type): Add PROCESSOR_INTEL.
11785 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
11786 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
11788 2014-01-17 Marek Polacek <polacek@redhat.com>
11791 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
11794 2014-01-17 Richard Biener <rguenther@suse.de>
11796 PR tree-optimization/46590
11797 * opts.c (default_options_table): Add entries for
11798 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
11799 all enabled at -O1 but not for -Og.
11800 * common.opt (fbranch-count-reg): Remove Init(1).
11801 (fmove-loop-invariants): Likewise.
11802 (ftree-pta): Likewise.
11804 2014-01-17 Jakub Jelinek <jakub@redhat.com>
11806 * config/i386/i386.c (ix86_data_alignment): For compatibility with
11807 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
11808 decls to at least the GCC 4.8 used alignments.
11811 * tree-nested.c (convert_nonlocal_reference_stmt,
11812 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
11813 of GIMPLE_BIND stmts, adjust associated decls.
11815 2014-01-17 Richard Biener <rguenther@suse.de>
11817 PR tree-optimization/46590
11818 * vec.h (vec<>::bseach): New member function implementing
11819 binary search according to C89 bsearch.
11820 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
11821 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
11822 bitmap pointer again. Make accesses_in_loop a flat array.
11823 (mem_ref_obstack): New global.
11824 (outermost_indep_loop): Adjust for mem_ref->stored changes.
11825 (mark_ref_stored): Likewise.
11826 (ref_indep_loop_p_2): Likewise.
11827 (set_ref_stored_in_loop): New helper function.
11828 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
11829 (memref_free): Adjust.
11830 (record_mem_ref_loc): Simplify.
11831 (gather_mem_refs_stmt): Adjust.
11832 (sort_locs_in_loop_postorder_cmp): New function.
11833 (analyze_memory_references): Sort accesses_in_loop after
11834 loop postorder number.
11835 (find_ref_loc_in_loop_cmp): New function.
11836 (for_all_locs_in_loop): Find relevant cluster of locs in
11837 accesses_in_loop and iterate without recursion.
11838 (execute_sm): Avoid uninit warning.
11839 (struct ref_always_accessed): Simplify.
11840 (ref_always_accessed::operator ()): Likewise.
11841 (ref_always_accessed_p): Likewise.
11842 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
11843 loop postorder numbers here.
11844 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
11847 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
11850 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
11851 on decls for which assemble_alias has been called.
11853 2014-01-17 Nick Clifton <nickc@redhat.com>
11855 * config/msp430/msp430.opt: (mcpu): New option.
11856 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
11857 (msp430_option_override): Parse target_cpu. If the MCU name
11858 matches a generic string, clear target_mcu.
11859 (msp430_attr): Allow numeric interrupt values up to 63.
11860 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
11861 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
11863 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
11865 * config/msp430/msp430.md (popm): Use %J rather than %I.
11866 (addsi3): Use msp430_nonimmediate_operand for operand 2.
11867 (addhi_cy_i): Use immediate_operand for operand 2.
11868 * doc/invoke.texi: Document -mcpu option.
11870 2014-01-17 Richard Biener <rguenther@suse.de>
11872 PR rtl-optimization/38518
11873 * df.h (df_analyze_loop): Declare.
11874 * df-core.c: Include cfgloop.h.
11875 (df_analyze_1): Split out main part of df_analyze.
11876 (df_analyze): Adjust.
11877 (loop_inverted_post_order_compute): New function.
11878 (loop_post_order_compute): Likewise.
11879 (df_analyze_loop): New function avoiding whole-function
11880 postorder computes.
11881 * loop-invariant.c (find_defs): Use df_analyze_loop.
11882 (find_invariants): Adjust.
11883 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
11885 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
11887 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
11888 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
11890 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
11892 * ipa-ref.c (ipa_remove_stmt_references): Fix references
11893 traversal when removing references.
11895 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
11898 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
11900 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
11902 PR middle-end/56791
11903 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
11904 pushing a reload for an autoinc when we had previously reloaded an
11905 inner part of the address.
11907 2014-01-16 Jakub Jelinek <jakub@redhat.com>
11909 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
11911 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
11912 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
11913 when not giving up or versioning for alias only because of
11915 (vect_analyze_data_ref_dependences): Set to true.
11916 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
11918 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
11919 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
11922 PR middle-end/58344
11923 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
11926 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
11927 operand 0 predicate for gathers, use a new pseudo as subtarget.
11929 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11931 PR middle-end/59609
11932 * lra-constraints.c (process_alt_operands): Add printing debug info.
11933 Check absence of input/output reloads for matched operands too.
11935 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
11937 PR rtl-optimization/59835
11938 * ira.c (ira_init_register_move_cost): Increase cost for
11941 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
11943 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
11945 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
11948 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
11949 non-register objects. Use gen_(high/low)part more consistently.
11952 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11955 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
11956 endian support, remove tests for WORDS_BIG_ENDIAN.
11957 (p8_mfvsrd_3_<mode>): Likewise.
11958 (reload_gpr_from_vsx<mode>): Likewise.
11959 (reload_gpr_from_vsxsf): Likewise.
11960 (p8_mfvsrd_4_disf): Likewise.
11962 2014-01-16 Richard Biener <rguenther@suse.de>
11964 PR rtl-optimization/46590
11965 * lcm.c (compute_antinout_edge): Use postorder iteration.
11966 (compute_laterin): Use inverted postorder iteration.
11968 2014-01-16 Nick Clifton <nickc@redhat.com>
11970 PR middle-end/28865
11971 * varasm.c (output_constant): Return the number of bytes actually
11973 (output_constructor_array_range): Update the field size with the
11974 number of bytes emitted by output_constant.
11975 (output_constructor_regular_field): Likewise. Also do not
11976 complain if the total number of bytes emitted is now greater
11977 than the expected fieldpos.
11978 * output.h (output_constant): Update prototype and descriptive comment.
11980 2014-01-16 Marek Polacek <polacek@redhat.com>
11982 PR middle-end/59827
11983 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
11984 it is error_mark_node.
11986 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
11988 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
11989 VALID_AVX256_REG_OR_OI_MODE.
11991 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
11993 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
11994 current procedure should be profiled.
11996 2014-01-15 Andrew Pinski <apinski@cavium.com>
11998 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
11999 of moving from/to the STACK_REG register class.
12001 2014-01-15 Richard Henderson <rth@redhat.com>
12004 * reginfo.c (global_regs_decl): Globalize.
12005 * rtl.h (global_regs_decl): Declare.
12006 * ira.c (do_reload): Diagnose frame_pointer_needed and it
12007 reserved via global_regs.
12009 2014-01-15 Teresa Johnson <tejohnson@google.com>
12011 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
12013 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12015 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
12016 and vmulosh rather than call gen_vec_widen_smult_*.
12017 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
12018 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
12019 (vec_widen_smult_even_v16qi): Likewise.
12020 (vec_widen_umult_even_v8hi): Likewise.
12021 (vec_widen_smult_even_v8hi): Likewise.
12022 (vec_widen_umult_odd_v16qi): Likewise.
12023 (vec_widen_smult_odd_v16qi): Likewise.
12024 (vec_widen_umult_odd_v8hi): Likewise.
12025 (vec_widen_smult_odd_v8hi): Likewise.
12026 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
12027 vmuloub rather than call gen_vec_widen_umult_*.
12028 (vec_widen_umult_lo_v16qi): Likewise.
12029 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
12030 vmulosb rather than call gen_vec_widen_smult_*.
12031 (vec_widen_smult_lo_v16qi): Likewise.
12032 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
12033 rather than call gen_vec_widen_umult_*.
12034 (vec_widen_umult_lo_v8hi): Likewise.
12035 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
12036 rather than call gen_vec_widen_smult_*.
12037 (vec_widen_smult_lo_v8hi): Likewise.
12039 2014-01-15 Jeff Law <law@redhat.com>
12041 PR tree-optimization/59747
12042 * ree.c (find_and_remove_re): Properly handle case where a second
12043 eliminated extension requires widening a copy created for elimination
12044 of a prior extension.
12045 (combine_set_extension): Ensure that the number of hard regs needed
12046 for a destination register does not change when we widen it.
12048 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
12050 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
12051 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
12052 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
12053 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
12054 (avr-*-rtems*): Likewise.
12055 (bfin*-rtems*): Likewise.
12056 (moxie-*-rtems*): Likewise.
12057 (h8300-*-rtems*): Likewise.
12058 (i[34567]86-*-rtems*): Likewise.
12059 (lm32-*-rtems*): Likewise.
12060 (m32r-*-rtems*): Likewise.
12061 (m68k-*-rtems*): Likewise.
12062 (microblaze*-*-rtems*): Likewise.
12063 (mips*-*-rtems*): Likewise.
12064 (powerpc-*-rtems*): Likewise.
12065 (sh-*-rtems*): Likewise.
12066 (sparc-*-rtems*): Likewise.
12067 (sparc64-*-rtems*): Likewise.
12068 (v850-*-rtems*): Likewise.
12069 (m32c-*-rtems*): Likewise.
12071 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
12073 PR rtl-optimization/59511
12074 * ira.c (ira_init_register_move_cost): Use memory costs for some
12075 cases of register move cost calculations.
12076 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
12077 instead of BB frequency.
12078 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
12079 * lra-assigns.c (find_hard_regno_for): Ditto.
12081 2014-01-15 Richard Biener <rguenther@suse.de>
12083 PR tree-optimization/59822
12084 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
12085 (vectorizable_load): Use it to hoist defs of uses of invariant
12086 loads out of the loop.
12088 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12089 Kugan Vivekanandarajah <kuganv@linaro.org>
12092 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
12095 2014-01-15 Richard Biener <rguenther@suse.de>
12097 PR rtl-optimization/59802
12098 * lcm.c (compute_available): Use inverted postorder to seed
12099 the initial worklist.
12101 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12104 * config/s390/s390.c (s390_preferred_reload_class): Don't return
12105 ADDR_REGS for invalid symrefs in non-PIC code.
12107 2014-01-15 Jakub Jelinek <jakub@redhat.com>
12110 * builtins.c (determine_block_size): Initialize *probable_max_size
12111 even if len_rtx is CONST_INT.
12113 2014-01-14 Andrew Pinski <apinski@cavium.com>
12115 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
12116 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
12117 (cortexa53_tunings): Likewise.
12118 (aarch64_sched_issue_rate): New function.
12119 (TARGET_SCHED_ISSUE_RATE): Define.
12121 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12123 * ira-costs.c (find_costs_and_classes): Add missed
12124 ira_init_register_move_cost_if_necessary.
12126 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
12129 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
12131 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
12134 * config/i386/i386.c (type_natural_mode): Add a bool parameter
12135 to indicate if type is used for function return value. Warn ABI
12136 change if the vector mode isn't available for function return value.
12137 (ix86_function_arg_advance): Pass false to type_natural_mode.
12138 (ix86_function_arg): Likewise.
12139 (ix86_gimplify_va_arg): Likewise.
12140 (function_arg_32): Don't warn ABI change.
12141 (ix86_function_value): Pass true to type_natural_mode.
12142 (ix86_return_in_memory): Likewise.
12143 (ix86_struct_value_rtx): Removed.
12144 (TARGET_STRUCT_VALUE_RTX): Likewise.
12146 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
12148 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
12149 converting a conditional jump into a conditional return.
12151 2014-01-14 Richard Biener <rguenther@suse.de>
12153 PR tree-optimization/58921
12154 PR tree-optimization/59006
12155 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
12156 hoisting invariant stmts.
12157 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
12158 invariant loads on the preheader edge if possible.
12160 2014-01-14 Joey Ye <joey.ye@arm.com>
12162 * doc/plugin.texi (Building GCC plugins): Update to C++.
12164 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
12166 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
12167 (_mm_rcp28_round_ss): Ditto.
12168 (_mm_rsqrt28_round_sd): Ditto.
12169 (_mm_rsqrt28_round_ss): Ditto.
12170 (_mm_rcp28_sd): Ditto.
12171 (_mm_rcp28_ss): Ditto.
12172 (_mm_rsqrt28_sd): Ditto.
12173 (_mm_rsqrt28_ss): Ditto.
12174 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
12175 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
12176 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
12177 (IX86_BUILTIN_RCP28SD): Ditto.
12178 (IX86_BUILTIN_RCP28SS): Ditto.
12179 (IX86_BUILTIN_RSQRT28SD): Ditto.
12180 (IX86_BUILTIN_RSQRT28SS): Ditto.
12181 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
12182 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
12183 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
12184 (ix86_expand_special_args_builtin): Expand new FTYPE.
12185 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
12186 (srcp14<mode>): Make insn unary.
12187 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
12188 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
12189 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
12190 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
12191 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
12192 Fix rounding: make it SAE only.
12193 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
12195 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
12197 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
12198 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
12199 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
12200 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
12201 (round_saeonly_mask_scalar_operand4): Ditto.
12202 (round_saeonly_mask_scalar_op3): Ditto.
12203 (round_saeonly_mask_scalar_op4): Ditto.
12205 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12207 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
12208 Implement -maltivec=be for vec_insert and vec_extract.
12210 2014-01-10 DJ Delorie <dj@redhat.com>
12212 * config/msp430/msp430.md (call_internal): Don't allow memory
12213 references with SP as the base register.
12214 (call_value_internal): Likewise.
12215 * config/msp430/constraints.md (Yc): New. For memory references
12216 that don't use SP as a base register.
12218 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
12219 "an integer without a # prefix"
12220 * config/msp430/msp430.md (epilogue_helper): Use it.
12222 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12225 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
12226 AVX512F gather builtins.
12227 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
12228 on gather decls with INTEGER_TYPE masktype.
12229 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
12230 directly into the builtin rather than hoisting it before loop.
12232 PR tree-optimization/59387
12233 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
12234 (scev_const_prop): If folded_casts and type has undefined overflow,
12235 use force_gimple_operand instead of force_gimple_operand_gsi and
12236 for each added stmt if it is assign with
12237 arith_code_with_undefined_signed_overflow, call
12238 rewrite_to_defined_overflow.
12239 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
12240 gimple-fold.h instead.
12241 (arith_code_with_undefined_signed_overflow,
12242 rewrite_to_defined_overflow): Moved to ...
12243 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
12244 rewrite_to_defined_overflow): ... here. No longer static.
12245 Include gimplify-me.h.
12246 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
12247 rewrite_to_defined_overflow): New prototypes.
12249 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12251 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
12253 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
12255 * builtins.c (get_object_alignment_2): Minor tweak.
12256 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
12258 2014-01-13 Christian Bruel <christian.bruel@st.com>
12260 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
12261 optimized non constant lengths.
12263 2014-01-13 Jakub Jelinek <jakub@redhat.com>
12266 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
12267 load as __atomic_load_N if possible.
12269 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12271 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
12273 (rs6000_expand_builtin): Adjust call.
12275 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
12278 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
12279 * config/rs6000/rs6000.c: Include target-globals.h.
12280 (rs6000_set_current_function): Instead of doing target_reinit
12281 unconditionally, use save_target_globals_default_opts and
12282 restore_target_globals.
12284 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
12286 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
12287 (rs6000_expand_builtin): Handle mffs and mtfsf.
12288 (rs6000_init_builtins): Define mffs and mtfsf.
12289 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
12290 (rs6000_mffs): New pattern.
12291 (rs6000_mtfsf): New pattern.
12293 2014-01-11 Bin Cheng <bin.cheng@arm.com>
12295 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
12296 Start narrowing with START. Apply candidate-use pair
12297 and check overall cost in narrowing.
12298 (iv_ca_prune): Pass new argument.
12300 2014-01-10 Jeff Law <law@redhat.com>
12302 PR middle-end/59743
12303 * ree.c (combine_reaching_defs): Ensure the defining statement
12304 occurs before the extension when optimizing extensions with
12305 different source and destination hard registers.
12307 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12310 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
12311 vtables into the type inheritance graph.
12313 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12315 PR rtl-optimization/59754
12316 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
12317 modes in the REGNO != REGNO case.
12319 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12321 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
12323 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12325 PR tree-optimization/59745
12326 * tree-predcom.c (tree_predictive_commoning_loop): Call
12327 free_affine_expand_cache if giving up because components is NULL.
12329 * target-globals.c (save_target_globals): Allocate < 4KB structs using
12330 GC in payload of target_globals struct instead of allocating them on
12331 the heap and the larger structs separately using GC.
12332 * target-globals.h (struct target_globals): Make regs, hard_regs,
12333 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
12334 of GTY((skip)) and change type to void *.
12335 (reset_target_globals): Cast loads from those fields to corresponding
12338 2014-01-10 Steve Ellcey <sellcey@mips.com>
12341 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
12342 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
12343 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
12345 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
12348 * aarch64-modes.def (CC_Zmode): New flags mode.
12349 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
12350 represents an equality.
12351 (aarch64_get_condition_code): Handle CC_Zmode.
12352 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
12354 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12356 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
12357 extraction in good case.
12359 2014-01-10 Richard Biener <rguenther@suse.de>
12361 PR tree-optimization/59374
12362 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
12363 checking after SLP discovery. Mark stmts not participating
12364 in any SLP instance properly.
12366 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12368 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
12369 when handling a SET rtx.
12371 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12373 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
12374 (cortex-a57): Likewise.
12375 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
12377 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12379 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
12380 non-iwmmxt builtins.
12382 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
12386 * ipa-devirt.c record_target_from_binfo): Take as argument
12387 stack of binfos and lookup matching one for virtual inheritance.
12388 (possible_polymorphic_call_targets_1): Update.
12390 2014-01-10 Huacai Chen <chenhc@lemote.com>
12392 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
12393 kernel strings for Loongson-2E/2F/3A.
12395 2014-01-10 Jakub Jelinek <jakub@redhat.com>
12397 PR middle-end/59670
12398 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
12399 is_gimple_call before calling gimple_call_internal_p.
12401 2014-01-09 Steve Ellcey <sellcey@mips.com>
12403 * Makefile.in (TREE_FLOW_H): Remove.
12404 (TREE_SSA_H): Add file names from tree-flow.h.
12405 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
12406 * tree.h: Remove tree-flow.h reference.
12407 * hash-table.h: Remove tree-flow.h reference.
12408 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
12409 reference with tree-ssa-loop.h.
12411 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12413 * doc/invoke.texi: Add -maltivec={be,le} options, and document
12414 default element-order behavior for -maltivec.
12415 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
12416 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
12417 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
12418 when targeting big endian, at least for now.
12419 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
12421 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12423 PR middle-end/47735
12424 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
12425 var satisfies use_register_for_decl, just take into account type
12426 alignment, rather than decl alignment.
12428 PR tree-optimization/59622
12429 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
12430 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
12431 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
12432 Don't devirtualize for inplace at all. For targets.length () == 1,
12433 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
12435 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12437 * config/i386/i386.md (cpu): Remove the unused btver1.
12439 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
12441 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
12443 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12446 * tree-core.h (struct target_globals): New forward declaration.
12447 (struct tree_target_option): Add globals field.
12448 * tree.h (TREE_TARGET_GLOBALS): Define.
12449 (prepare_target_option_nodes_for_pch): New prototype.
12450 * target-globals.h (struct target_globals): Define even if
12451 !SWITCHABLE_TARGET.
12452 * tree.c (prepare_target_option_node_for_pch,
12453 prepare_target_option_nodes_for_pch): New functions.
12454 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
12455 * config/i386/i386.c: Include target-globals.h.
12456 (ix86_set_current_function): Instead of doing target_reinit
12457 unconditionally, use save_target_globals_default_opts and
12458 restore_target_globals.
12460 2014-01-09 Richard Biener <rguenther@suse.de>
12462 PR tree-optimization/59715
12463 * tree-cfg.h (split_critical_edges): Declare.
12464 * tree-cfg.c (split_critical_edges): Export.
12465 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
12467 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
12469 * cfgexpand.c (expand_stack_vars): Optionally disable
12470 asan stack protection.
12471 (expand_used_vars): Likewise.
12472 (partition_stack_vars): Likewise.
12473 * asan.c (asan_emit_stack_protection): Optionally disable
12474 after return stack usage.
12475 (instrument_derefs): Optionally disable memory access instrumentation.
12476 (instrument_builtin_call): Likewise.
12477 (instrument_strlen_call): Likewise.
12478 (asan_protect_global): Optionally disable global variables protection.
12479 * doc/invoke.texi: Added doc for new options.
12480 * params.def: Added new options.
12481 * params.h: Likewise.
12483 2014-01-09 Jakub Jelinek <jakub@redhat.com>
12485 PR rtl-optimization/59724
12486 * ifcvt.c (cond_exec_process_if_block): Don't call
12487 flow_find_head_matching_sequence with 0 longest_match.
12488 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
12489 non-active insns if !stop_after.
12490 (try_head_merge_bb): Revert 2014-01-07 changes.
12492 2014-01-08 Jeff Law <law@redhat.com>
12494 * ree.c (get_sub_rtx): New function, extracted from...
12495 (merge_def_and_ext): Here.
12496 (combine_reaching_defs): Use get_sub_rtx.
12498 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
12500 * cgraph.h (varpool_variable_node): Do not choke on null node.
12502 2014-01-08 Catherine Moore <clm@codesourcery.com>
12504 * config/mips/mips.md (simple_return): Attempt to use JRC
12506 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
12508 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12510 PR rtl-optimization/59137
12511 * reorg.c (steal_delay_list_from_target): Call update_block for
12513 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
12515 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12517 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
12518 two duplicate entries.
12520 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
12523 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
12525 * config/mips/mips.c (mips_truncated_op_cost): New function.
12526 (mips_rtx_costs): Adjust test for BADDU.
12527 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
12529 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
12531 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
12532 (*baddu_si): ...this new pattern.
12534 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12537 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
12539 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12541 PR middle-end/57748
12542 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
12544 (expand_expr, expand_normal): Adjust.
12545 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
12546 inner_reference_p. Use inner_reference_p to expand inner references.
12547 (store_expr): Adjust.
12548 * cfgexpand.c (expand_call_stmt): Adjust.
12550 2014-01-08 Rong Xu <xur@google.com>
12552 * gcov-io.c (gcov_var): Move from gcov-io.h.
12553 (gcov_position): Ditto.
12554 (gcov_is_error): Ditto.
12555 (gcov_rewrite): Ditto.
12556 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
12557 only part to libgcc/libgcov.h.
12559 2014-01-08 Marek Polacek <polacek@redhat.com>
12561 PR middle-end/59669
12562 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
12564 2014-01-08 Marek Polacek <polacek@redhat.com>
12567 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
12569 2014-01-08 Jakub Jelinek <jakub@redhat.com>
12571 PR rtl-optimization/59649
12572 * stor-layout.c (get_mode_bounds): For BImode return
12573 0 and STORE_FLAG_VALUE.
12575 2014-01-08 Richard Biener <rguenther@suse.de>
12577 PR middle-end/59630
12578 * gimple.h (is_gimple_builtin_call): Remove.
12579 (gimple_builtin_call_types_compatible_p): New.
12580 (gimple_call_builtin_p): New overload.
12581 * gimple.c (is_gimple_builtin_call): Remove.
12582 (validate_call): Rename to ...
12583 (gimple_builtin_call_types_compatible_p): ... this and export. Also
12584 check return types.
12585 (validate_type): New static function.
12586 (gimple_call_builtin_p): New overload and adjust.
12587 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
12588 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
12589 (gimple_fold_stmt_to_constant_1): Likewise.
12590 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
12592 2014-01-08 Richard Biener <rguenther@suse.de>
12594 PR middle-end/59471
12595 * gimplify.c (gimplify_expr): Gimplify register-register type
12596 VIEW_CONVERT_EXPRs to separate stmts.
12598 2014-01-07 Jeff Law <law@redhat.com>
12600 PR middle-end/53623
12601 * ree.c (combine_set_extension): Handle case where source
12602 and destination registers in an extension insn are different.
12603 (combine_reaching_defs): Allow source and destination registers
12604 in extension to be different under limited circumstances.
12605 (add_removable_extension): Remove restriction that the
12606 source and destination registers in the extension are the same.
12607 (find_and_remove_re): Emit a copy from the extension's
12608 destination to its source after the defining insn if
12609 the source and destination registers are different.
12611 PR middle-end/59285
12612 * ifcvt.c (merge_if_block): If we are merging a block with more than
12613 one successor with a block with no successors, remove any BARRIER
12614 after the second block.
12616 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
12618 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
12620 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
12623 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
12624 for 14-bit register offsets when INT14_OK_STRICT is false.
12626 2014-01-07 Roland Stigge <stigge@antcom.de>
12627 Michael Meissner <meissner@linux.vnet.ibm.com>
12630 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
12631 Only check TFmode for SPE constants. Don't check TImode or TDmode.
12633 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
12635 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
12638 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
12640 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
12641 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
12642 rtx is const0_rtx or not.
12644 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
12647 * target-globals.c (save_target_globals): Remove this_fn_optab
12649 * toplev.c: Include optabs.h.
12650 (target_reinit): Temporarily restore the global options if another
12651 set of options are in force.
12653 2014-01-07 Jakub Jelinek <jakub@redhat.com>
12655 PR rtl-optimization/58668
12656 * cfgcleanup.c (flow_find_cross_jump): Don't count
12657 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
12658 to determine what is counted.
12659 (flow_find_head_matching_sequence): Use active_insn_p to determine
12661 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
12663 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
12664 determine what is counted.
12666 PR tree-optimization/59643
12667 * tree-predcom.c (split_data_refs_to_components): If one dr is
12668 read and one write, determine_offset fails and the write isn't
12669 in the bad component, just put the read into the bad component.
12671 2014-01-07 Mike Stump <mikestump@comcast.net>
12672 Jakub Jelinek <jakub@redhat.com>
12675 * tree-core.h (struct tree_optimization_option): Change optabs
12676 type from unsigned char * to void *.
12677 * optabs.c (init_tree_optimization_optabs): Adjust
12678 TREE_OPTIMIZATION_OPTABS initialization.
12680 2014-01-06 Jakub Jelinek <jakub@redhat.com>
12683 * config/i386/i386.h (struct machine_function): Add
12684 no_drap_save_restore field.
12685 * config/i386/i386.c (ix86_save_reg): Use
12686 !cfun->machine->no_drap_save_restore instead of
12687 crtl->stack_realign_needed.
12688 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
12689 this function clears frame_pointer_needed. Set
12690 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
12691 and DRAP reg is needed.
12693 2014-01-06 Marek Polacek <polacek@redhat.com>
12696 * doc/implement-c.texi: Mention that other integer types are
12697 permitted as bit-field types in strictly conforming mode.
12699 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
12701 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
12702 is newly allocated.
12704 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
12706 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
12708 2014-01-06 Martin Jambor <mjambor@suse.cz>
12711 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
12713 * ipa-prop.c (ipa_print_node_params): Fix indentation.
12715 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
12719 * var-tracking.c (add_stores): Preserve the value of the source even if
12720 we don't record the store.
12722 2014-01-06 Terry Guo <terry.guo@arm.com>
12724 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
12726 2014-01-05 Iain Sandoe <iain@codesourcery.com>
12729 * config/darwin.c (darwin_function_section): Adjust return values to
12730 correspond to optimisation changes made in r206070.
12732 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
12734 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
12735 from prefetch_block tune setting.
12736 (nocona_cost): Correct size of prefetch block to 64.
12738 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
12740 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
12741 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
12742 used to save the static chain register in the computation of the offset
12743 from which the FP registers need to be restored.
12745 2014-01-04 Jakub Jelinek <jakub@redhat.com>
12747 PR tree-optimization/59519
12748 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
12749 ICE if get_current_def (current_new_name) is already non-NULL, as long
12750 as it is a phi result of some other phi in *new_exit_bb that has
12753 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
12754 or vmovdqu* for misaligned_operand.
12755 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
12756 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
12757 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
12758 aligned_mem for AVX512F masked aligned load and store builtins and for
12759 non-temporal moves.
12761 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
12763 PR tree-optimization/59651
12764 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
12765 Address range for negative step should be added by TYPE_SIZE_UNIT.
12767 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
12769 * config/m68k/m68k.c (handle_move_double): Handle pushes with
12770 overlapping registers also for registers other than the stack pointer.
12772 2014-01-03 Marek Polacek <polacek@redhat.com>
12775 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
12778 2014-01-03 Jakub Jelinek <jakub@redhat.com>
12781 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
12784 * config/i386/i386.md (MODE_SIZE): New mode attribute.
12785 (push splitter): Use <P:MODE_SIZE> instead of
12786 GET_MODE_SIZE (<P:MODE>mode).
12787 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
12788 (mov -1, reg peephole2): Likewise.
12789 * config/i386/sse.md (*mov<mode>_internal,
12790 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
12791 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
12792 *<code><mode>3, *andnot<mode>3<mask_name>,
12793 <mask_codefor><code><mode>3<mask_name>): Likewise.
12794 * config/i386/subst.md (mask_mode512bit_condition,
12795 sd_mask_mode512bit_condition): Likewise.
12797 2014-01-02 Xinliang David Li <davidxl@google.com>
12799 PR tree-optimization/59303
12800 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
12801 (dump_predicates): Better output format.
12802 (pred_equal_p): New function.
12803 (is_neq_relop_p): Ditto.
12804 (is_neq_zero_form_p): Ditto.
12805 (pred_expr_equal_p): Ditto.
12806 (pred_neg_p): Ditto.
12807 (simplify_pred): Ditto.
12808 (simplify_preds_2): Ditto.
12809 (simplify_preds_3): Ditto.
12810 (simplify_preds_4): Ditto.
12811 (simplify_preds): Ditto.
12812 (push_pred): Ditto.
12813 (push_to_worklist): Ditto.
12814 (get_pred_info_from_cmp): Ditto.
12815 (is_degenerated_phi): Ditto.
12816 (normalize_one_pred_1): Ditto.
12817 (normalize_one_pred): Ditto.
12818 (normalize_one_pred_chain): Ditto.
12819 (normalize_preds): Ditto.
12820 (normalize_cond_1): Remove function.
12821 (normalize_cond): Ditto.
12822 (is_gcond_subset_of): Ditto.
12823 (is_subset_of_any): Ditto.
12824 (is_or_set_subset_of): Ditto.
12825 (is_and_set_subset_of): Ditto.
12826 (is_norm_cond_subset_of): Ditto.
12827 (pred_chain_length_cmp): Ditto.
12828 (convert_control_dep_chain_into_preds): Type change.
12829 (find_predicates): Ditto.
12830 (find_def_preds): Ditto.
12831 (destroy_predicates_vecs): Ditto.
12832 (find_matching_predicates_in_rest_chains): Ditto.
12833 (use_pred_not_overlap_with_undef_path_pred): Ditto.
12834 (is_pred_expr_subset): Ditto.
12835 (is_pred_chain_subset_of): Ditto.
12836 (is_included_in): Ditto.
12837 (is_superset_of): Ditto.
12839 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12841 Update copyright years.
12843 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
12845 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
12846 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
12847 config/arc/arc.md, config/arc/arc.opt,
12848 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
12849 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
12850 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
12851 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
12852 config/linux-protos.h, config/linux.c, config/winnt-c.c,
12853 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
12854 vtable-verify.c, vtable-verify.h: Use the standard form for the
12857 2014-01-02 Tobias Burnus <burnus@net-b.de>
12859 * gcc.c (process_command): Update copyright notice dates.
12860 * gcov-dump.c: Ditto.
12862 * doc/cpp.texi: Bump @copying's copyright year.
12863 * doc/cppinternals.texi: Ditto.
12864 * doc/gcc.texi: Ditto.
12865 * doc/gccint.texi: Ditto.
12866 * doc/gcov.texi: Ditto.
12867 * doc/install.texi: Ditto.
12868 * doc/invoke.texi: Ditto.
12870 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12872 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
12874 2014-01-01 Jakub Jelinek <jakub@redhat.com>
12876 * config/i386/sse.md (*mov<mode>_internal): Guard
12877 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
12879 PR rtl-optimization/59647
12880 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
12881 new_rtx into UNSIGNED_FLOAT rtxes.
12883 Copyright (C) 2014 Free Software Foundation, Inc.
12885 Copying and distribution of this file, with or without modification,
12886 are permitted in any medium without royalty provided the copyright
12887 notice and this notice are preserved.